preliminary_analysis_SCIFAR10.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preliminary analysis on  Split-CIFAR10"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The purpose of this notebook is to perform preliminary analysis to the DER++ implementation with SplitCIFAR10, reproducing the paper experiments.\n",
    "\n",
    "Evaluation is performed directly on the test set,  as we use the paper best found hyperparameters.\n",
    "\n",
    "5 experiences of 2 classes each are used."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "import torch\n",
    "\n",
    "from torchvision.transforms import Compose, Normalize, ToTensor\n",
    "\n",
    "from avalanche.benchmarks.classic import SplitCIFAR10\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Following the DER paper, 5 experiences  were used, each with 2 new classes.\n",
    "\n",
    "We are in a Class-Incremental setting, without task labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_experiences = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "# No transformations as they are applied in the DER plugin\n",
    "split_cifar10 = SplitCIFAR10(n_experiences=num_experiences, seed=42, train_transform=Compose([ToTensor(), Normalize(mean=(0.4914, 0.4822, 0.4465), std=(0.2023, 0.1994, 0.201))]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000\n",
      "2000\n"
     ]
    }
   ],
   "source": [
    "print(len(split_cifar10.train_stream[0].dataset))\n",
    "print(len(split_cifar10.test_stream[0].dataset))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "# With transformations for Naive and Replay training\n",
    "split_cifar10_transform = SplitCIFAR10(n_experiences=num_experiences, seed=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000\n",
      "2000\n"
     ]
    }
   ],
   "source": [
    "print(len(split_cifar10_transform.train_stream[0].dataset))\n",
    "print(len(split_cifar10_transform.test_stream[0].dataset))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Show some samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaEAAAGdCAYAAAC7EMwUAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAm2UlEQVR4nO3df3TU9Z3v8ddIw0g0mV0aSSYl5EQKUg1yVrERpBrokmt0uVpaL5W73bDbdYsCXQ562Ea612x3l3DZK9f2UumpeimcSmF3K9SzoJAtJEgpyM8SUVmEaOIxMReOzoSASQif+wfr1AjC9w0zfJLJ83HOnENm3nnn853PTF7nS2beE3LOOQEA4MFVvhcAAOi/CCEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3nzO9wI+7cyZM3rvvfeUlZWlUCjkezkAACPnnNra2pSfn6+rrrrwuU6vC6H33ntPBQUFvpcBALhMTU1NGjp06AVrUvbfcU8//bSKiop09dVX69Zbb9Urr7wS6PuysrJStSQAwBUU5Pd5SkJozZo1mjt3rhYsWKB9+/bpK1/5isrLy9XY2HjR7+W/4AAgPQT5fR5KxQDTkpIS3XLLLVq2bFniui996Uu6//77VV1dfcHvjcfjikQiyV4SAOAKi8Viys7OvmBN0s+EOjs7tWfPHpWVlfW4vqysTNu3bz+nvqOjQ/F4vMcFANA/JD2Ejh07pu7ubuXm5va4Pjc3Vy0tLefUV1dXKxKJJC68KAEA+o+UvTDh0/8X6Jw77/8PVlZWKhaLJS5NTU2pWhIAoJdJ+ku0c3JyNGDAgHPOelpbW885O5KkcDiscDic7GUAAPqApJ8JDRw4ULfeeqtqamp6XF9TU6Px48cn+8cBAPqwlLxZdd68efrWt76lsWPHaty4cfrpT3+qxsZGzZw5MxU/DgDQR6UkhKZNm6bjx4/rBz/4gZqbm1VcXKwNGzaosLAwFT8OANBHpeR9QpeD9wmloVGG/Xwzlrp1pNCND0RN9RNGDzfVXz8heP3Rkxd/U/gnHTlyJHDtr9fZeudmBK99f5Opdf9heWiNMvbOMdSeNNR2Sdrk6X1CAAAERQgBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALxJyew44JOe3bw8cO2GbbWm3i/M+FHw4kxTa/39Ow8Hrv1+5t/Zmmunsf6Ysd7CMBfmu9axSsE/H6xRwccHSVJTLPh9khGxzbPJlG10WEzNht42gxR89tEw5Zl6Zxp6d6krcG08/pFyIt8LVMuZEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOANIQQA8IbZcUi5v8yfmrrmtwQv/dP/9V1T6+9nfttQfZ2pt1RirLewzT2TYSaYjPPdpGGGykG2zpFThurPm3rLMFPtrHsMtZb7W5IaU9j7ZODKDFNtKHAtZ0IAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN4ztQd+2N3jpzyf9yNT6y/8WfETNnHv/zdTbPuYnlQYaaq0jgQwjZ/b+xNQ59qP6wLVNW0ytFbzzWQ9uHhW8ePhoY/c7DLXW/bGMJ7L0Dt6XMyEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOBNyDnnfC/ik+LxuCKRiL4bu1fh7GDzh3Y+tC5w/63PXuLCgCRy7qfG73goJeuwO2Ar3/I/g9dOWmVqXWOofdPUWaoy1ncZanOMvf9oevDaXz5fbex+vaH2ZODKePyUIpFHFIvFlJ2dfcFazoQAAN4kPYSqqqoUCoV6XPLy8pL9YwAAaSAlH+Vw00036d///d8TXw8YMCAVPwYA0MelJIQ+97nPcfYDALiolPxN6PDhw8rPz1dRUZG++c1v6ujRo59Z29HRoXg83uMCAOgfkh5CJSUlWrlypTZu3KhnnnlGLS0tGj9+vI4fP37e+urqakUikcSloKAg2UsCAPRSSQ+h8vJyff3rX9fo0aP1x3/8x1q/fr0kacWKFeetr6ysVCwWS1yampqSvSQAQC+Vkr8JfdI111yj0aNH6/Dhw+e9PRwOKxwOp3oZAIBeKOXvE+ro6NAbb7yhaDSa6h8FAOhjkh5Cjz32mOrq6tTQ0KCdO3fqG9/4huLxuCoqKpL9owAAfVzS/zvu3Xff1YMPPqhjx47puuuu0+23364dO3aosLDQ1Ofv9YSydW2g2ucePRK479Zn603rwOW72lA7zNg7Zqi9wdi7xFC73th7fuivTPWL/3ll8OIHymyLebMxeG1jsFFaCfWWgTY2e1PW2fa4kqRuQ22bsXezZZrR8zuN3T9vrA8q+L4nPYRWr16d7JYAgDTF7DgAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAm5R/lMOly5WUHajyOebBnWOoobbY2NswaUyS9Nmfq3uu/zD2HmCovd7Y+48MtbONva0z8vTX24LX/shQK9kGpWXaWhtGiEkRW+uJhnUHny55lmXvJWm3sd7CNvNwuLH7+T9s9PJ9FLiSMyEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAm148tucPFHRsT8axlC6kT2o21H7Z2DtqrH/dUGsZwyNJpYZa67pfNdZbmMf2WDbUMoZHkk4a6w1+Y6i1jsqxPG4td58kjTLWp3Jsz+8Mta9tW2bqXTzhnwLXnjQM4DqpjsC1nAkBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvevHsOPefl4uLWmdl9QPdhlrrjLTPG+st5hnr7zHUPmfsbZk3Zh1faJ1lZpp7l8JZcK8Z62sMtV3G3qWGWuv9bZl515v86STb5u/vPBW4NlORwLWn9VHgWs6EAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN714dtxaSZmBKl9dl9KFpD3rXK08Y/1QQ+19xt53GGo3GHuvNNQGn6p11kxjvUUKR8ep0Vi/LUW1knS9odY6XrLBWN9b/M46gM/07Lc07whcyZkQAMAbcwht3bpVU6ZMUX5+vkKhkNatW9fjduecqqqqlJ+fr0GDBqm0tFQHDx5M1noBAGnEHELt7e0aM2aMli5det7bFy9erCVLlmjp0qXatWuX8vLyNHnyZLW1tV32YgEA6cX8N6Hy8nKVl5ef9zbnnJ566iktWLBAU6dOlSStWLFCubm5WrVqlb7zne9c3moBAGklqX8TamhoUEtLi8rKyhLXhcNh3XXXXdq+fft5v6ejo0PxeLzHBQDQPyQ1hFpaWiRJubm5Pa7Pzc1N3PZp1dXVikQiiUtBQUEylwQA6MVS8uq4UCjU42vn3DnXfayyslKxWCxxaWpqSsWSAAC9UFLfJ5SXd/YdJC0tLYpGo4nrW1tbzzk7+lg4HFY4HE7mMgAAfURSz4SKioqUl5enmpqaxHWdnZ2qq6vT+PHjk/mjAABpwHwmdOLECb311luJrxsaGrR//34NHjxYw4YN09y5c7Vw4UKNGDFCI0aM0MKFC5WZmanp06cndeEAgL7PHEK7d+/WxIkTE1/PmzdPklRRUaGf/exnmj9/vk6dOqVHHnlEH3zwgUpKSrRp0yZlZWUZf9L7kq4OVGkdO4Oeuo311tcv/pmh1jKGR5JpXs5k4+yWfzLU7rS1No/WsdQfSmHvfzX2/rWx3mK+oXZYylbR1wUfxRMzDG2KG/qaQ6i0tFTOuc+8PRQKqaqqSlVVVdbWAIB+htlxAABvCCEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDdJ/SiH5Br0n5eLi168JKHhktaSIhnBS588YDlK6dEvpW6iXrBd+b17DLXBp1Od1WWYB2ed12bxO2P9ypSs4qzgU7vOetVQ+ytj71RaY6j905Stom+zPCcyDI8sSy1nQgAAbwghAIA3hBAAwBtCCADgDSEEAPCGEAIAeEMIAQC8IYQAAN4QQgAAbwghAIA3vXhszxhJ1waqnD8seNeHrXNhUuiNzu8Grh2lvzP1/jN3Z+Da60L1pt7WETV7DbW/MPa2bKd1nE0qbTDWDzfUGiYZSbKN4vnI2Lu3+LnvBfRSmbrBUB38GXRanYFrORMCAHhDCAEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADe9OLZce9LagtUmTl8YvC2jVsubTkBPP6GrX6Uvm2otk0+yzFMeHvDjTH1/pJx1lzwCXn9R46x3nKPbzX2Rhq51/oNtxhqXzXUMjsOANAHEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG967diemP63XMDlVWzZlrJ1OPdrQ/UkY/d2Q+1OY+9Y4MpR2m3qfOfwsKl+6xFTeZ9UlOL+v0lxf6SJDOs31AaubFTwJ3KbTgeu5UwIAOANIQQA8MYcQlu3btWUKVOUn5+vUCikdevW9bh9xowZCoVCPS633357stYLAEgj5hBqb2/XmDFjtHTp0s+sufvuu9Xc3Jy4bNiw4bIWCQBIT+YXJpSXl6u8vPyCNeFwWHl5eZe8KABA/5CSvwnV1tZqyJAhGjlypB566CG1trZ+Zm1HR4fi8XiPCwCgf0h6CJWXl+v555/X5s2b9eSTT2rXrl2aNGmSOjo6zltfXV2tSCSSuBQUFCR7SQCAXirp7xOaNm1a4t/FxcUaO3asCgsLtX79ek2dOvWc+srKSs2bNy/xdTweJ4gAoJ9I+ZtVo9GoCgsLdfjw4fPeHg6HFQ7b3vwIAEgPKX+f0PHjx9XU1KRoNJrqHwUA6GPMZ0InTpzQW2+9lfi6oaFB+/fv1+DBgzV48GBVVVXp61//uqLRqN5++209/vjjysnJ0de+9rWkLhwA0PeZQ2j37t2aOHFi4uuP/55TUVGhZcuWqb6+XitXrtSHH36oaDSqiRMnas2aNcrKyjL9nIj+Qdm6JmD1uMB9p/3MtAzZ5sF1Gnu/aahtNva2DJEaaOr847dKTPWjQ9a5d33P/zXWW985122s7w/+/IHgtcv/JXXrkKSsh4PXti1L3Tq0zvoNXYErMxQx1Abvaw6h0tJSOec+8/aNGzdaWwIA+ilmxwEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADehNyFZvB4EI/HFYlEFIvFlJ2dHfC73rp4SUKmcUXB5yVJMWPv4POVbLPgJNtacoy9t5mqQ6FzP0cKOMcoW7l743eBa0OhMcbF2HzVMK3M+mxb81+M32DwL25C4Npv6BuBa+PxjxSJfC/Q73HOhAAA3hBCAABvCCEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvevHYnnplZ2cF/K5Cw0/4f8YVWUbrWEcCnUxhb8vYnkZj79Gm6mNdcwPXXjdwhXEt6NWCT4WRe+XHxuZlgStDfzDC1to6gcswVct1/qWpdSj0rHExwRU9H7z26PRfB66Nx9sVifxXxvYAAHo3QggA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADw5nO+F/DZ3pV0TQr6WodCRQy1ljlzkpRjqB1o7G2ZS2eplaQjpuqcjOWG6l40O2548FL31hlT65j+wVT/f2LrAtfW7mw29T6WcSxw7cKJtrln92iOodq2bmll4Mqpz9g6v/DfjEsxPfWNzzfDXDrrr6CGnYbi6ZZ1nwpcyZkQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4E0vHtuToeDzKizjPixjeKxs42xsY3sstZLtPokae1uFUtw/NaattVTbjjGivzXVfz8SvP77Zc7U2/a4DT7i5yzLmCzr+KjrA1d++4HRps4vqN64Fot7TNU7O4PfhyWh9balmCYlWX4HtQeu5EwIAOCNKYSqq6t12223KSsrS0OGDNH999+vQ4cO9ahxzqmqqkr5+fkaNGiQSktLdfDgwaQuGgCQHkwhVFdXp1mzZmnHjh2qqanR6dOnVVZWpvb23596LV68WEuWLNHSpUu1a9cu5eXlafLkyWpra0v64gEAfZvpb0Ivv/xyj6+XL1+uIUOGaM+ePbrzzjvlnNNTTz2lBQsWaOrUqZKkFStWKDc3V6tWrdJ3vvOd5K0cANDnXdbfhGKxs38wGzx4sCSpoaFBLS0tKisrS9SEw2Hddddd2r59+3l7dHR0KB6P97gAAPqHSw4h55zmzZunCRMmqLi4WJLU0tIiScrNze1Rm5ubm7jt06qrqxWJRBKXgoKCS10SAKCPueQQmj17tg4cOKBf/OIX59wWCvV8uapz7pzrPlZZWalYLJa4NDU1XeqSAAB9zCW9T2jOnDl68cUXtXXrVg0dOjRxfV5enqSzZ0TR6O/fe9La2nrO2dHHwuGwwuHwpSwDANDHmc6EnHOaPXu2XnjhBW3evFlFRUU9bi8qKlJeXp5qamoS13V2dqqurk7jx49PzooBAGnDdCY0a9YsrVq1Sr/61a+UlZWV+DtPJBLRoEGDFAqFNHfuXC1cuFAjRozQiBEjtHDhQmVmZmr69OkpOQAAQN9lCqFly5ZJkkpLS3tcv3z5cs2YMUOSNH/+fJ06dUqPPPKIPvjgA5WUlGjTpk3KyspKyoIBAOkj5JyzDppKqXg8rkgkoljsJWVnXxPwuzINP2GYcUVdhtptxt5HDbWDjL1PG2rvMPa23N+SZeZU6L9/wdZ6lXEpBuvcvYFr79O/pW4hkqR3DLWmgWCyzXdrNPYOOv9Rsj3XJMtzOabXTJ3/IPSYcS3BOXfY+B3Bf6+EQn9ua1128ZKPuY27A9fG4ycUiZQqFospOzv7grXMjgMAeEMIAQC8IYQAAN4QQgAAbwghAIA3hBAAwBtCCADgDSEEAPCGEAIAeEMIAQC8uaSPcuh9ThpqrWNHjhhqrWNHLB/gZxmtItnG9ljvE8v9bTMm+KQcSdLvUji2Z7i+nbrmpjE8krTXUBsx9rawjOGRpOOBKxtNxygNM4zt+YV2mnqbmaaBrTe1jimFnzZtulss46CC/47gTAgA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHjTa2fHtei42nUqUO1xw+yzzyvTtI5a/SpwbalGmXpH9WeGautcOku9bXbca3rNVF9gmCMVT91YOrO/WvGvgWu3V0w0dt9mrG8x1GabOjcb9vOQaZai1GV4HDabZpNJow29Z//1FlNvs2jw0lr9xtQ60zaYzsYwkrLLsD9dAX93S5wJAQA8IoQAAN4QQgAAbwghAIA3hBAAwBtCCADgDSEEAPCGEAIAeEMIAQC8IYQAAN702rE9r2iHMhUOVDtMBYH7Xm+ZryFpmIYHrs1Ujqm3TPVvGnvvC1zZbJndcQksA4SardOJUui3M1YFrq2qsI1WqdIdxtUEH9vTpeOmzvsMo3hqjeOGjhkeW1Hj8yfT8Fzu/pGptV3wXxNqND7fYsZRSamy0jBu6JQ6A9dyJgQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOANIQQA8IYQAgB4QwgBALzptbPjbtJQXaurA9XeoOBzuzKUaVrHHbrHUG2bH2aZB3dEG0ydX9PewLV3qMzUu8s4+2qfYS3G7ek1/i60yFZvfKgseif4Ht1rGWQmqUuRwLWf1yhT7wzDY+WYjpl6/yS2yVSfUoaZhzFlmFq/pkbjYlLjVcM6OnU6cC1nQgAAb0whVF1drdtuu01ZWVkaMmSI7r//fh06dKhHzYwZMxQKhXpcbr/99qQuGgCQHkwhVFdXp1mzZmnHjh2qqanR6dOnVVZWpvb29h51d999t5qbmxOXDRts/5UEAOgfTH8Tevnll3t8vXz5cg0ZMkR79uzRnXfembg+HA4rLy8vOSsEAKSty/qbUCx29o+OgwcP7nF9bW2thgwZopEjR+qhhx5Sa2vrZ/bo6OhQPB7vcQEA9A+XHELOOc2bN08TJkxQcXFx4vry8nI9//zz2rx5s5588knt2rVLkyZNUkdHx3n7VFdXKxKJJC4FBcE/JRUA0Ldd8ku0Z8+erQMHDmjbtp4f9ztt2rTEv4uLizV27FgVFhZq/fr1mjp16jl9KisrNW/evMTX8XicIAKAfuKSQmjOnDl68cUXtXXrVg0dOvSCtdFoVIWFhTp8+PB5bw+HwwqHw5eyDABAH2cKIeec5syZo7Vr16q2tlZFRUUX/Z7jx4+rqalJ0Wj0khcJAEhPpr8JzZo1Sz//+c+1atUqZWVlqaWlRS0tLTp16pQk6cSJE3rsscf029/+Vm+//bZqa2s1ZcoU5eTk6Gtf+1pKDgAA0HeZzoSWLVsmSSotLe1x/fLlyzVjxgwNGDBA9fX1WrlypT788ENFo1FNnDhRa9asUVZWVtIWDQBID+b/jruQQYMGaePGjZe1oI/lKq4sdQaqzTDNVhptXIlhKJROpqz3cN1g6jxMwV/ckaESU+8u/cZUHzPMjrvhFlNr/c5SbB3tZ6nfdvGSHozjwL4XCj4nLcPZZsedVE7g2uPGuYGWWY1dxplqv/6RqTy1DPu/+M31ptaf7yXzFLft3Rm4tvvEhbPik5gdBwDwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwhhACAHhzyZ8nlGrHdVgdAcd4HDOMvxllXollzI91LkzwcSnW3hmmcUNHTL1PGke3ZCr4BPXjtqXYGEfl/NU79wau/WnINoolleavst2JwycEr79n2ERT70bDnV6zt9nU2zwqycI4PkoPBC991/hQedcytucHtt6GiVp6vdEwlsxQypkQAMAbQggA4A0hBADwhhACAHhDCAEAvCGEAADeEEIAAG8IIQCAN4QQAMAbQggA4A0hBADwptfOjtujFg0KuLxmBZ851SLLICbpjoDz6yTrvDZJphlswddxlmUunW0WXIvh/pakvToWuPbdSlNrG8tdImlvs2EG29O23nrEWG/QvclW/x+Gp0TBsOB7KUmxruDPiZzhtg06WR18LcX/bOvd1Gg7zgzDffiucUSeLPs5wdjbMEwzyzBPz7VJJwLWciYEAPCGEAIAeEMIAQC8IYQAAN4QQgAAbwghAIA3hBAAwBtCCADgDSEEAPCGEAIAeNNrx/as10FlpCAj42o01e81jLTJ0C9MvY+qPnBtxDhz5nrDPI5hGm7qvU1vmupbLOOMpptaS6uCl4792b2m1pnR4KNbbpw+zNT7dePjcOiw4Ps/uazM1PtICh+HGYZ5NsMjEVNvDdsbuPSQcQxPm3G0ztDRwWvvnGB7vh0dHvyxkh21jicKfqAFw6KBa7vjZ3RI7weq5UwIAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOANIQQA8IYQAgB402tnx51Sgbo0IFBtVMFnGslUK/3GMFcrFjti6v3qtuBz6SI5wWsl6RslwedwRQxz5s6yzb66wXCf//nf2mZf1f7ltsC1o6PB55hJUkzB63Mitvsk5wHb47AkpyRwbZ5xvlvzyeCP28yMDFPvY7Hgj9uaLTtNvbst4/dsd4lkGwWorpPBa4cbnz850eCPlVeP2X4HZWYG7/36NsNAvfbgpZwJAQC8MYXQsmXLdPPNNys7O1vZ2dkaN26cXnrppcTtzjlVVVUpPz9fgwYNUmlpqQ4ePJj0RQMA0oMphIYOHapFixZp9+7d2r17tyZNmqT77rsvETSLFy/WkiVLtHTpUu3atUt5eXmaPHmy2traUrJ4AEDfZgqhKVOm6J577tHIkSM1cuRI/eM//qOuvfZa7dixQ845PfXUU1qwYIGmTp2q4uJirVixQidPntSqVYYPfQEA9BuX/Deh7u5urV69Wu3t7Ro3bpwaGhrU0tKisk98oFY4HNZdd92l7du3f2afjo4OxePxHhcAQP9gDqH6+npde+21CofDmjlzptauXasbb7xRLS0tkqTc3Nwe9bm5uYnbzqe6ulqRSCRxKSgosC4JANBHmUPohhtu0P79+7Vjxw49/PDDqqio0Ouvv564PRQK9ah3zp1z3SdVVlYqFoslLk1NTdYlAQD6KPP7hAYOHKgvfvGLkqSxY8dq165d+uEPf6i/+Zu/kSS1tLQo+onXtbe2tp5zdvRJ4XBY4XDYugwAQBq47PcJOefU0dGhoqIi5eXlqaamJnFbZ2en6urqNH78+Mv9MQCANGQ6E3r88cdVXl6ugoICtbW1afXq1aqtrdXLL7+sUCikuXPnauHChRoxYoRGjBihhQsXKjMzU9OnT0/V+gEAfZgphN5//31961vfUnNzsyKRiG6++Wa9/PLLmjx5siRp/vz5OnXqlB555BF98MEHKikp0aZNm5SVlWVeWFOsSwPcmUC1xxR8nES9oVaSdu/cG7i2KGqb9RHJCT5aJ9s4dqRLwXsfkW0k0FEZZpRIalbwcSxdxrVcHw1+xxzTm6bejYa1xI/Y7pOGncdM9U2R4DNq8nJsI4F+uyr4Y3zMA7YH4qmursC13X9vai1Zxvb8rbG3bTqR3t8UvHZDhaFYUmYk+HP53f9he/6YnsqW2uDbbguh55577oK3h0IhVVVVqaqqytIWANBPMTsOAOANIQQA8IYQAgB4QwgBALwhhAAA3hBCAABvCCEAgDeEEADAG0IIAOCNeYp2qjnnJEndbd2Bv6c7lVnaHrz0zIlgY4YS9adc4Npu46DxzvjpwLUfqdPW2zKTQ9JpBV/LaQXf97OC3+ddxt7dht5n2mx7b5x8pDMZwfufbjfehx3BS7vbbcfZ3RX8MZ7CrZc+MvY27o/lPjzTZmt95irDfWh7KtvqLU/7/6z9+Pf5hYRckKor6N133+WD7QAgDTQ1NWno0KEXrOl1IXTmzBm99957ysrK6vFhePF4XAUFBWpqalJ2drbHFaYWx5k++sMxShxnuknGcTrn1NbWpvz8fF111YX/p6rX/XfcVVdddcHkzM7OTusHwMc4zvTRH45R4jjTzeUeZyTg9G9emAAA8IYQAgB402dCKBwO64knnlA4bHyZWB/DcaaP/nCMEseZbq70cfa6FyYAAPqPPnMmBABIP4QQAMAbQggA4A0hBADwps+E0NNPP62ioiJdffXVuvXWW/XKK6/4XlJSVVVVKRQK9bjk5eX5XtZl2bp1q6ZMmaL8/HyFQiGtW7eux+3OOVVVVSk/P1+DBg1SaWmpDh486Gexl+Fixzljxoxz9vb222/3s9hLVF1drdtuu01ZWVkaMmSI7r//fh06dKhHTTrsZ5DjTIf9XLZsmW6++ebEG1LHjRunl156KXH7ldzLPhFCa9as0dy5c7VgwQLt27dPX/nKV1ReXq7GxkbfS0uqm266Sc3NzYlLfX297yVdlvb2do0ZM0ZLly497+2LFy/WkiVLtHTpUu3atUt5eXmaPHmy2tqMEx49u9hxStLdd9/dY283bNhwBVd4+erq6jRr1izt2LFDNTU1On36tMrKytTe/vsJv+mwn0GOU+r7+zl06FAtWrRIu3fv1u7duzVp0iTdd999iaC5onvp+oAvf/nLbubMmT2uGzVqlPve977naUXJ98QTT7gxY8b4XkbKSHJr165NfH3mzBmXl5fnFi1alLjuo48+cpFIxP3kJz/xsMLk+PRxOudcRUWFu++++7ysJ1VaW1udJFdXV+ecS9/9/PRxOpee++mcc3/4h3/onn322Su+l73+TKizs1N79uxRWVlZj+vLysq0fft2T6tKjcOHDys/P19FRUX65je/qaNHj/peUso0NDSopaWlx76Gw2HdddddabevklRbW6shQ4Zo5MiReuihh9Ta2up7SZclFotJkgYPHiwpfffz08f5sXTaz+7ubq1evVrt7e0aN27cFd/LXh9Cx44dU3d3t3Jzc3tcn5ubq5aWFk+rSr6SkhKtXLlSGzdu1DPPPKOWlhaNHz9ex48f9720lPh479J9XyWpvLxczz//vDZv3qwnn3xSu3bt0qRJk9TRYfgQml7EOad58+ZpwoQJKi4ulpSe+3m+45TSZz/r6+t17bXXKhwOa+bMmVq7dq1uvPHGK76XvW6K9mf55Mc6SGcfIJ++ri8rLy9P/Hv06NEaN26chg8frhUrVmjevHkeV5Za6b6vkjRt2rTEv4uLizV27FgVFhZq/fr1mjp1qseVXZrZs2frwIED2rZt2zm3pdN+ftZxpst+3nDDDdq/f78+/PBD/fKXv1RFRYXq6uoSt1+pvez1Z0I5OTkaMGDAOQnc2tp6TlKnk2uuuUajR4/W4cOHfS8lJT5+5V9/21dJikajKiws7JN7O2fOHL344ovasmVLj49cSbf9/KzjPJ++up8DBw7UF7/4RY0dO1bV1dUaM2aMfvjDH17xvez1ITRw4EDdeuutqqmp6XF9TU2Nxo8f72lVqdfR0aE33nhD0WjU91JSoqioSHl5eT32tbOzU3V1dWm9r5J0/PhxNTU19am9dc5p9uzZeuGFF7R582YVFRX1uD1d9vNix3k+fXE/z8c5p46Ojiu/l0l/qUMKrF692mVkZLjnnnvOvf76627u3LnummuucW+//bbvpSXNo48+6mpra93Ro0fdjh073J/8yZ+4rKysPn2MbW1tbt++fW7fvn1OkluyZInbt2+fe+edd5xzzi1atMhFIhH3wgsvuPr6evfggw+6aDTq4vG455XbXOg429ra3KOPPuq2b9/uGhoa3JYtW9y4cePcF77whT51nA8//LCLRCKutrbWNTc3Jy4nT55M1KTDfl7sONNlPysrK93WrVtdQ0ODO3DggHv88cfdVVdd5TZt2uScu7J72SdCyDnnfvzjH7vCwkI3cOBAd8stt/R4yWQ6mDZtmotGoy4jI8Pl5+e7qVOnuoMHD/pe1mXZsmWLk3TOpaKiwjl39mW9TzzxhMvLy3PhcNjdeeedrr6+3u+iL8GFjvPkyZOurKzMXXfddS4jI8MNGzbMVVRUuMbGRt/LNjnf8Ulyy5cvT9Skw35e7DjTZT//4i/+IvH79LrrrnNf/epXEwHk3JXdSz7KAQDgTa//mxAAIH0RQgAAbwghAIA3hBAAwBtCCADgDSEEAPCGEAIAeEMIAQC8IYQAAN4QQgAAbwghAIA3hBAAwJv/D9QeFu+aE5MBAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for experience in split_cifar10.train_stream:\n",
    "    # Show the first image for each experience\n",
    "    for sample in experience.dataset:\n",
    "        plt.imshow(sample[0].permute(1, 2, 0))\n",
    "        plt.show()\n",
    "        break"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torchvision.models import resnet18\n",
    "from torch.nn import CrossEntropyLoss\n",
    "from avalanche.training import Naive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 1 GPU(s) available.\n",
      "Device name: NVIDIA GeForce RTX 3080 Ti Laptop GPU\n"
     ]
    }
   ],
   "source": [
    "if torch.cuda.is_available():       \n",
    "        device = torch.device(\"cuda\")\n",
    "        print(f'There are {torch.cuda.device_count()} GPU(s) available.')\n",
    "        print('Device name:', torch.cuda.get_device_name(0))\n",
    "\n",
    "else:\n",
    "    print('No GPU available, using the CPU instead.')\n",
    "    device = torch.device(\"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from avalanche.training.plugins import EvaluationPlugin\n",
    "from avalanche.evaluation.metrics import accuracy_metrics, loss_metrics, forgetting_metrics, bwt_metrics\n",
    "from avalanche.logging import InteractiveLogger, TextLogger, TensorboardLogger"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_folder = './results/SCIFAR10'\n",
    "if not os.path.exists(save_folder):\n",
    "    os.makedirs(save_folder)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DER++"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_plugin = EvaluationPlugin(\n",
    "    accuracy_metrics(trained_experience=True, stream=True, epoch_running=True),\n",
    "    loss_metrics(stream=True, epoch_running=True, epoch=True),\n",
    "    forgetting_metrics(experience=True, stream=True),\n",
    "    bwt_metrics(experience=True, stream=True),\n",
    "    loggers=[InteractiveLogger(),\n",
    "                TextLogger(open(os.path.join(save_folder, 'exploratory_SCIFAR10_der.txt'), 'a')),\n",
    "                TensorboardLogger(os.path.join(save_folder, 'der_tb'))\n",
    "              ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from der import DerPlugin"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Choose best hyperparams from DER paper. We use a buffer size of 500."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = 0.03\n",
    "buffer_size = 500\n",
    "alpha = 0.2\n",
    "beta = 0.5\n",
    "epochs = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = resnet18(weights=None, num_classes=10)\n",
    "optimizer = torch.optim.SGD(model.parameters(), lr=lr)\n",
    "criterion = CrossEntropyLoss()\n",
    "\n",
    "der_plugin = DerPlugin(mem_size=buffer_size, alpha=alpha, beta=beta)\n",
    "\n",
    "cl_strategy = Naive(model, optimizer, criterion, train_mb_size=64, train_epochs=epochs, eval_mb_size=64,\n",
    "                    eval_every=0, device=device, evaluator=eval_plugin, plugins=[der_plugin])"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training loop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_results = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting experiment...\n",
      "Start of experience:  0\n",
      "Current Classes:  [2, 6]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.29it/s]\n",
      "> Eval on experience 0 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 2.3523\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.65it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 1.5909\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 1.5909\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7150\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.46it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.7646\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.7646\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8025\n",
      "100%|██████████| 157/157 [00:03<00:00, 47.50it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.6239\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.6239\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8341\n",
      "100%|██████████| 157/157 [00:03<00:00, 47.68it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5583\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5583\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8503\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.90it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5128\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5128\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8625\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.17it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4682\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4682\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8749\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.53it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4515\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4515\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8806\n",
      "100%|██████████| 157/157 [00:03<00:00, 45.77it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4264\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4264\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8831\n",
      "100%|██████████| 157/157 [00:03<00:00, 45.34it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4126\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4126\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8870\n",
      "100%|██████████| 157/157 [00:03<00:00, 45.23it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3909\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3909\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8959\n",
      "100%|██████████| 157/157 [00:03<00:00, 45.55it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3897\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3897\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8934\n",
      "100%|██████████| 157/157 [00:03<00:00, 44.31it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3681\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3681\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9014\n",
      "100%|██████████| 157/157 [00:03<00:00, 45.16it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3537\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3537\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9056\n",
      "100%|██████████| 157/157 [00:03<00:00, 44.87it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3382\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3382\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9096\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.75it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3303\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3303\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9111\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.42it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3164\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3164\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9134\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.37it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3082\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3082\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9172\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.76it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2954\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2954\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9208\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.79it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2952\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2952\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9207\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.56it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2839\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2839\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9253\n",
      "100%|██████████| 157/157 [00:03<00:00, 44.02it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2723\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2723\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9311\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.68it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2717\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2717\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9307\n",
      "100%|██████████| 157/157 [00:03<00:00, 43.03it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2665\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2665\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9308\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.86it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2601\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2601\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9343\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.68it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2574\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2574\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9339\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.91it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2471\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2471\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9362\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.27it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2390\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2390\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9403\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.54it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2322\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2322\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9413\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.62it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2298\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2298\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9398\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.73it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2337\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2337\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9397\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.88it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2141\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2141\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9500\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.24it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2070\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2070\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9497\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.33it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2178\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2178\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9469\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.30it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1997\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1997\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9526\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.82it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1994\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1994\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9538\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.03it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2001\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2001\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9514\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.30it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1907\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1907\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9534\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.44it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1829\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1829\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9584\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.18it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1832\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1832\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9584\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.09it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1889\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1889\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9562\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.51it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1893\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1893\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9563\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.92it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1820\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1820\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9563\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.85it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1672\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1672\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9627\n",
      "100%|██████████| 157/157 [00:04<00:00, 35.16it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1680\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1680\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9611\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.76it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1606\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1606\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9660\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.88it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1562\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1562\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9657\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.08it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1613\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1613\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9632\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.15it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1647\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1647\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9614\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.29it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1534\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1534\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9663\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.61it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1484\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1484\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9675\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 47.73it/s]\n",
      "> Eval on experience 0 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9635\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.1087\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9635\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 51.20it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 48.33it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 51.56it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 48.53it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 53.07it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.8970\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 6.6483\n",
      "\tStreamBWT/eval_phase/test_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1794\n",
      "Start of experience:  1\n",
      "Current Classes:  [8, 1]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 1 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.99it/s]\n",
      "> Eval on experience 1 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 8.1095\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.33it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 1.2180\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 1.2180\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7491\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.35it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5984\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5984\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8789\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.12it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3368\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3368\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9119\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.73it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2938\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2938\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9233\n",
      "100%|██████████| 157/157 [00:04<00:00, 36.25it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3580\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3580\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9220\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.86it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2489\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2489\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9390\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.58it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2419\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2419\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9412\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.18it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2293\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2293\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9472\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.76it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2231\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2231\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9495\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.56it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2107\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2107\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9530\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.02it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1986\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1986\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9569\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.16it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1841\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1841\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9586\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.69it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1848\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1848\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9572\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.81it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1818\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1818\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9623\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.39it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1806\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1806\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9601\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.90it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2842\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2842\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9555\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.03it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1688\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1688\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9684\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.94it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1798\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1798\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9642\n",
      "100%|██████████| 157/157 [00:05<00:00, 30.32it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1558\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1558\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9681\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.24it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1552\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1552\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9679\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.97it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1589\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1589\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9723\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.45it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1402\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1402\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9745\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.88it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1394\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1394\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9743\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.27it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1316\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1316\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9748\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.57it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2095\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2095\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9671\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.05it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1430\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1430\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9760\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.05it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1343\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1343\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9775\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.19it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1271\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1271\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9780\n",
      "100%|██████████| 157/157 [00:04<00:00, 37.76it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1321\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1321\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9785\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.46it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1299\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1299\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9775\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.38it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1358\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1358\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9778\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.99it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1184\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1184\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9802\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.97it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1126\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1126\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9815\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.99it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1088\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1088\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9837\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.88it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1063\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1063\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9826\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.89it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1040\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1040\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9845\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.36it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1072\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1072\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9841\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.29it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1038\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1038\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9840\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.52it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1041\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1041\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9843\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.74it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0999\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0999\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9861\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.80it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1020\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1020\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9844\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.02it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0903\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0903\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9895\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.70it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0957\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0957\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9861\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.29it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0885\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0885\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9894\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.38it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1050\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1050\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9868\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.97it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1031\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1031\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9846\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.99it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0939\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0939\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9872\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.50it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0931\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0931\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9868\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.23it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0949\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0949\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9871\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.81it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0950\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0950\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9883\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 1 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.19it/s]\n",
      "> Eval on experience 1 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.8253\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.5039\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.8253\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 47.59it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.2465\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.2465\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 48.23it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 49.69it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 45.51it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 47.06it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.7260\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 4.8825\n",
      "\tStreamBWT/eval_phase/test_stream = -0.2465\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.2465\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.2904\n",
      "Start of experience:  2\n",
      "Current Classes:  [4, 5]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 2 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 47.56it/s]\n",
      "> Eval on experience 2 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 6.3891\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.00it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 1.5268\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 1.5268\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7098\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.57it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5411\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5411\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8413\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.26it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4725\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4725\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8493\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.44it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4280\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4280\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8755\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.43it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4098\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4098\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8815\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.81it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3796\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3796\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8864\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.49it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3566\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3566\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8969\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.28it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3510\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3510\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8971\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.71it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3287\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3287\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9086\n",
      "100%|██████████| 157/157 [00:04<00:00, 36.20it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3072\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3072\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9126\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.20it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2985\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2985\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9139\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.47it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2893\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2893\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9194\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.23it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2887\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2887\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9198\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.10it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2737\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2737\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9259\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.61it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2865\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2865\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9257\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.17it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2561\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2561\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9303\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.18it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2507\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2507\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9326\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.93it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2414\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2414\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9357\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.78it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2297\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2297\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9405\n",
      "100%|██████████| 157/157 [00:04<00:00, 36.86it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2307\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2307\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9395\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.20it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2329\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2329\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9392\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.94it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2219\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2219\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9454\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.79it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2302\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2302\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9438\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.90it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2155\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2155\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9466\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.82it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2103\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2103\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9472\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.59it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2137\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2137\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9468\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.07it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2060\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2060\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9530\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.52it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1947\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1947\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9554\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.78it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1854\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1854\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9554\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.89it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1927\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1927\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9552\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.82it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1781\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1781\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9581\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.52it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1829\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1829\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9608\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.58it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1726\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1726\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9618\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.77it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1654\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1654\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9643\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.03it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1663\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1663\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9623\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.38it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1603\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1603\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9650\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.90it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1497\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1497\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9712\n",
      "100%|██████████| 157/157 [00:04<00:00, 35.86it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1730\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1730\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9616\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.79it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1658\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1658\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9649\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.43it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1642\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1642\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9679\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.52it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1616\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1616\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9649\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.06it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1532\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1532\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9707\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.87it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1499\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1499\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9704\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.63it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1504\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1504\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9717\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.38it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1354\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1354\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9733\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.46it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1397\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1397\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9716\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.73it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1588\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1588\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9692\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.02it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1389\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1389\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9735\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.34it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1305\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1305\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9759\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.58it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1248\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1248\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9777\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 2 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.00it/s]\n",
      "> Eval on experience 2 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.8689\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.3243\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.8689\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 48.41it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.4395\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.4395\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 44.53it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = 0.0690\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = -0.0690\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 44.16it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 48.46it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 50.04it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.7073\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 3.9423\n",
      "\tStreamBWT/eval_phase/test_stream = -0.1852\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.1852\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.4244\n",
      "Start of experience:  3\n",
      "Current Classes:  [0, 9]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 3 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.14it/s]\n",
      "> Eval on experience 3 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 7.4032\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.40it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 1.1422\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 1.1422\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8310\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.73it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5909\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5909\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8934\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.28it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3433\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3433\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9174\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.76it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3101\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3101\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9306\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.73it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2523\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2523\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9404\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.42it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2502\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2502\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9410\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.64it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2393\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2393\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9438\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.23it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2397\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2397\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9462\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.95it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2099\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2099\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9520\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.84it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1856\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1856\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9587\n",
      "100%|██████████| 157/157 [00:04<00:00, 36.53it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1808\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1808\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9622\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.45it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1804\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1804\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9588\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.63it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1734\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1734\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9634\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.23it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1691\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1691\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9635\n",
      "100%|██████████| 157/157 [00:05<00:00, 30.58it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1752\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1752\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9638\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.78it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1529\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1529\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9716\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.49it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1549\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1549\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9698\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.85it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1461\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1461\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9715\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.49it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1461\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1461\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9713\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.93it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1389\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1389\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9722\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.70it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1402\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1402\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9730\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.14it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1359\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1359\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9746\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.59it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1453\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1453\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9770\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.11it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1406\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1406\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9762\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.82it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1178\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1178\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9810\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.00it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1243\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1243\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9788\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.81it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1269\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1269\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9772\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.70it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1161\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1161\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9832\n",
      "100%|██████████| 157/157 [00:04<00:00, 35.69it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1146\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1146\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9814\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.02it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1085\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1085\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9848\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.97it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1110\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1110\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9830\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.23it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1192\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1192\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9819\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.50it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1081\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1081\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9854\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.13it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1219\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1219\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9813\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.81it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1193\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1193\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9853\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.40it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1042\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1042\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9879\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.91it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1098\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1098\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9847\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.10it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1110\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1110\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9858\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.91it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1013\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1013\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9862\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.48it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0928\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0928\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9882\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.77it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0913\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0913\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9897\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.38it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0929\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0929\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9891\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.85it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0978\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0978\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9888\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.76it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0906\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0906\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9908\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.86it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1161\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1161\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9850\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.44it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1994\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1994\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9804\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.43it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0968\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0968\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9887\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.27it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0917\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0917\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9882\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.67it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1031\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1031\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9872\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.98it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0844\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0844\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9922\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 3 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.72it/s]\n",
      "> Eval on experience 3 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.8133\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.5738\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.8133\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:01<00:00, 25.70it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.3490\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.3490\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 48.16it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.1470\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.1470\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 50.02it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp002 = -0.2450\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp002 = 0.2450\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 47.07it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 47.15it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.6316\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 2.6753\n",
      "\tStreamBWT/eval_phase/test_stream = -0.2470\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.2470\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.5053\n",
      "Start of experience:  4\n",
      "Current Classes:  [3, 7]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 4 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.80it/s]\n",
      "> Eval on experience 4 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 6.5151\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.94it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 1.1611\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 1.1611\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7951\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.71it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5034\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5034\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8711\n",
      "100%|██████████| 157/157 [00:03<00:00, 40.04it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3964\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3964\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8874\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.32it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3716\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3716\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8935\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.05it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3326\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3326\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9029\n",
      "100%|██████████| 157/157 [00:04<00:00, 37.96it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3012\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3012\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9136\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.78it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2874\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2874\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9207\n",
      "100%|██████████| 157/157 [00:04<00:00, 35.45it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2840\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2840\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9219\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.21it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2774\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2774\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9222\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.64it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2588\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2588\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9330\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.48it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2488\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2488\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9345\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.61it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2271\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2271\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9413\n",
      "100%|██████████| 157/157 [00:05<00:00, 30.60it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2276\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2276\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9402\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.98it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2232\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2232\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9459\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.06it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2170\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2170\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9469\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.73it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2077\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2077\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9485\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.08it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1940\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1940\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9541\n",
      "100%|██████████| 157/157 [00:05<00:00, 30.27it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1970\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1970\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9540\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.45it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2048\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2048\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9559\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.47it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1921\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1921\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9558\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.58it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1793\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1793\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9590\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.51it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1752\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1752\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9650\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.70it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1665\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1665\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9641\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.19it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1629\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1629\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9653\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.30it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1614\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1614\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9687\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.84it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1693\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1693\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9658\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.82it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1544\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1544\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9697\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.14it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1609\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1609\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9702\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.30it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1607\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1607\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9698\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.79it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1470\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1470\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9708\n",
      "100%|██████████| 157/157 [00:04<00:00, 36.14it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1485\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1485\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9729\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.52it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1968\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1968\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9636\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.17it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1401\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1401\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9742\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.22it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1478\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1478\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9703\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.55it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1426\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1426\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9749\n",
      "100%|██████████| 157/157 [00:05<00:00, 30.47it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1294\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1294\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9759\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.10it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1247\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1247\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9757\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.36it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1188\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1188\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9785\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.85it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1216\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1216\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9803\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.82it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1073\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1073\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9829\n",
      "100%|██████████| 157/157 [00:04<00:00, 37.92it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1074\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1074\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9824\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.17it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1043\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1043\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9840\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.62it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1058\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1058\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9833\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.84it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1183\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1183\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9806\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.28it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1130\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1130\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9820\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.54it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1130\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1130\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9826\n",
      "100%|██████████| 157/157 [00:04<00:00, 38.81it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1189\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1189\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9818\n",
      "100%|██████████| 157/157 [00:03<00:00, 39.38it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1121\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1121\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9836\n",
      "100%|██████████| 157/157 [00:04<00:00, 39.03it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1101\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1101\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9831\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.18it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1036\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1036\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9855\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 4 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.28it/s]\n",
      "> Eval on experience 4 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.8398\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.4417\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.8398\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 45.24it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.5265\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.5265\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 43.40it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.0885\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.0885\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 50.15it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp002 = -0.4720\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp002 = 0.4720\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 42.80it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp003 = -0.2390\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp003 = 0.2390\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 45.04it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.5425\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 2.3678\n",
      "\tStreamBWT/eval_phase/test_stream = -0.3315\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.3315\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.5425\n"
     ]
    }
   ],
   "source": [
    "# TRAINING LOOP WITH EVALUATION AT EACH EXPERIENCE\n",
    "\n",
    "print('Starting experiment...')\n",
    "results = []\n",
    "for experience in split_cifar10.train_stream:\n",
    "    print(\"Start of experience: \", experience.current_experience)\n",
    "    print(\"Current Classes: \", experience.classes_in_this_experience)\n",
    "\n",
    "    cl_strategy.train(experience)\n",
    "    print('Training completed')\n",
    "\n",
    "    exp_results.append(cl_strategy.eval(split_cifar10.test_stream))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get the accuracy on current and previous experiences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "aggregated_results = {}\n",
    "\n",
    "for i, exp_dict in enumerate(exp_results):\n",
    "   for metric_name, metric_value in exp_dict.items():\n",
    "        if metric_name not in aggregated_results:\n",
    "            aggregated_results[metric_name] = []\n",
    "            aggregated_results[metric_name].append(metric_value)\n",
    "        else:\n",
    "            aggregated_results[metric_name].append(metric_value)\n",
    "\n",
    "acc_stream_der = aggregated_results['Top1_Acc_Stream/eval_phase/test_stream/Task000']\n",
    "acc_trained_exps_der = aggregated_results['Accuracy_On_Trained_Experiences/eval_phase/test_stream/Task000']\n",
    "loss_stream_der = aggregated_results['Loss_Stream/eval_phase/test_stream/Task000']"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot the training loss curve for each experience"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(os.path.join(save_folder, 'exploratory_SCIFAR10_der.txt'), 'r') as f:\n",
    "# In the log file find all the lines starting with \"Loss_Epoch\"\n",
    "    lines = [line for line in f.readlines() if line.startswith('\\tLoss_Epoch')]\n",
    "    # Get the final loss at each line after the = sign\n",
    "    epoch_losses_list = [float(line.split('=')[1]) for line in lines]\n",
    "    \n",
    "# Divide by experiences\n",
    "exp_epoch_losses = [epoch_losses_list[i:i+epochs] for i in range(0, len(epoch_losses_list), epochs)]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the loss for each experience\n",
    "for i, exp_epoch_loss in enumerate(exp_epoch_losses):\n",
    "    plt.plot(exp_epoch_loss, label=f'Experience {i}')\n",
    "plt.xticks(range(0, epochs+1, 10))\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot the full loss curve across the experiences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "full_loss_stream = []\n",
    "for i, exp_epoch_loss in enumerate(exp_epoch_losses):\n",
    "    full_loss_stream += exp_epoch_loss\n",
    "\n",
    "plt.plot(full_loss_stream)\n",
    "plt.xticks(range(0, len(full_loss_stream), epochs))\n",
    "plt.gcf().set_size_inches(20, 5)\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check labels occurrences in the buffer are balanced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label: 0 - Occurrences: 51\n",
      "Label: 1 - Occurrences: 54\n",
      "Label: 2 - Occurrences: 54\n",
      "Label: 3 - Occurrences: 50\n",
      "Label: 4 - Occurrences: 61\n",
      "Label: 5 - Occurrences: 55\n",
      "Label: 6 - Occurrences: 56\n",
      "Label: 7 - Occurrences: 37\n",
      "Label: 8 - Occurrences: 52\n",
      "Label: 9 - Occurrences: 30\n"
     ]
    }
   ],
   "source": [
    "labels, occurrences = der_plugin.get_buffer_labels()\n",
    "for label, occ in zip(labels, occurrences):\n",
    "    print(f'Label: {label} - Occurrences: {occ}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(labels, occurrences)\n",
    "plt.xlabel('Classes')\n",
    "plt.xticks([])\n",
    "plt.show()\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Replay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_plugin = EvaluationPlugin(\n",
    "    accuracy_metrics(trained_experience=True, stream=True, epoch_running=True),\n",
    "    loss_metrics(stream=True, epoch_running=True, epoch=True),\n",
    "    forgetting_metrics(experience=True, stream=True),\n",
    "    bwt_metrics(experience=True, stream=True),\n",
    "    loggers=[InteractiveLogger(),\n",
    "                TensorboardLogger(os.path.join(save_folder, 'replay_tb'))\n",
    "              ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from avalanche.training import Replay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = 0.1\n",
    "buffer_size = 500\n",
    "epochs = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = resnet18(weights=None, num_classes=10)\n",
    "optimizer = torch.optim.SGD(model.parameters(), lr=lr)\n",
    "criterion = CrossEntropyLoss()\n",
    "\n",
    "cl_strategy = Replay(model, optimizer, criterion, mem_size=buffer_size, train_mb_size=64, train_epochs=epochs, eval_mb_size=64,\n",
    "                    eval_every=0, device=device, evaluator=eval_plugin)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_results = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting experiment...\n",
      "Start of experience:  0\n",
      "Current Classes:  [2, 6]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from train stream --\n",
      "0it [00:00, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100%|██████████| 157/157 [00:02<00:00, 53.78it/s]\n",
      "> Eval on experience 0 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 2.4063\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 45.99it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 1.1118\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 1.1118\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.6475\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.15it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5323\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5323\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7764\n",
      "100%|██████████| 157/157 [00:03<00:00, 42.06it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4878\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4878\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7960\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.28it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4252\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4252\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8247\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.02it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3986\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3986\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8374\n",
      "100%|██████████| 157/157 [00:03<00:00, 47.71it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3770\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3770\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8466\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.19it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3391\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3391\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8565\n",
      "100%|██████████| 157/157 [00:03<00:00, 46.17it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3909\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3909\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8397\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.04it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3210\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3210\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8679\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.10it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2969\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2969\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8784\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.51it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2921\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2921\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8800\n",
      "100%|██████████| 157/157 [00:03<00:00, 48.64it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2660\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2660\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8900\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.11it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2627\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2627\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8924\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.32it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2665\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2665\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8925\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.48it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2431\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2431\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9021\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.47it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2409\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2409\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8989\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.37it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2402\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2402\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9019\n",
      "100%|██████████| 157/157 [00:04<00:00, 37.13it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2259\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2259\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9062\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.80it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2132\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2132\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9150\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.95it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2153\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2153\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9116\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.84it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2094\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2094\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9144\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.54it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2030\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2030\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9161\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.96it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1947\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1947\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9202\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.38it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1931\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1931\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9218\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.01it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1806\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1806\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9258\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.90it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1735\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1735\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9284\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.93it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1758\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1758\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9306\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.15it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1816\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1816\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9289\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.02it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1763\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1763\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9297\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.31it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1643\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1643\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9351\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.93it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1640\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1640\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9354\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.13it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1551\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1551\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9378\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.03it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1450\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1450\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9407\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.98it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1425\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1425\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9431\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.25it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1396\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1396\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9445\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.99it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1374\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1374\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9456\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.13it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1344\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1344\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9470\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.16it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1357\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1357\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9474\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.62it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1260\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1260\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9513\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.83it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1271\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1271\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9514\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.54it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1246\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1246\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9524\n",
      "100%|██████████| 157/157 [00:03<00:00, 49.89it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1242\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1242\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9490\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.75it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1247\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1247\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9524\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.11it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1111\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1111\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9556\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.52it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1128\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1128\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9554\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.50it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1125\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1125\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9546\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.78it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1071\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1071\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9603\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.57it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1025\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1025\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9580\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.76it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1010\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1010\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9602\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.63it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1004\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1004\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9626\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 60.47it/s]\n",
      "> Eval on experience 0 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9460\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.1296\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9460\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 59.35it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.82it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.34it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.95it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.92it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.9155\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 13.1494\n",
      "\tStreamBWT/eval_phase/test_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1831\n",
      "Start of experience:  1\n",
      "Current Classes:  [8, 1]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 1 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 57.30it/s]\n",
      "> Eval on experience 1 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 14.5243\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.50it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.7112\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.7112\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7729\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.83it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2706\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2706\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9015\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.20it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1873\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1873\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9319\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.97it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1636\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1636\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9420\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.38it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1353\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1353\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9520\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.20it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1200\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1200\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9572\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.29it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1123\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1123\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9625\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.37it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1095\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1095\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9629\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.32it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0924\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0924\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9677\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.27it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0849\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0849\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9700\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.42it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0803\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0803\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9696\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.32it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0778\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0778\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9718\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.59it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0744\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0744\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9736\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.33it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0705\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0705\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9755\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.75it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0644\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0644\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9764\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.15it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0602\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0602\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9781\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.35it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0632\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0632\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9784\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.04it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0608\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0608\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9776\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.89it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0533\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0533\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9808\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.32it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0548\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0548\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9801\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.81it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0514\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0514\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9807\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.76it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0476\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0476\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9829\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.69it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0468\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0468\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9830\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.80it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0467\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0467\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9836\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.63it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0421\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0421\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9859\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.11it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0400\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0400\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9853\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.29it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0433\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0433\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9853\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.95it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0402\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0402\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9864\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.56it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0394\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0394\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9858\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.02it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0353\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0353\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9876\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.93it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0376\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0376\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9862\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.02it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0356\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0356\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9873\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.04it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0362\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0362\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9868\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.32it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0327\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0327\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9883\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.87it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0338\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0338\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9871\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.06it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0302\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0302\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9892\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.14it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0285\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0285\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9904\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.00it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0338\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0338\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9881\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.81it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0276\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0276\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9893\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.93it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0290\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0290\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9897\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.87it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0259\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0259\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9903\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.72it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0269\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0269\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9906\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.64it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0325\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0325\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9884\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.93it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0321\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0321\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9891\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.74it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0235\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0235\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9919\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.82it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0223\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0223\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9914\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.33it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0207\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0207\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9926\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.98it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0232\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0232\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9908\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.64it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0224\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0224\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9919\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.54it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0203\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0203\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9932\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 1 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 57.49it/s]\n",
      "> Eval on experience 1 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9925\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0241\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9925\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.45it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.2185\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.2185\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 61.45it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.02it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 56.80it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.98it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.8320\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 10.7006\n",
      "\tStreamBWT/eval_phase/test_stream = -0.2185\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.2185\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.3328\n",
      "Start of experience:  2\n",
      "Current Classes:  [4, 5]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 2 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 55.51it/s]\n",
      "> Eval on experience 2 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 19.1055\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.89it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.7300\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.7300\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7653\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.11it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3307\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3307\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8725\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.19it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2664\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2664\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8968\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.41it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2237\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2237\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9145\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.40it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2081\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2081\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9213\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.40it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1819\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1819\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9295\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.12it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1723\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1723\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9341\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.45it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1574\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1574\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9377\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.19it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1513\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1513\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9413\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.16it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1384\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1384\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9468\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.06it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1363\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1363\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9482\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.15it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1264\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1264\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9520\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.35it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1250\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1250\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9526\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.14it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1194\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1194\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9542\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.49it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1177\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1177\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9565\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.25it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1099\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1099\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9588\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.35it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1057\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1057\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9599\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.18it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1061\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1061\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9581\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.32it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1022\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1022\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9613\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.33it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0970\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0970\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9626\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.40it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0863\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0863\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9664\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.30it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0912\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0912\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9667\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.70it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0912\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0912\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9661\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.83it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0871\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0871\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9677\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.07it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0814\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0814\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9703\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.73it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0853\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0853\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9674\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.87it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0807\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0807\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9697\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.73it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0761\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0761\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9717\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.28it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0721\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0721\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9727\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.18it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0702\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0702\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9743\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.15it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0686\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0686\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9753\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.01it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0722\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0722\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9739\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.07it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0624\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0624\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9756\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.24it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0699\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0699\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9738\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.88it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0593\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0593\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9777\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.97it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0621\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0621\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9776\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.14it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0600\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0600\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9781\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.20it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0627\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0627\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9759\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.78it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0575\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0575\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9785\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.16it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0525\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0525\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9797\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.14it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0526\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0526\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9807\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.74it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0552\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0552\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9795\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.09it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0529\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0529\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9800\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.66it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0471\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0471\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9826\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.50it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0480\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0480\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9823\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.97it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0449\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0449\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9828\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.94it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0448\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0448\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9846\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.81it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0429\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0429\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9852\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.17it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0393\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0393\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9861\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.79it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0428\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0428\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9838\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 2 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 58.14it/s]\n",
      "> Eval on experience 2 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9713\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0792\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9713\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 52.48it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.5000\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.5000\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 58.41it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.2595\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.2595\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 63.49it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.52it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 61.12it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.6752\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 7.8900\n",
      "\tStreamBWT/eval_phase/test_stream = -0.3797\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.3797\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.4051\n",
      "Start of experience:  3\n",
      "Current Classes:  [0, 9]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 3 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 57.80it/s]\n",
      "> Eval on experience 3 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 17.6396\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.19it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5025\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5025\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8441\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.63it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2152\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2152\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9221\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.52it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1699\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1699\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9387\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.90it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1495\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1495\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9467\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.44it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1256\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1256\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9547\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.26it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1172\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1172\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9576\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.30it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1049\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1049\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9617\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.25it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1028\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1028\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9622\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.84it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0906\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0906\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9661\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.79it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0875\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0875\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9672\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.68it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0817\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0817\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9693\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.23it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0759\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0759\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9721\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.44it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0748\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0748\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9724\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.77it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0692\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0692\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9754\n",
      "100%|██████████| 157/157 [00:04<00:00, 31.82it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0657\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0657\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9763\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.83it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0616\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0616\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9771\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.91it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0575\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0575\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9783\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.66it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0537\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0537\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9806\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.99it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0512\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0512\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9811\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.77it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0503\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0503\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9820\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.01it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0504\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0504\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9822\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.88it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0489\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0489\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9831\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.90it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0462\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0462\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9834\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.85it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0404\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0404\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9852\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.48it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0478\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0478\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9831\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.11it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0424\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0424\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9854\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.63it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0355\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0355\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9870\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.77it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0409\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0409\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9857\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.58it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0360\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0360\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9865\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.88it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0386\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0386\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9866\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.69it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0354\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0354\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9876\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.90it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0365\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0365\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9871\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.36it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0355\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0355\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9875\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.43it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0263\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0263\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9903\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.55it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0326\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0326\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9874\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.96it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0291\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0291\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9901\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.60it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0295\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0295\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9895\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.25it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0330\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0330\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9878\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.57it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0230\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0230\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9916\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.37it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0277\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0277\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9901\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.26it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0303\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0303\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9896\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.33it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0245\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0245\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9913\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.02it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0215\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0215\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9917\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.77it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0235\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0235\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9919\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.74it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0178\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0178\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9936\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.44it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0228\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0228\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9919\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.34it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0248\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0248\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9919\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.86it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0218\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0218\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9925\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.32it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0196\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0196\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9933\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.67it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0175\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0175\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9936\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 3 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 58.48it/s]\n",
      "> Eval on experience 3 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9664\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.1665\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9664\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 59.70it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.5870\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.5870\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 63.12it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.5410\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.5410\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 61.05it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp002 = -0.4045\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp002 = 0.4045\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 61.19it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 56.43it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.5403\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 6.4321\n",
      "\tStreamBWT/eval_phase/test_stream = -0.5108\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.5108\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.4322\n",
      "Start of experience:  4\n",
      "Current Classes:  [3, 7]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 4 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 57.91it/s]\n",
      "> Eval on experience 4 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 17.0554\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:05<00:00, 31.37it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4935\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4935\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8438\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.55it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2267\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2267\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9135\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.90it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1900\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1900\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9262\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.10it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1737\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1737\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9337\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.17it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1487\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1487\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9432\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.26it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1368\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1368\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9449\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.32it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1244\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1244\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9527\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.41it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1196\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1196\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9529\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.07it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1132\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1132\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9557\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.53it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1067\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1067\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9588\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.68it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1088\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1088\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9578\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.59it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0930\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0930\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9651\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.31it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0895\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0895\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9648\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.41it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0846\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0846\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9669\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.55it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0831\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0831\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9679\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.63it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0797\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0797\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9698\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.79it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0763\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0763\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9714\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.99it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0748\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0748\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9717\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.86it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0705\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0705\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9723\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.68it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0650\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0650\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9749\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.80it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0671\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0671\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9758\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.03it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0654\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0654\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9756\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.91it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0647\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0647\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9749\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.71it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0604\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0604\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9782\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.08it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0635\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0635\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9768\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.48it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0575\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0575\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9790\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.86it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0556\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0556\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9794\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.77it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0493\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0493\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9815\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.25it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0454\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0454\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9838\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.83it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0475\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0475\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9820\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.85it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0436\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0436\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9835\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.00it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0424\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0424\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9830\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.19it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0427\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0427\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9847\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.22it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0420\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0420\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9838\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.26it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0388\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0388\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9856\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.83it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0411\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0411\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9847\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.17it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0417\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0417\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9848\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.94it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0439\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0439\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9846\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.09it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0326\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0326\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9874\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.28it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0420\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0420\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9845\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.30it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0329\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0329\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9882\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.85it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0304\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0304\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9895\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.12it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0305\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0305\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9891\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.07it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0302\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0302\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9897\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.21it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0319\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0319\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9890\n",
      "100%|██████████| 157/157 [00:04<00:00, 34.09it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0328\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0328\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9871\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.70it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0340\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0340\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9876\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.59it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0280\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0280\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9896\n",
      "100%|██████████| 157/157 [00:04<00:00, 33.68it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0296\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0296\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9896\n",
      "100%|██████████| 157/157 [00:04<00:00, 32.83it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0324\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0324\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9891\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 4 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 60.73it/s]\n",
      "> Eval on experience 4 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9933\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0242\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9933\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 58.29it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.6340\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.6340\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 64.63it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.4945\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.4945\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.99it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp002 = -0.7755\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp002 = 0.7755\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 59.86it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp003 = -0.3970\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp003 = 0.3970\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 64.12it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.4590\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 4.9973\n",
      "\tStreamBWT/eval_phase/test_stream = -0.5753\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.5753\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.4590\n"
     ]
    }
   ],
   "source": [
    "# TRAINING LOOP WITH EVALUATION AT EACH EXPERIENCE\n",
    "\n",
    "print('Starting experiment...')\n",
    "results = []\n",
    "for experience in split_cifar10_transform.train_stream:\n",
    "    print(\"Start of experience: \", experience.current_experience)\n",
    "    print(\"Current Classes: \", experience.classes_in_this_experience)\n",
    "\n",
    "    cl_strategy.train(experience)\n",
    "    print('Training completed')\n",
    "\n",
    "    exp_results.append(cl_strategy.eval(split_cifar10_transform.test_stream))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "aggregated_results = {}\n",
    "\n",
    "for i, exp_dict in enumerate(exp_results):\n",
    "   for metric_name, metric_value in exp_dict.items():\n",
    "        if metric_name not in aggregated_results:\n",
    "            aggregated_results[metric_name] = []\n",
    "            aggregated_results[metric_name].append(metric_value)\n",
    "        else:\n",
    "            aggregated_results[metric_name].append(metric_value)\n",
    "\n",
    "acc_stream_er = aggregated_results['Top1_Acc_Stream/eval_phase/test_stream/Task000']\n",
    "acc_trained_exps_er = aggregated_results['Accuracy_On_Trained_Experiences/eval_phase/test_stream/Task000']\n",
    "loss_stream_er = aggregated_results['Loss_Stream/eval_phase/test_stream/Task000']"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Naive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_plugin = EvaluationPlugin(\n",
    "    accuracy_metrics(trained_experience=True, stream=True, epoch_running=True),\n",
    "    loss_metrics(stream=True, epoch_running=True, epoch=True),\n",
    "    forgetting_metrics(experience=True, stream=True),\n",
    "    bwt_metrics(experience=True, stream=True),\n",
    "    loggers=[InteractiveLogger(),\n",
    "                TensorboardLogger(os.path.join(save_folder, 'naive_tb'))\n",
    "              ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = 0.1\n",
    "epochs = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = resnet18(weights=None, num_classes=10)\n",
    "optimizer = torch.optim.SGD(model.parameters(), lr=lr)\n",
    "criterion = CrossEntropyLoss()\n",
    "\n",
    "cl_strategy = Naive(model, optimizer, criterion, train_mb_size=64, train_epochs=epochs, eval_mb_size=64,\n",
    "                    eval_every=0, device=device, evaluator=eval_plugin)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_results = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting experiment...\n",
      "Start of experience:  0\n",
      "Current Classes:  [2, 6]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from train stream --\n",
      "  1%|▏         | 2/157 [00:00<00:02, 62.82it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100%|██████████| 157/157 [00:02<00:00, 60.50it/s]\n",
      "> Eval on experience 0 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0249\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 2.2649\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0249\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.24it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.9254\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.9254\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.6815\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.76it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.5088\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.5088\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7802\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.91it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.6159\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.6159\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7700\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.41it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4578\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4578\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8026\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.90it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3863\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3863\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8389\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.90it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3543\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3543\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8465\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.52it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3438\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3438\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8556\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.54it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3168\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3168\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8720\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.60it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2964\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2964\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8779\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.52it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2890\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2890\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8786\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.55it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2640\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2640\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8920\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.98it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2650\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2650\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8886\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.16it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2547\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2547\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8912\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.34it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2444\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2444\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9006\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.83it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2378\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2378\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9049\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.53it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2296\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2296\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9086\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.15it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2228\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2228\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9122\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.22it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2114\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2114\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9154\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.53it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2219\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2219\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9085\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.16it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1963\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1963\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9213\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.86it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1917\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1917\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9220\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.52it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1937\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1937\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9209\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.55it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1887\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1887\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9235\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.26it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1742\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1742\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9295\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.44it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1681\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1681\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9345\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.41it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1689\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1689\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9336\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.41it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1689\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1689\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9313\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.67it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1562\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1562\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9387\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.95it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1555\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1555\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9398\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.61it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1563\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1563\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9379\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.18it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1507\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1507\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9403\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.03it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1386\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1386\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9440\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.54it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1403\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1403\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9462\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.76it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1374\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1374\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9472\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.78it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1395\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1395\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9452\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.86it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1411\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1411\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9434\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.22it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1271\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1271\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9518\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.22it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1265\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1265\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9530\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.38it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1199\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1199\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9534\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.62it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1217\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1217\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9538\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.42it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1134\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1134\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9566\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.49it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1155\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1155\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9557\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.97it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1059\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1059\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9574\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.46it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1014\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1014\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9620\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.62it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1107\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1107\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9592\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.38it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1022\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1022\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9601\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.85it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1041\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1041\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9604\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.05it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0958\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0958\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9629\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.58it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1039\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1039\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9622\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.66it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0946\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0946\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9631\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 62.06it/s]\n",
      "> Eval on experience 0 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9823\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0500\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9823\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.69it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.15it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 62.62it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 61.72it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 64.36it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.9340\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 12.4685\n",
      "\tStreamBWT/eval_phase/test_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1868\n",
      "Start of experience:  1\n",
      "Current Classes:  [8, 1]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 1 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 60.03it/s]\n",
      "> Eval on experience 1 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 15.3110\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.95it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.6451\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.6451\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8003\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.62it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2845\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2845\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8837\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.26it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2454\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2454\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9023\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.85it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2137\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2137\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9176\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.18it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1861\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1861\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9265\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.46it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1716\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1716\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9347\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.36it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1572\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1572\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9373\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.73it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1588\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1588\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9362\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.73it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1447\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1447\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9459\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.11it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1421\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1421\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9450\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.00it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1318\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1318\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9484\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.90it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1223\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1223\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9514\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.69it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1174\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1174\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9554\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.51it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1070\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1070\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9564\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.25it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1061\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1061\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9619\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.64it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1013\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1013\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9596\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.88it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0974\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0974\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9628\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.11it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0940\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0940\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9627\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.63it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0932\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0932\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9669\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.32it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0929\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0929\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9664\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.76it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0886\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0886\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9648\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.45it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0951\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0951\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9627\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.30it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0804\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0804\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9700\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.52it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0834\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0834\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9668\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.27it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0719\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0719\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9733\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.76it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0685\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0685\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9745\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.44it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0695\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0695\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9742\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.32it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0678\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0678\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9723\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.60it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0621\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0621\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9769\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.43it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0679\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0679\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9742\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.90it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0674\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0674\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9736\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.52it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0637\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0637\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9768\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.99it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0558\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0558\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9788\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.88it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0584\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0584\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9792\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.18it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0576\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0576\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9784\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.57it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0602\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0602\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9766\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.65it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0503\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0503\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9816\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.38it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0500\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0500\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9807\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.51it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0489\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0489\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9823\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.51it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0439\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0439\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9829\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.82it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0481\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0481\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9818\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.76it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0531\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0531\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9807\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.46it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0481\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0481\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9799\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.03it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0442\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0442\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9829\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.13it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0402\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0402\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9850\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.67it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0418\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0418\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9857\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.88it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0376\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0376\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9851\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.45it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0380\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0380\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9859\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.65it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0428\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0428\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9855\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.51it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0381\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0381\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9847\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 1 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 62.30it/s]\n",
      "> Eval on experience 1 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9946\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0179\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9946\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 58.43it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.9340\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.9340\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 66.76it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 62.22it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 58.57it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 62.94it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.4848\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 12.3892\n",
      "\tStreamBWT/eval_phase/test_stream = -0.9340\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.9340\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1939\n",
      "Start of experience:  2\n",
      "Current Classes:  [4, 5]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 2 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 63.68it/s]\n",
      "> Eval on experience 2 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 15.1720\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.32it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.7245\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.7245\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.7484\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.73it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.4017\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.4017\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8342\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.37it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3531\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3531\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8530\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.71it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3239\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3239\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8663\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.46it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3035\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3035\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8778\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.42it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2767\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2767\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8890\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.04it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2578\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2578\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8980\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.25it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2551\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2551\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8991\n",
      "100%|██████████| 157/157 [00:03<00:00, 41.47it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2368\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2368\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9033\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.84it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2317\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2317\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9078\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.10it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2222\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2222\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9094\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.21it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2210\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2210\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9092\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.50it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2076\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2076\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9160\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.94it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2050\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2050\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9177\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.39it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2007\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2007\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9182\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.82it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1841\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1841\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9265\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.72it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1799\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1799\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9295\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.64it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1848\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1848\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9276\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.33it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1728\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1728\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9318\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.19it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1659\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1659\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9349\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.77it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1626\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1626\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9374\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.07it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1632\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1632\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9369\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.33it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1655\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1655\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9349\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.78it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1505\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1505\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9403\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.06it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1532\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1532\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9430\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.18it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1361\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1361\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9472\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.88it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1380\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1380\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9448\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.96it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1423\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1423\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9464\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.36it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1374\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1374\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9438\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.18it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1314\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1314\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9476\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.72it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1250\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1250\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9516\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.29it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1393\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1393\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9450\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.07it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1222\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1222\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9535\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.20it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1211\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1211\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9553\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.06it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1168\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1168\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9539\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.22it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1209\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1209\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9508\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.14it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1255\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1255\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9521\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.14it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1119\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1119\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9578\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.96it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1023\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1023\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9615\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.55it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1061\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1061\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9591\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.22it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1028\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1028\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9602\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.20it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0992\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0992\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9591\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.88it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1065\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1065\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9602\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.84it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1008\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1008\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9623\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.74it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0969\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0969\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9613\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.01it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0867\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0867\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9660\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.99it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0890\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0890\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9664\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.82it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0902\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0902\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9665\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.91it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0832\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0832\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9691\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.55it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0889\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0889\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9644\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 2 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 59.95it/s]\n",
      "> Eval on experience 2 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9840\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0456\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9840\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.37it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.9340\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.9340\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.13it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.9695\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.9695\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 66.65it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 57.30it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 65.01it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.3102\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 12.9997\n",
      "\tStreamBWT/eval_phase/test_stream = -0.9518\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.9518\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1861\n",
      "Start of experience:  3\n",
      "Current Classes:  [0, 9]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 3 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 60.54it/s]\n",
      "> Eval on experience 3 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 17.2388\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.08it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.6185\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.6185\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8093\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.20it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2614\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2614\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8992\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.02it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2184\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2184\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9155\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.54it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1931\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1931\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9235\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.66it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1725\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1725\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9310\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.67it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1672\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1672\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9329\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.04it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1451\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1451\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9406\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.08it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1447\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1447\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9398\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.78it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1406\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1406\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9432\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.80it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1281\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1281\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9506\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.68it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1317\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1317\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9492\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.83it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1167\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1167\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9533\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.70it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1105\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1105\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9581\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.55it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1083\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1083\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9562\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.49it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0984\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0984\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9620\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.66it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1009\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1009\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9611\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.51it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0935\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0935\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9645\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.55it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0985\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0985\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9636\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.98it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0863\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0863\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9679\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.75it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0933\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0933\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9642\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.84it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0822\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0822\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9695\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.09it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0786\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0786\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9698\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.05it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0750\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0750\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9698\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.67it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0709\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0709\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9720\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.45it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0775\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0775\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9700\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.07it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0738\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0738\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9723\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.33it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0685\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0685\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9744\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.44it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0674\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0674\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9745\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.99it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0673\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0673\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9745\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.77it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0646\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0646\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9759\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.58it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0626\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0626\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9761\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.31it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0603\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0603\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9765\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.57it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0632\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0632\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9758\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.40it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0510\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0510\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9803\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.43it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0507\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0507\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9798\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.41it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0479\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0479\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9826\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.70it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0590\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0590\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9793\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.93it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0570\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0570\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9785\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.19it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0427\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0427\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9841\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.76it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0416\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0416\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9844\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.59it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0463\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0463\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9808\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.28it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0478\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0478\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9822\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.96it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0426\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0426\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9827\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.74it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0445\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0445\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9825\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.99it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0431\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0431\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9841\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.61it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0396\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0396\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9845\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.99it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0366\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0366\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9856\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.26it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0459\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0459\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9828\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.64it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0377\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0377\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9847\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.50it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0330\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0330\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9868\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 3 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 58.84it/s]\n",
      "> Eval on experience 3 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9893\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.0276\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9893\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 62.72it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.9340\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.9340\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 59.37it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.9695\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.9695\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 61.51it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp002 = -0.9305\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp002 = 0.9305\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 62.21it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 58.08it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.2351\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 11.7899\n",
      "\tStreamBWT/eval_phase/test_stream = -0.9447\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.9447\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1881\n",
      "Start of experience:  4\n",
      "Current Classes:  [3, 7]\n",
      "-- >> Start of training phase << --\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 4 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 62.43it/s]\n",
      "> Eval on experience 4 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.0000\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 15.0263\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.0000\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.09it/s]\n",
      "Epoch 0 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.8014\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.8014\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.6809\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.34it/s]\n",
      "Epoch 1 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3796\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3796\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8324\n",
      "100%|██████████| 157/157 [00:03<00:00, 50.93it/s]\n",
      "Epoch 2 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.3142\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.3142\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8634\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.33it/s]\n",
      "Epoch 3 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2908\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2908\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8749\n",
      "100%|██████████| 157/157 [00:03<00:00, 51.28it/s]\n",
      "Epoch 4 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2722\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2722\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8857\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.80it/s]\n",
      "Epoch 5 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2509\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2509\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.8974\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.44it/s]\n",
      "Epoch 6 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2384\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2384\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9034\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.43it/s]\n",
      "Epoch 7 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2376\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2376\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9015\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.29it/s]\n",
      "Epoch 8 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2173\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2173\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9122\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.63it/s]\n",
      "Epoch 9 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.2137\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.2137\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9129\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.37it/s]\n",
      "Epoch 10 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1968\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1968\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9189\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.60it/s]\n",
      "Epoch 11 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1980\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1980\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9195\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.99it/s]\n",
      "Epoch 12 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1909\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1909\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9206\n",
      "100%|██████████| 157/157 [00:02<00:00, 54.13it/s]\n",
      "Epoch 13 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1838\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1838\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9247\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.33it/s]\n",
      "Epoch 14 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1777\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1777\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9282\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.99it/s]\n",
      "Epoch 15 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1676\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1676\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9317\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.02it/s]\n",
      "Epoch 16 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1655\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1655\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9347\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.81it/s]\n",
      "Epoch 17 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1625\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1625\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9349\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.31it/s]\n",
      "Epoch 18 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1537\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1537\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9393\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.81it/s]\n",
      "Epoch 19 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1476\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1476\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9420\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.14it/s]\n",
      "Epoch 20 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1435\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1435\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9412\n",
      "100%|██████████| 157/157 [00:02<00:00, 54.02it/s]\n",
      "Epoch 21 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1423\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1423\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9413\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.25it/s]\n",
      "Epoch 22 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1335\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1335\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9452\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.95it/s]\n",
      "Epoch 23 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1372\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1372\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9456\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.93it/s]\n",
      "Epoch 24 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1347\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1347\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9461\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.94it/s]\n",
      "Epoch 25 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1267\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1267\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9502\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.96it/s]\n",
      "Epoch 26 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1346\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1346\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9462\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.95it/s]\n",
      "Epoch 27 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1278\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1278\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9487\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.71it/s]\n",
      "Epoch 28 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1190\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1190\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9537\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.88it/s]\n",
      "Epoch 29 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1109\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1109\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9572\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.60it/s]\n",
      "Epoch 30 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1081\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1081\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9563\n",
      "100%|██████████| 157/157 [00:02<00:00, 54.06it/s]\n",
      "Epoch 31 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1102\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1102\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9581\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.78it/s]\n",
      "Epoch 32 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1114\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1114\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9554\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.74it/s]\n",
      "Epoch 33 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1074\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1074\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9579\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.51it/s]\n",
      "Epoch 34 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.1042\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.1042\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9587\n",
      "100%|██████████| 157/157 [00:03<00:00, 52.07it/s]\n",
      "Epoch 35 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0994\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0994\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9628\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.91it/s]\n",
      "Epoch 36 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0915\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0915\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9638\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.03it/s]\n",
      "Epoch 37 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0938\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0938\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9636\n",
      "100%|██████████| 157/157 [00:02<00:00, 52.98it/s]\n",
      "Epoch 38 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0928\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0928\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9623\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.65it/s]\n",
      "Epoch 39 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0869\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0869\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9660\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.40it/s]\n",
      "Epoch 40 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0930\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0930\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9644\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.28it/s]\n",
      "Epoch 41 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0829\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0829\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9663\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.62it/s]\n",
      "Epoch 42 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0845\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0845\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9669\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.47it/s]\n",
      "Epoch 43 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0840\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0840\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9695\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.73it/s]\n",
      "Epoch 44 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0718\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0718\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9742\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.81it/s]\n",
      "Epoch 45 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0771\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0771\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9692\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.61it/s]\n",
      "Epoch 46 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0745\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0745\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9719\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.08it/s]\n",
      "Epoch 47 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0731\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0731\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9727\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.45it/s]\n",
      "Epoch 48 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0680\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0680\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9742\n",
      "100%|██████████| 157/157 [00:02<00:00, 53.63it/s]\n",
      "Epoch 49 ended.\n",
      "\tLoss_Epoch/train_phase/train_stream/Task000 = 0.0680\n",
      "\tRunningLoss_Epoch/train_phase/train_stream/Task000 = 0.0680\n",
      "\tTop1_RunningAcc_Epoch/train_phase/train_stream/Task000 = 0.9749\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 4 (Task 0) from train stream --\n",
      "100%|██████████| 157/157 [00:02<00:00, 64.73it/s]\n",
      "> Eval on experience 4 (Task 0) from train stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/train_stream/Task000 = 0.9432\n",
      "\tLoss_Stream/eval_phase/train_stream/Task000 = 0.1632\n",
      "\tStreamBWT/eval_phase/train_stream = 0.0000\n",
      "\tStreamForgetting/eval_phase/train_stream = 0.0000\n",
      "\tTop1_Acc_Stream/eval_phase/train_stream/Task000 = 0.9432\n",
      "-- >> End of training phase << --\n",
      "Training completed\n",
      "-- >> Start of eval phase << --\n",
      "-- Starting eval on experience 0 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 64.14it/s]\n",
      "> Eval on experience 0 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp000 = -0.9340\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp000 = 0.9340\n",
      "-- Starting eval on experience 1 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 73.87it/s]\n",
      "> Eval on experience 1 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp001 = -0.9695\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp001 = 0.9695\n",
      "-- Starting eval on experience 2 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 68.86it/s]\n",
      "> Eval on experience 2 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp002 = -0.9305\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp002 = 0.9305\n",
      "-- Starting eval on experience 3 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 65.16it/s]\n",
      "> Eval on experience 3 (Task 0) from test stream ended.\n",
      "\tExperienceBWT/eval_phase/test_stream/Task000/Exp003 = -0.9405\n",
      "\tExperienceForgetting/eval_phase/test_stream/Task000/Exp003 = 0.9405\n",
      "-- Starting eval on experience 4 (Task 0) from test stream --\n",
      "100%|██████████| 32/32 [00:00<00:00, 60.77it/s]\n",
      "> Eval on experience 4 (Task 0) from test stream ended.\n",
      "-- >> End of eval phase << --\n",
      "\tAccuracy_On_Trained_Experiences/eval_phase/test_stream/Task000 = 0.1780\n",
      "\tLoss_Stream/eval_phase/test_stream/Task000 = 12.4839\n",
      "\tStreamBWT/eval_phase/test_stream = -0.9436\n",
      "\tStreamForgetting/eval_phase/test_stream = 0.9436\n",
      "\tTop1_Acc_Stream/eval_phase/test_stream/Task000 = 0.1780\n"
     ]
    }
   ],
   "source": [
    "# TRAINING LOOP WITH EVALUATION AT EACH EXPERIENCE\n",
    "\n",
    "print('Starting experiment...')\n",
    "results = []\n",
    "for experience in split_cifar10_transform.train_stream:\n",
    "    print(\"Start of experience: \", experience.current_experience)\n",
    "    print(\"Current Classes: \", experience.classes_in_this_experience)\n",
    "\n",
    "    cl_strategy.train(experience)\n",
    "    print('Training completed')\n",
    "\n",
    "    exp_results.append(cl_strategy.eval(split_cifar10_transform.test_stream))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "aggregated_results = {}\n",
    "\n",
    "for i, exp_dict in enumerate(exp_results):\n",
    "   for metric_name, metric_value in exp_dict.items():\n",
    "        if metric_name not in aggregated_results:\n",
    "            aggregated_results[metric_name] = []\n",
    "            aggregated_results[metric_name].append(metric_value)\n",
    "        else:\n",
    "            aggregated_results[metric_name].append(metric_value)\n",
    "\n",
    "acc_stream_naive = aggregated_results['Top1_Acc_Stream/eval_phase/test_stream/Task000']\n",
    "acc_trained_exps_naive = aggregated_results['Accuracy_On_Trained_Experiences/eval_phase/test_stream/Task000']\n",
    "loss_stream_naive = aggregated_results['Loss_Stream/eval_phase/test_stream/Task000']"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compare DER++, ER and Naive on test stream"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Final average accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DER++:  0.5425\n",
      "Replay:  0.459\n",
      "Naive:  0.178\n"
     ]
    }
   ],
   "source": [
    "print('DER++: ', acc_stream_der[-1])\n",
    "print('Replay: ', acc_stream_er[-1])\n",
    "print('Naive: ', acc_stream_naive[-1])"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Incremental average accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DER++:  0.7008916666666667\n",
      "Replay:  0.6843833333333333\n",
      "Naive:  0.4284083333333334\n"
     ]
    }
   ],
   "source": [
    "print('DER++: ', np.mean(acc_trained_exps_der))\n",
    "print('Replay: ', np.mean(acc_trained_exps_er))\n",
    "print('Naive: ', np.mean(acc_trained_exps_naive))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plots over all experiences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiwAAAHFCAYAAADR1KI+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB5rElEQVR4nO3dd3RU1d7G8e9k0isJgQRIKALSkQ4BQXpTX9GLYENREZCiiFyKgDS9wYIiSrWAFbkKlquIBgRBQKWLgogIJoSETgrpM+f9Y2BgCCUJSWaSPJ+1ZjlzZp99fpMg87DP2fuYDMMwEBEREXFhbs4uQERERORaFFhERETE5SmwiIiIiMtTYBERERGXp8AiIiIiLk+BRURERFyeAouIiIi4PAUWERERcXkKLCIiIuLyFFhEitGcOXMwmUw0bNjQ2aVIAa1cuZKpU6c6uwyRMkeBRaQYvfPOOwD8/vvv/Pzzz06uRgpi5cqVTJs2zdlliJQ5CiwixWTr1q3s2rWLW2+9FYC3337byRVdWVpamrNLKBUMwyA9Pd3ZZYiUCgosIsXkfECZOXMmbdu25eOPP75sMIiPj2fw4MFERkbi6elJ5cqV6du3L0ePHrW3OXPmDE8//TQ33HADXl5eVKxYkd69e/PHH38AsG7dOkwmE+vWrXPo+9ChQ5hMJpYsWWLfNnDgQPz9/dm9ezfdu3cnICCALl26ABATE8Mdd9xBREQE3t7e1KpViyFDhnDixIlcdf/xxx/ce++9hIWF4eXlRdWqVXnwwQfJzMzk0KFDuLu7Ex0dnWu/9evXYzKZ+OSTT67684uNjeWBBx6gYsWKeHl5Ua9ePWbNmoXVas31+V5++WVeeeUVatSogb+/P1FRUfz0009X7R9sQW3MmDHUqFEDb29vQkJCaNGiBUuXLrX/rObOnQuAyWSyPw4dOmTfNmLECBYsWEC9evXw8vLi3XffBWD//v3cd999DvWf7+u8jIwMnn76aZo0aUJQUBAhISFERUXxxRdf5Kr1/LEWL15MnTp18PHxoUWLFvz0008YhsFLL71k//ydO3fmr7/+uubnF3Fl7s4uQKQsSE9PZ+nSpbRs2ZKGDRvyyCOPMGjQID755BMeeughe7v4+HhatmxJdnY2zzzzDI0bN+bkyZN8++23nD59mrCwMFJSUrj55ps5dOgQ48aNo3Xr1qSmprJ+/XoSEhKoW7duvuvLysri//7v/xgyZAjjx48nJycHgAMHDhAVFcWgQYMICgri0KFDvPLKK9x8883s3r0bDw8PAHbt2sXNN99MaGgo06dPp3bt2iQkJPDll1+SlZVF9erV+b//+z8WLFjA2LFjMZvN9mO/8cYbVK5cmTvvvPOK9R0/fpy2bduSlZXFjBkzqF69Ol999RVjxozhwIEDzJs3z6H93LlzqVu3LrNnzwZg8uTJ9O7dm4MHDxIUFHTF44wePZr333+f5557jqZNm3L27Fl+++03Tp48ae/n7NmzfPrpp2zevNm+X6VKlezPP//8czZs2MCzzz5LeHg4FStWZM+ePbRt25aqVasya9YswsPD+fbbb3niiSc4ceIEU6ZMASAzM5NTp04xZswYqlSpQlZWFqtXr+auu+5i8eLFPPjggw71fvXVV+zYsYOZM2diMpkYN24ct956Kw899BB///03b7zxBklJSYwePZp//etf7Ny5E5PJdMXPL+LSDBEpcu+9954BGAsWLDAMwzBSUlIMf39/o3379g7tHnnkEcPDw8PYs2fPFfuaPn26ARgxMTFXbLN27VoDMNauXeuw/eDBgwZgLF682L7toYceMgDjnXfeuepnsFqtRnZ2tvHPP/8YgPHFF1/Y3+vcubNRrlw549ixY9es6bPPPrNvi4+PN9zd3Y1p06Zd9djjx483AOPnn3922P74448bJpPJ2Ldvn8Pna9SokZGTk2Nv98svvxiAsXTp0qsep2HDhkafPn2u2mb48OHGlf7qBIygoCDj1KlTDtt79OhhREREGElJSQ7bR4wYYXh7e+dqf15OTo6RnZ1tPProo0bTpk1zHSs8PNxITU21b/v8888NwGjSpIlhtVrt22fPnm0Axq+//nrVzybiynRKSKQYvP322/j4+HDPPfcA4O/vz913382GDRvYv3+/vd0333xDp06dqFev3hX7+uabb7jxxhvp2rVrodb4r3/9K9e2Y8eOMXToUCIjI3F3d8fDw4Nq1aoBsHfvXsB2GuWHH36gX79+VKhQ4Yr9d+zYkZtuusnhNMiCBQswmUwMHjz4qrV9//331K9fn1atWjlsHzhwIIZh8P333ztsv/XWWx1GcRo3bgzAP//8c9XjtGrVim+++Ybx48ezbt26Al1/0rlzZ4KDg+2vMzIyWLNmDXfeeSe+vr7k5OTYH7179yYjI8PhdNUnn3xCu3bt8Pf3t//M3377bfvP+2KdOnXCz8/P/vr8n5tevXo5jKSc336tzy/iyhRYRIrYX3/9xfr167n11lsxDIMzZ85w5swZ+vbtC1yYOQS2Ux8RERFX7S8vbfLL19eXwMBAh21Wq5Xu3buzYsUKxo4dy5o1a/jll1/sX67nv8xPnz6NxWLJU01PPPEEa9asYd++fWRnZ/Pmm2/St29fwsPDr7rfyZMnHU67nFe5cmX7+xcrX768w2svLy+Hmq9kzpw5jBs3js8//5xOnToREhJCnz59HELltVxa58mTJ8nJyeH111/Hw8PD4dG7d28A+zVBK1asoF+/flSpUoUPPviAzZs3s2XLFh555BEyMjJyHSskJMThtaen51W3X64PkZJC17CIFLF33nkHwzD49NNP+fTTT3O9/+677/Lcc89hNpupUKEChw8fvmp/eWnj7e0N2K6JuNjlLpYFLntdw2+//cauXbtYsmSJw3U2l168GRISgtlsvmZNAPfddx/jxo1j7ty5tGnThsTERIYPH37N/cqXL09CQkKu7UeOHAEgNDT0mn3khZ+fH9OmTWPatGkcPXrUPtpy++232y9ovpZLf5bBwcGYzWYGDBhwxc9ao0YNAD744ANq1KjBsmXLHPq59PcoUhZphEWkCFksFt59911q1qzJ2rVrcz2efvppEhIS+OabbwDbUP7atWvZt2/fFfvs1asXf/75Z67TIBerXr06AL/++qvD9i+//DLPtZ//wjw/OnHewoULHV77+Phwyy238Mknn1wxEJ3n7e3N4MGDeffdd3nllVdo0qQJ7dq1u2YtXbp0Yc+ePWzfvt1h+3vvvYfJZKJTp055+Uj5EhYWxsCBA7n33nvZt2+ffUZXXkdrzvP19aVTp07s2LGDxo0b06JFi1yP8yNCJpMJT09Ph7CSmJh42VlCImWNRlhEitA333zDkSNHeOGFF+jYsWOu9xs2bMgbb7zB22+/zW233cb06dP55ptv6NChA8888wyNGjXizJkzrFq1itGjR1O3bl1GjRrFsmXLuOOOOxg/fjytWrUiPT2dH374gdtuu41OnToRHh5O165diY6OJjg4mGrVqrFmzRpWrFiR59rr1q1LzZo1GT9+PIZhEBISwv/+9z9iYmJytT0/c6h169aMHz+eWrVqcfToUb788ksWLlxIQECAve2wYcN48cUX2bZtG2+99Vaeannqqad47733uPXWW5k+fTrVqlXj66+/Zt68eTz++OPceOONef5cV9O6dWtuu+02GjduTHBwMHv37uX9998nKioKX19fABo1agTACy+8QK9evTCbzTRu3Nh+2uVyXnvtNW6++Wbat2/P448/TvXq1UlJSeGvv/7if//7nz183nbbbaxYsYJhw4bRt29f4uLimDFjBpUqVcrXaSmRUsm51/yKlG59+vQxPD09rzp75p577jHc3d2NxMREwzAMIy4uznjkkUeM8PBww8PDw6hcubLRr18/4+jRo/Z9Tp8+bTz55JNG1apVDQ8PD6NixYrGrbfeavzxxx/2NgkJCUbfvn2NkJAQIygoyHjggQeMrVu3XnaWkJ+f32Vr27Nnj9GtWzcjICDACA4ONu6++24jNjbWAIwpU6bkanv33Xcb5cuXNzw9PY2qVasaAwcONDIyMnL127FjRyMkJMRIS0vLy4/RMAzD+Oeff4z77rvPKF++vOHh4WHUqVPHeOmllwyLxWJvc36W0EsvvZRr/8vVfKnx48cbLVq0MIKDgw0vLy/jhhtuMJ566injxIkT9jaZmZnGoEGDjAoVKhgmk8kAjIMHD9qPMXz48Mv2ffDgQeORRx4xqlSpYnh4eBgVKlQw2rZtazz33HMO7WbOnGlUr17d8PLyMurVq2e8+eabxpQpU3LNTLrcsa70+c/P0Prkk0+u+vlFXJnJMAzDWWFJRMqeY8eOUa1aNUaOHMmLL77o7HJEpITQKSERKRaHDx/m77//5qWXXsLNzY0nn3zS2SWJSAmii25FpFi89dZbdOzYkd9//50PP/yQKlWqOLskESlBdEpIREREXJ5GWERERMTlKbCIiIiIy1NgEREREZdXamYJWa1Wjhw5QkBAgG6fLiIiUkIYhkFKSgqVK1fGze3K4yilJrAcOXKEyMhIZ5chIiIiBRAXF3fVm6iWmsByfunvuLi4XHedFREREdeUnJxMZGSkwy08LqfUBJbzp4ECAwMVWEREREqYa13OoYtuRURExOUpsIiIiIjLU2ARERERl1ega1jmzZvHSy+9REJCAg0aNGD27Nm0b9/+sm3XrVtHp06dcm3fu3cvdevWBWDJkiU8/PDDudqkp6fj7e1dkBIvy2q1kpWVVWj9yQWenp5XnY4mIiJyPfIdWJYtW8aoUaOYN28e7dq1Y+HChfTq1Ys9e/ZQtWrVK+63b98+h4thK1So4PB+YGAg+/btc9hWmGElKyuLgwcPYrVaC61PucDNzY0aNWrg6enp7FJERKQUyndgeeWVV3j00UcZNGgQALNnz+bbb79l/vz5REdHX3G/ihUrUq5cuSu+bzKZCA8Pz285eWIYBgkJCZjNZiIjIzUSUMjOL9qXkJBA1apVtXCfiIgUunwFlqysLLZt28b48eMdtnfv3p1NmzZddd+mTZuSkZFB/fr1mTRpUq7TRKmpqVSrVg2LxUKTJk2YMWMGTZs2vWJ/mZmZZGZm2l8nJydfsW1OTg5paWlUrlwZX1/fq9YpBVOhQgWOHDlCTk4OHh4ezi5HRERKmXwNNZw4cQKLxUJYWJjD9rCwMBITEy+7T6VKlVi0aBHLly9nxYoV1KlThy5durB+/Xp7m7p167JkyRK+/PJLli5dire3N+3atWP//v1XrCU6OpqgoCD742qr3FosFgCdrihC53+253/WIiIihalAF91eOuRvGMYVTwPUqVOHOnXq2F9HRUURFxfHyy+/TIcOHQBo06YNbdq0sbdp164dzZo14/XXX2fOnDmX7XfChAmMHj3a/vr8Snn5qVsKj362IiJSlPI1whIaGorZbM41mnLs2LFcoy5X06ZNm6uOnri5udGyZcurtvHy8rKvaqvVbUVEREq3fAUWT09PmjdvTkxMjMP2mJgY2rZtm+d+duzYQaVKla74vmEY7Ny586ptJO86duzIqFGjnF2GiIhIgeX7lNDo0aMZMGAALVq0ICoqikWLFhEbG8vQoUMB26ma+Ph43nvvPcA2i6h69eo0aNCArKwsPvjgA5YvX87y5cvtfU6bNo02bdpQu3ZtkpOTmTNnDjt37mTu3LmF9DFLpoEDB/Luu+8SHR3tcKHz559/zp133olhGHnqZ8WKFboQVkRESrR8B5b+/ftz8uRJpk+fTkJCAg0bNmTlypVUq1YNgISEBGJjY+3ts7KyGDNmDPHx8fj4+NCgQQO+/vprevfubW9z5swZBg8eTGJiIkFBQTRt2pT169fTqlWrQviIJZu3tzcvvPACQ4YMITg4uEB9hISEFHJVIiJSlmRbsvn1xK80D2vutBoKtCDJsGHDOHToEJmZmWzbts1+8SzYVq1dt26d/fXYsWP566+/SE9P59SpU2zYsMEhrAC8+uqr/PPPP2RmZnLs2DG+/fZboqKiCvaJSpmuXbsSHh5+xTVuTp48yb333ktERAS+vr40atSIpUuXOrS5+JTQhAkTHC5wPq9x48ZMmTLF/nrx4sXUq1cPb29v6taty7x58wrvQ4mIiMtLykziq7+/YswPY+iwrAMDVw0kNjn22jsWkQLNEirpDMMgPds50299PMz5mlFjNpv5z3/+w3333ccTTzxBRESEw/sZGRk0b96ccePGERgYyNdff82AAQO44YYbaN26da7+7r//fmbOnMmBAweoWbMmAL///ju7d+/m008/BeDNN99kypQpvPHGGzRt2pQdO3bw2GOP4efnx0MPPXQdn15ERFxZbHIs6+LWse7wOrYf3Y7FuPBdGeIdwuGUw1QNvPKq9kWpTAaW9GwL9Z/91inH3jO9B76e+fux33nnnTRp0oQpU6bw9ttvO7xXpUoVxowZY389cuRIVq1axSeffHLZwNKwYUMaN27MRx99xOTJkwH48MMPadmyJTfeeCMAM2bMYNasWdx1110A1KhRgz179rBw4UIFFhGRUsRitbD7xG7Wxq1lXdw6/k762+H9WuVq0SmyEx0jO9IwtCFuJuetFF8mA0tJ9MILL9C5c2eefvpph+0Wi4WZM2eybNky4uPj7SsA+/n5XbGv+++/n3feeYfJkydjGAZLly61nzI6fvw4cXFxPProozz22GP2fXJycggKCiqSzyYiIsUnLTuNzQmbWRe3jvWH13Mq45T9PXeTO83DmtMxsiO3RN5CZMDV1zcrTmUysPh4mNkzvYfTjl0QHTp0oEePHjzzzDMMHDjQvn3WrFm8+uqrzJ49m0aNGuHn58eoUaOuelfq++67j/Hjx7N9+3bS09OJi4vjnnvuAbDfHPLNN9/MNUJjNhesdhERca5jacdsp3ri1vFzws9kWS98RwR4BHBzxM10iuxEuyrtCPR0zXXNymRgMZlM+T4t4wpmzpxJkyZN7KduADZs2MAdd9zBAw88ANgCx/79+6lXr94V+4mIiKBDhw58+OGHpKen07VrV/vCf2FhYVSpUoW///6b+++/v2g/kIiIFAnDMPjz9J/2Uz2/n/zd4f0q/lXsp3qahTXDw831l74oed/aZVijRo24//77ef311+3batWqxfLly9m0aRPBwcG88sorJCYmXjWwgO200NSpU8nKyuLVV191eG/q1Kk88cQTBAYG0qtXLzIzM9m6dSunT592uB2CiIi4jmxLNlsSt7DusG0kJeFsgsP7jUMb0zGyIx0jO1KrXK0Sd0sVBZYSZsaMGfz3v/+1v548eTIHDx6kR48e+Pr6MnjwYPr06UNSUtJV+7n77rsZOXIkZrOZPn36OLw3aNAgfH19eemllxg7dix+fn40atRIq+WKiLiYpMwk1h9ez7q4dWw8spGz2Wft73mbvWlTuQ2dIjvRIaIDoT6hziu0EJiMvC6X6uKSk5MJCgoiKSkp132FMjIyOHjwIDVq1MDb29tJFZZu+hmLiBSPf5L/sV+PsuPYDoepx+W9y9tHUVpXao2Pu4/T6syrq31/X0wjLCIiIi7MYrXw64lf7dejHEw66PB+7eDadIzo6BJTj4uSAouIiIiLSctOY/ORzayNW8v6w+s5nXna/p67yZ3m4c3pFNmJWyJuISIg4io9lR4KLCIiIi7g6Nmj/HD4h8tPPfYMoH2V9nSM7OjSU4+LkgKLiIiIExiGwb7T++ynevac3OPw/vmpx50iO9E0rGmJmHpclBRYREREikmWJYutiVttIeXwOhLPJtrfM2GiUYVGtvVRIjpSs1zNEjf1uCgpsIiIiBShMxln2BC/gbVxa9l0ZFOuqcdRlaPoFNmJ9hHtS/zU46KkwCIiIlLIzk89Xhu3lh3HdmA1rPb3Qn1CuSXiFjpFdqJ1pdZ4u2spiLxQYBEREblOFquFXcd32VeZvXTq8Y3BN9pDSoPQBqV26nFRUmAREREpgLTsNDYd2cTauLVsOLwh19TjFuEt7Iu4VfGv4sRKSwcFFhERkTxKPJvI+sPrWRu3lp8Tfibbmm1/7/zU4/N3PQ7wDHBipaWPAosLGzhwIO+++26u7T169GDVqlVUr16df/75BwBvb2+qVavGo48+ypgxY3RluYhIITAMgz9O/WFbCv9w7qnHEf4RdKpqm3rcpGKTMj/1uCgpsLi4nj17snjxYodtXl5e9ufTp0/nscceIyMjg9WrV/P4448TGBjIkCFDirtUEZFSIcuSxZbELayNW8sPh3/INfW4cQXbXY87RXbihqAb9A/EYqLA4uK8vLwIDw+/4vsBAQH29wcNGsT8+fP57rvvFFhERPLh4qnHG+M3kpaTZn/Px92HqEpRdIzsSIeIDpT3Ke/ESsuushlYDAOy067drih4+EIRpHHDMPjhhx/Yu3cvtWvXLvT+RURKm0NJh+xTj3ce3+kw9biCTwVuibTN6mkV3kpTj11A2Qws2Wnwn8rOOfYzR8DTL8/Nv/rqK/z9/R22jRs3jsmTJ9ufT5o0iaysLLKzs/H29uaJJ54o1JJFREoD+9TjOFtIOZR8yOH9OsF17CGlfvn6mnrsYspmYClBOnXqxPz58x22hYSE2J//+9//ZuDAgRw/fpyJEyfSuXNn2rZtW9xlioi4pLPZZ9l0ZBPr4tax/vB6zmSesb/n7uZOy7CW9qnHlf2d9A9ZyZOyGVg8fG0jHc46dj74+flRq1atK74fGhpKrVq1qFWrFsuXL6dWrVq0adOGrl27Xm+lIiIlUuLZRH6I+4G1h9fyS8IvDlOPAz0DaR9x7q7HlTX1uCQpm4HFZMrXaZmSIjg4mJEjRzJmzBh27NihK9dFpEwwDIO9p/baQkrcWvae2uvwfmRApO2GgpEdaVqxKe5uZfOrr6TTb83FZWZmkpiY6LDN3d2d0NDL3yBr+PDhvPDCCyxfvpy+ffsWR4kiIsUuy5LFL4m/2NZHiVvH0bSj9vdMmLipwk32qcc1gmroH3ClgAKLi1u1ahWVKlVy2FanTh3++OOPy7avUKECAwYMYOrUqdx11124uemiMREpHU5nnGZD/AbWxa3T1OMyyGQYhuHsIgpDcnIyQUFBJCUlERgY6PBeRkYGBw8epEaNGnh7a2paUdDPWESKwsGkg/ZRlMtNPT5/waymHpdcV/v+vphGWERExGXkWHPsU4/Xxa277NTj86d66pWvp6nHZYgCi4iIONW1ph63Cm/FLRG3aOpxGafAIiIixS7xbKJ9FOWXxNxTjztEdLBPPfb39L9iP1J2KLCIiEiROz/1+HxIuXTqcdWAqnSK7MQtkbdo6rFclv5EiIhIkbEaVj7Z9wlv7n4z19TjJhWb2C+arRGoqcdydQosIiJSJOJS4piyaQpbErcAtqnHbSu3tU89DvEOuUYPIhcU6PLqefPm2aevNm/enA0bNlyx7bp16zCZTLkel64jsnz5curXr4+Xlxf169fns88+K0hpIiLiZFbDyod7P+RfX/6LLYlb8DZ7M67lODbcs4HZnWbTp1YfhRXJt3wHlmXLljFq1CgmTpzIjh07aN++Pb169SI2Nvaq++3bt4+EhAT7o3bt2vb3Nm/eTP/+/RkwYAC7du1iwIAB9OvXj59//jn/n0hERJzmn+R/eHjVw8z8ZSbpOem0CGvBiv9bwQP1H8DL7OXs8qQEy/fCca1bt6ZZs2YOdxCuV68effr0ITo6Olf7devW0alTJ06fPk25cuUu22f//v1JTk7mm2++sW/r2bMnwcHBLF26NE91aeE459LPWKRss1gtfLj3Q17f8ToZlgx83H0Y3Xw0/er001opclV5XTguX3+KsrKy2LZtG927d3fY3r17dzZt2nTVfZs2bUqlSpXo0qULa9eudXhv8+bNufrs0aPHVfvMzMwkOTnZ4SEiIsXvYNJBHlr1EC9tfYkMSwatw1uz4v9WcE/dexRWpNDk60/SiRMnsFgshIWFOWwPCwvLdYO+8ypVqsSiRYtYvnw5K1asoE6dOnTp0oX169fb2yQmJuarT4Do6GiCgoLsj8jIyPx8lBJh4MCB9mt+PDw8CAsLo1u3brzzzjtYrReWp65evfplrxOaOXMmAIcOHXLYHhQURJs2bfjf//7nrI8mIqWAxWph8W+L6ftlX3Yd34Wfhx/PRj3Lm93fJCIgwtnlSSlToFlCl049MwzjitPR6tSpQ506deyvo6KiiIuL4+WXX6ZDhw4F6hNgwoQJjB492v46OTm5VIaWnj17snjxYiwWC0ePHmXVqlU8+eSTfPrpp3z55Ze4u9t+hdOnT+exxx5z2DcgIMDh9erVq2nQoAFnzpxh3rx5/Otf/2L79u00bNiw2D6PiJQOB84cYPLGyew+sRuAtpXbMjVqKpX8K11jT5GCyVdgCQ0NxWw25xr5OHbsWK4Rkqtp06YNH3zwgf11eHh4vvv08vLCy6v0X8Dl5eVFeHg4AFWqVKFZs2a0adOGLl26sGTJEgYNGgTYwsn5dldSvnx5wsPDCQ8P5/nnn+f1119n7dq1Ciwikmc51hyW/L6EeTvnkW3Nxt/Dn7Etx9KnVh+toyJFKl+nhDw9PWnevDkxMTEO22NiYmjbtm2e+9mxYweVKl1I4VFRUbn6/O677/LVZ34YhkFadppTHoVxc+zOnTtz0003sWLFigLtn52dzZtvvgmAh4fHddcjImXDn6f/5P6V9/Pa9tfItmbTvkp7PrvjM+6sfafCihS5fJ8SGj16NAMGDKBFixZERUWxaNEiYmNjGTp0KGA7VRMfH897770HwOzZs6levToNGjQgKyuLDz74gOXLl7N8+XJ7n08++SQdOnTghRde4I477uCLL75g9erV/Pjjj4X0MR2l56TT+qPWRdL3tfx838/4evhedz9169bl119/tb8eN24ckyZNcmjz1Vdf0bFjR/vrtm3b4ubmRnp6OlarlerVq9OvX7/rrkVESrdsazZv736bhb8uJMeaQ4BnAONbjef2G25XUJFik+/A0r9/f06ePMn06dNJSEigYcOGrFy5kmrVqgGQkJDgsCZLVlYWY8aMIT4+Hh8fHxo0aMDXX39N79697W3atm3Lxx9/zKRJk5g8eTI1a9Zk2bJltG7tnFBRElx6jc+///1vBg4c6NCmSpUqDq+XLVtG3bp1+fPPPxk1ahQLFiwgJESLN4nIlf1x6g8mb5zMH6dsi312jOzI5DaTqehb0cmVSVlToItuhw0bxrBhwy773pIlSxxejx07lrFjx16zz759+9K3b9+ClJNvPu4+/Hyfcxal83H3KZR+9u7dS40aNeyvQ0NDqVWr1lX3iYyMpHbt2tSuXRt/f3/+9a9/sWfPHipW1F88IuIo25LNot2LeOvXt8gxcgjyCmJCqwn0rtFboyriFGXyXkImk6lQTss4y/fff8/u3bt56qmnCtzHLbfcQsOGDXn++ed57bXXCrE6ESnpfj/5O5M3Tmb/6f0AdK3alYltJhLqE+rkyqQsK5OBpSTJzMwkMTHRYVpzdHQ0t912Gw8++KC9XUpKSq6ZVr6+vlddNfDpp5/m7rvvZuzYsblOH4lI2ZNlyWLBrgW889s7WAwLwV7BPNPmGXpU66FRFXE6LUHo4latWkWlSpWoXr06PXv2ZO3atcyZM4cvvvgCs9lsb/fss89SqVIlh8e1TsXddtttVK9eneeff76oP4aIuLjdx3fT73/9eHP3m1gMCz2q9+DzPp/Ts3pPhRVxCfm+l5Cr0r2EnEs/Y5GSKdOSydydc3n393exGlZCvEOY1GYS3ap1c3ZpUkbk9V5COiUkIlJG7Ty2k8kbJ3Mo+RAAvWv0Znyr8QR7Bzu3MJHLUGARESlj0nPSeWPHG7y/530MDEJ9QpncZjKdq3Z2dmkiV6TAIiJShmw7uo1nNz5LbIptvaz/q/l/jG05liCvICdXJnJ1CiwiImVAWnYac3bM4aO9H2FgUNGnIlPaTqFDRIdr7yziAspUYCkl1xe7JP1sRVzXlsQtPLvxWQ6nHgbgzlp3MqblGAI9r3yBo4irKROB5fz036ysLHx8CmelWXGUlZUF4DDVWkScKy07jVe2vcKyfcsACPMNY2rbqdxc5WYnVyaSf2UisLi7u+Pr68vx48fx8PDAzU3LzxQmq9XK8ePH8fX1xd29TPyREnF5PyX8xJSNUzhy9ggAfW/sy9PNn8bf09/JlYkUTJn4djGZTFSqVImDBw/yzz//OLucUsnNzY2qVatqgSkRJ0vNSmXWtll8+uenAFT2q8zUtlOJqhzl5MpErk+ZCCwAnp6e1K5d237qQgqXp6enRq5EnGxj/Eambp5K4lnbbTr61+nPU82fws/Dz8mViVy/MhNYwDYKoFVYRaS0Sc5K5uUtL/PZX58BUMW/CtPbTqdVpVZOrkyk8JSpwCIiUtqsP7yeaZuncSztGAD317ufJ5o+UaLvSC9yOQosIiIlUFJmEi9ueZEvD3wJQNWAqkxvN53mYc2dXJlI0VBgEREpYdbGrmX6T9M5kX4CEyYG1B/AiKYj8HHXsg1SeimwiIiUEGcyzhD9SzQrD64EoHpgdWa0m0GTik2cW5hIMVBgEREpAVb/s5oZP83gVMYp3ExuPNTgIYbdNAxvd00kkLJBgUVExIWdyjhF9M/RrDq0CoCaQTWZ3m46jSs0dnJlIsVLgUVExEV9e+hb/vPzfziVcQqzycwjDR9h6E1D8TR7Ors0kWKnwCIi4mJOpJ/gPz//h5h/YgCoVa4Wz938HA3KN3ByZSLOo8AiIuIiDMPgm4PfEP1LNGcyz+BucmdQ40EMbjQYD7OHs8sTcSoFFhERF3A87TgzfprB2ri1ANQJrsNzNz9H3ZC6Tq5MxDUosIiIOJFhGHz191fM/GUmyVnJuLu5M7jxYAY1GoSHm0ZVRM5TYBERcZKjZ48y/afprD+8HoB6IfV47ubnuDH4RidXJuJ6FFhERIqZYRh8/tfnvLTlJVKyU/Bw8+Dxmx5nYMOBGlURuQIFFhGRYpR4NpGpm6ay8chGABqFNmJ62+nUCq7l5MpEXJsCi4hIMTAMg+X7l/Py1pc5m30WTzdPhjcdzoP1H8TdTX8Vi1yL/i8RESli8anxTN00lZ8SfgLgpgo3Mb3ddG4IusHJlYnkUdZZOLIDqrYFNzenlKDAIiJSRKyGlU/2fcIr214hLScNL7MXI5uO5IF6D2B2Mzu7PJHLMww4+Rcc3nLhcXQPGBYYvgUqOOeicAUWEZEiEJcSx5RNU9iSuAWAZhWbMb3ddKoFVnNyZSKXSD8D8dvg8NYLASXjTO52gVUgNVGBRUSkNLAaVpb+sZTXtr9Gek463mZvRjUfxb1178XN5JyhdBE7qwWO7T0XTM4FlBP7crdz94bKTSGiBUS0hCotIKhK8dd7cUlOPbqISCnyT/I/PLvxWbYf2w5Ai7AWTG87ncjASCdXJmVW6nGIv2jkJH47ZKXmbhdcwxZMIlraQkp4I3Cx20EosIiIXCeL1cKHez/k9R2vk2HJwMfdh9HNR9OvTj+NqkjxycmCo7sdT+2cPpS7nac/VGnuGFD8Qou93PwqUGCZN28eL730EgkJCTRo0IDZs2fTvn37a+63ceNGbrnlFho2bMjOnTvt25csWcLDDz+cq316ejre3t4FKVFEpFgcTDrI5I2T2XV8FwCtw1szte1UIgIinFyZlHpJ8RddGLsVEnZCTkbudhXqXji1E9EKKtSBEnjRd74Dy7Jlyxg1ahTz5s2jXbt2LFy4kF69erFnzx6qVq16xf2SkpJ48MEH6dKlC0ePHs31fmBgIPv2OZ5HU1gREVdlsVp4b897vLHjDbKsWfh5+PF0i6fpW7svJpPJ2eVJaZOdDkd2OgaUlCO52/kEO46cVG4GPuWKu9oike/A8sorr/Doo48yaNAgAGbPns23337L/PnziY6OvuJ+Q4YM4b777sNsNvP555/net9kMhEeHp7fckREit2BMweYvHEyu0/sBqBt5bZMjZpKJf9KTq5MSgXDgNMHL5zaifsFjv4G1hzHdiYzhDWAyFYXQkrIDVBKA3O+AktWVhbbtm1j/PjxDtu7d+/Opk2brrjf4sWLOXDgAB988AHPPffcZdukpqZSrVo1LBYLTZo0YcaMGTRt2vSKfWZmZpKZmWl/nZycnJ+PIiKSbznWHJb8voR5O+eRbc3G38OfsS3H0qdWH42qSMFlJMOR7Y4zd9JO5m7nH3bR6ElLqNwEPP2KvVxnyVdgOXHiBBaLhbCwMIftYWFhJCYmXnaf/fv3M378eDZs2IC7++UPV7duXZYsWUKjRo1ITk7mtddeo127duzatYvatWtfdp/o6GimTZuWn/JFRArsz9N/MnnjZPac3ANA+yrteTbqWcL9NDIs+WC1wok/HRdlO7YXMBzbmT2h0k22a07OX38SFFFqR0/yokAX3V76LwnDMC77rwuLxcJ9993HtGnTuPHGKy8006ZNG9q0aWN/3a5dO5o1a8brr7/OnDlzLrvPhAkTGD16tP11cnIykZGaOigihSvbms3bu99m4a8LybHmEOAZwPhW47n9hts1qiLXlnbKcdZO/DbIvMwZgXJVHUdPwhuBu1fx1+vC8hVYQkNDMZvNuUZTjh07lmvUBSAlJYWtW7eyY8cORowYAYDVasUwDNzd3fnuu+/o3Llzrv3c3Nxo2bIl+/fvv2ItXl5eeHnplykiReePU38weeNk/jj1BwAdIzsyuc1kKvpWdHJl4pIsOXDs9wunduJ+gVMHcrfz8D03rfiiRdkCcn+HiqN8BRZPT0+aN29OTEwMd955p317TEwMd9xxR672gYGB7N6922HbvHnz+P777/n000+pUaPGZY9jGAY7d+6kUaNG+SlPRKRQZFuyWbR7EW/9+hY5Rg5BXkFMaDWB3jV6a1RFLkhJdJy1c2QHZKflble+9oVZOxEtoWJ9MGsZtPzK909s9OjRDBgwgBYtWhAVFcWiRYuIjY1l6NChgO1UTXx8PO+99x5ubm40bNjQYf+KFSvi7e3tsH3atGm0adOG2rVrk5yczJw5c9i5cydz5869zo8nIpI/v5/8nckbJ7P/tG2Et2vVrkxsM5FQH9dfWEuKUE4mJPzqeO1JUlzudl5BF6150hKqNAPfkOKvtxTKd2Dp378/J0+eZPr06SQkJNCwYUNWrlxJtWq2G3olJCQQGxubrz7PnDnD4MGDSUxMJCgoiKZNm7J+/XpatWqV3/JERAoky5LFgl0LeOe3d7AYFoK9gnmmzTP0qNZDoypljWHAmVjHWTuJv4Ily7Gdyc02WnLxomzla4GbVjcuCibDMIxrN3N9ycnJBAUFkZSURGBgoLPLEZESZPfx3UzeOJkDSbbrDXpU78EzrZ8hxFv/Mi4Tss7aTudcHFBScy9wim/ouTVPzgWUyk3BK6D46y1l8vr9rZNoIlJmZVoymbtzLu/+/i5Ww0qIdwiT2kyiW7Vuzi5NiophwMm/HE/tHN0DhsWxnZs7hDe2BZPzIaVctTI9rdjZFFhEpEzaeWwnkzdO5lDyIQB61+jN+FbjCfYOdm5hUrjSz9imEl88tTjjTO52gRGO155UagwePsVdrVyFAouIlCnpOem8seMN3t/zPgYGoT6hTG4zmc5Vcy+xICWM1WJbhO3iUzsn9uVu5+5tO50T0eLCwmyBlYu/XskXBRYRKTO2Hd3GsxufJTbFNjHg/2r+H2NbjiXIK8jJlUmBpB6H+IsXZdsOWam524Xc4HhDwLCGYPYo/nrluiiwiEipl5adxpwdc/ho70cYGFT0qciUtlPoENHB2aVJXuVkwdHdjqd2Th/K3c4zACKaXzStuAX4lS/2cqXwKbCISKm2JXELz258lsOphwG4s9adjGk5hkBPzSZ0aUnxjhfGHtkJlsxLGpmgQl3Ha08q1AE3szMqliKmwCIipVJadhqvbHuFZfuWARDmG8bUtlO5ucrNTq5McslOtwWSi1eNTTmSu51PiOOpnSrNwFun88oKBRYRKXV+SviJqZumEp8aD0DfG/vydPOn8ff0d3JlgmHA6YMQd/G04t/AmuPYzmSG8IaONwQMuUHTisswBRYRKTVSs1J5ZdsrfPLnJwBU9qvM1LZTiaoc5eTKyrDMlHPTii+auZN2Mnc7/3CIvCicVGoCnr7FXq64LgUWESkVNsVvYsrmKSSetd1Nvn+d/jzV/Cn8PPycXFkZcn5Rtrhf4PAvtoBybA8YVsd2Zk9bIIloeSGkBFbR6IlclQKLiJRoKVkpvLz1ZVbsXwFAFf8qTG87nVaVdC+yIpeRfGH0JO6XKy/KVq7qhXvtRLS0nepx9yr2cqVkU2ARkRJr/eH1TNs8jWNpxwC4v979PNH0CXw9dCqh0FmtcHK/Yzg5the45HZ07t5QuZntotjIcwElINwpJUvposAiIiVOUmYSL255kS8PfAlA1YCqTG83neZhzZ1cWSmSkXTumpOtF07vXHb0pNq5YHJuxdjwRlqUTYqEAouIlChrY9cy46cZHE8/jgkTA+oPYETTEfi4674vBWa1wok/zwWTLbYZPMf/IPfoiY9tKrH9hoAtwb+iU0qWskeBRURKhDMZZ5i5ZSZf//01ANUDqzOj3QyaVGzi3MJKovQztiXt47acCynbIDMpd7vgGo7hJKyBRk/EaRRYRMTlrflnDTN+msHJjJO4mdx4qMFDDLtpGN7u3s4uzfVZrbYbAJ6fuRN3hRsCevhCleYX3RCwJfhXKP56Ra5AgUVEXNqHez9k5i8zAagZVJPp7abTuEJjJ1flwtJP20ZMDv9iCynx2yAzOXe7kBsuXHcS2QoqNgCzvhLEdelPp4i4rK///toeVh6o9wCjmo/Cy6zpsHZWi+1ak7hfLlwce+LP3O08/GzXnpw/tRPREvxCi79ekeugwCIiLmlj/EYm/TgJgPvq3sfYlmMxlfWFxdJOXTRrZ4ttJCUrJXe7kJoXwklkK6hQT6MnUuLpT7CIuJxfj//KU+ueIsfIoVf1XoxrNa7shRWrxbbOyfnrTg7/YltF9lKe/ueuPbno4ljfkOKvV6SIKbCIiEv5+8zfDF8znPScdNpWbsvzNz+Pm8nN2WUVvbMnz83cOXdxbPx2yErN3a587XPB5NzFsRXrgZu5+OsVKWYKLCLiMhLPJjJk9RDOZJ6hUWgjXu34Kh6lcRqtJcd2j53zC7LF/QKnDuRu5xkAEc0vzNqJaKHREymzFFhExCWcyTjDkJghJJ5NpHpgdeZ2mVt6ltg/e8JxSfv47ZB9Nne70Btt4eT8DQEr1NXoicg5Ciwi4nRp2WkM/344fyf9TUXfiizstpBg72Bnl1Uwlhw49vuFcHJ4C5z6O3c7r0DbtSfnl7Wv0kyjJyJXocAiIk6Vbc3m6R+e5tfjvxLoGcjCrgup7F/Z2WXlXerxc8Hk3MWxR7ZDdlrudhXqXrjuJLIVhNYBtzJwbY5IIVFgERGnsRpWnt34LD/G/4i32Zu5XeZSK7iWs8u6Mks2HP3twnUnh3+B04dyt/MKOhdOWtpO71RpAT7lirtakVJFgUVEnMIwDF7e+jJf/f0VZpOZWR1nud59gVKPXQgmh7farj3JSb+kkck2enL+upOIVrZrUTR6IlKoFFhExCne+e0d3t/zPgAz2s2gQ0QH5xZkyYbE3Y4Xx575J3c776ALwSSype06FO+g4q9XpIxRYBGRYrdi/wpmb58NwJgWY7i95u3FX0RKomM4ObIDcjIuaWSyrXNiX5StFZSvpdETESdQYBGRYvV97PdM2zwNgEcaPsJDDR4q+oPmZF0YPTl/cWxSbO523uUcV4yt0hy8A4u+PhG5JgUWESk2WxO3Mnb9WKyGlT61+jCq2aiiOVBywoX77cRtgYSduUdPTG5Qsf6FmwFGnhs9KWu3ABApIRRYRKRY7Du1jye+f4JMSyYdIzsyJWpK4dwfKCcLEn91XPckKS53O5+QC7N2zo+eeAVc//FFpFgosIhIkTuccpihq4eSkp1Cs4rNeKnDS7i7FfCvn+QjjuHkyE6wZDq2MblBxQbnwsm50zvla2r0RKQEU2ARkSJ1Iv0EQ2KGcCL9BLWDa/N6l9fxdvfOewdZabDzQzj0oy2gJMfnbuNb3vHUTuWmGj0RKWUUWESkyKRmpTJs9TBiU2Kp4l+FBV0XEOiZx4tYDQP2/g++fcbxFI/JDGENHC+ODblBoycipVyB5ubNmzePGjVq4O3tTfPmzdmwYUOe9tu4cSPu7u40adIk13vLly+nfv36eHl5Ub9+fT777LOClCYiLiLTksmTa59k76m9hHiHsLDbQir6Vszbzsf3wft94L8DbGElKBI6T4aHvoLxsTB0A9z2Ctx0j071iJQR+Q4sy5YtY9SoUUycOJEdO3bQvn17evXqRWzsZaYIXiQpKYkHH3yQLl265Hpv8+bN9O/fnwEDBrBr1y4GDBhAv379+Pnnn/Nbnoi4AIvVwoQNE/gl8Rd83X2Z13Ue1QKrXXvHjGT4diLMbwt/rwOzF3QYC8N/gQ5joEZ78PIv8vpFxPWYDMMw8rND69atadasGfPnz7dvq1evHn369CE6OvqK+91zzz3Url0bs9nM559/zs6dO+3v9e/fn+TkZL755hv7tp49exIcHMzSpUvzVFdycjJBQUEkJSURGKh1E0ScxTAMZvw0g0/+/AQPNw/md51P60qtr76T1Qq/LoPVUyD1qG1bnd7Q4z8QUqPoixYRp8nr93e+RliysrLYtm0b3bt3d9jevXt3Nm3adMX9Fi9ezIEDB5gyZcpl39+8eXOuPnv06HHVPjMzM0lOTnZ4iIjzzds1j0/+/AQTJma2n3ntsHJkJyzuCZ8PtYWVkJpw/3K4d6nCiojY5eui2xMnTmCxWAgLC3PYHhYWRmJi4mX32b9/P+PHj2fDhg24u1/+cImJifnqEyA6Oppp06blp3wRKWIf7f2IBbsWADCx9US6V+9+5cZpp2DNdNi2BDDAww9u+Te0GQbuXsVSr4iUHAW66PbSxZ4Mw7jsAlAWi4X77ruPadOmceONNxZKn+dNmDCBpKQk+yMu7jILRYlIsVl1cBUzf5kJwLAmw+hft//lG1otsOUteL0ZbFsMGNDobhi5FW5+SmFFRC4rXyMsoaGhmM3mXCMfx44dyzVCApCSksLWrVvZsWMHI0aMAMBqtWIYBu7u7nz33Xd07tyZ8PDwPPd5npeXF15e+otNxBVsOrKJCT9OwMDgnjr3MLTx0Ms3jP0JVo6x3dcHIKwh9HoRqrcrvmJFpETK1wiLp6cnzZs3JyYmxmF7TEwMbdu2zdU+MDCQ3bt3s3PnTvtj6NCh1KlTh507d9K6te3cdlRUVK4+v/vuu8v2KSKu5bcTvzFq7ShyrDn0qN6D8a3G5x4dTUmEFYPhnR62sOIdBL1fhsE/KKyISJ7ke+G40aNHM2DAAFq0aEFUVBSLFi0iNjaWoUNt/6KaMGEC8fHxvPfee7i5udGwYUOH/StWrIi3t7fD9ieffJIOHTrwwgsvcMcdd/DFF1+wevVqfvzxx+v8eCJSlA4mHWTY6mGk56TTplIb/nPzfzC7mS80yMmCnxfADy9AVipggmYPQpdnwS/UaXWLSMmT78DSv39/Tp48yfTp00lISKBhw4asXLmSatVsaywkJCRcc02WS7Vt25aPP/6YSZMmMXnyZGrWrMmyZcvsIzAi4nqOnj3KkJghnM48TYPyDZjdaTaeZs8LDf5aA9+Mg5P7ba+rtIDeL0GVZs4pWERKtHyvw+KqtA6LSPFJykxi4KqB/HXmL6oHVufdXu8S4h1ie/P0P7bl9P/4yvbarwJ0nQY33QtuBbrOX0RKsbx+f+teQiKSL+k56YxYM4K/zvxFRZ+KLOi2wBZWstNh42vw46uQk2G750/rIdBxvO2aFRGR66DAIiJ5lm3NZswPY9h5fCcBngEs6LaAKn6VL9yk8My508HV29tO/1Ss59yCRaTUUGARkTyxGlambprK+sPr8TJ7MbfLXGrnGPDBXXDge1ujwAjo8RzU76MbEopIoVJgEZE8eXXbq3x54EvMJjOz2j5H053L4af5YM0Bsye0fQLajwZPP2eXKiKlkAKLiFzT4t8Ws+T3JQBMi+zNLSuehNRziz3e2At6/gdCbnBegSJS6imwiMhVff7X57yy7RUAnrYEcscPc21vhNwAPV+AG69yvyARkUKiwCIiV7Qubh1TN00FYGBSCgNPxYKHL3T4N0QN131/RKTYKLCIyGVtT9zKmLWjsBgW/i8lladOnYaG/4JuMyCoirPLE5EyRoFFRHL5c89yRvwyjUyTwS1p6Ux1C8PtofegRntnlyYiZZQCi4hckHKU+O/GMfTMz6S4u9M0M4eXmozCo9VQMOuvCxFxHv0NJCJgyYafF3Jy/QsMCfXjuIcHtdz8eL3vR/ho9o+IuAAFFpGy7sBa+GYcZ0/+ybBKFfnHw4PK3uVZcNsygvzCnF2diAigwCJSdp2JhW8nwt4vyQKerFyFPV5mgr2CWdBzMWEKKyLiQhRYRMqa7HTYOOfcTQrTsZjMTKjTnJ8zE/F192V+1/nUCKrh7CpFRBwosIiUFYYB+1bCqglw5h/bpuo3E12tLt/Ffoe7mzuzO82mQWgDJxcqIpKbAotIWXBiP3wzDg6ssb0OrALdZ7Ag5yjLds3HhIno9tFEVY5ybp0iIlfg5uwCRKQIZaZAzLMwL8oWVsye0P5pGLGFZe45zNs1H4BnWj9Dz+o9nVysiMiVaYRFpDQyDNj9KcRMhpQE27baPaBnNJSvybeHvuX5n58HYOhNQ7mn7j1OLFZE5NoUWERKm8TdsHIsxG6yvQ6uAT1nQh3bCMpPCT8xYcMEDAz63diPYTcNc2KxIiJ5o8AiUlqknYK1/4Gtb4Nhtd2ksP3TEDUCPLwB+P3k7zz5/ZNkW7PpVq0bz7R+BpPJ5OTCRUSuTYFFpKSzWmDH+7B6GqSfsm1rcCd0fw6CIuzN/kn+h2Grh5GWk0br8NbMbD8Ts5vZSUWLiOSPAotISRa3BVaOgYSdttcV6kHvF6FGB4dmx9KOMSRmCKcyTlEvpB6zO83G0+xZ/PWKiBSQAotISZR6DFZPhZ0f2l57BUKnZ6DlIDB7ODRNykxiSMwQ4lPjqRpQlfld5+Pv6V/8NYuIXAcFFpGSxJINvyyCdTMhM9m2rckD0HUK+FfM1TwjJ4Mnvn+Cv878RahPKAu7LaS8T/liLlpE5PopsIiUFH//AN+MheN/2F5Xbgq9X4aIFpdtnmPN4d8//Jvtx7YT4BHAgq4LiAiIuGxbERFXp8Ai4urOxMF3E2HPF7bXvuWhyxRoOgDcLr/2o2EYTNs8jXWH1+Fl9uL1Lq9TJ6ROMRYtIlK4FFhEXFV2Bmx6HTbMgpx0MLlBy8eg0wTwCb7qrrO3z+bzvz7HbDLzUoeXaB7WvJiKFhEpGgosIq7GMODPVbBqPJw+ZNtWrR30ehHCG15z93d/f5d3fnsHgClRU+hUtVMRFisiUjwUWERcyckDtpsU/hVjex1QybaeSsN/QR4WePvywJe8vPVlAEY1G8Wdte8sympFRIqNAouIK8hMhQ0vw+a5YMkCNw9oOwLajwGvvE1BXn94Pc9ufBaAB+s/yCMNHynKikVEipUCi4gzGQb8thy+mwwpR2zbanWz3fsntFaeu9l5bCdPr3sai2Hhthtu4+kWT2vJfREpVRRYRJwl8Tfb6Z9/frS9Dq5uCyo39szT6Z/z/jr9F8PXDCfDksHNVW5mervpuJkuP3tIRKSkUmARKW7pp2FtNGx503aTQncf200K246036Qwr46kHmHI6iEkZyVzU4WbmHXLLDzcPK69o4hICaPAIlJcrFbbTQrXTIO0k7Zt9e+A7s9Duch8d3c64zRDYoZwLO0YNYNqMrfLXHw9fAu5aBER16DAIlIcDm+z3aTwyHbb6wp1odcLcEPHAnWXlp3GsNXDOJR8iHC/cBZ0W0CQV1Dh1Ssi4mIKdKJ73rx51KhRA29vb5o3b86GDRuu2PbHH3+kXbt2lC9fHh8fH+rWrcurr77q0GbJkiWYTKZcj4yMjIKUJ+I6Uo/DF8Phrc62sOIVCD3+A0N/LHBYybZkM2rtKH47+RvlvMqxsNtCwv3CC7duEREXk+8RlmXLljFq1CjmzZtHu3btWLhwIb169WLPnj1UrVo1V3s/Pz9GjBhB48aN8fPz48cff2TIkCH4+fkxePBge7vAwED27dvnsK+3d/7O54u4DEuO7RqVtdGQmWTb1uR+25L6AWEF7tZqWJn440Q2J2zGx92HeV3mcUPQDYVUtIiI6zIZhmHkZ4fWrVvTrFkz5s+fb99Wr149+vTpQ3R0dJ76uOuuu/Dz8+P9998HbCMso0aN4syZM/kpxUFycjJBQUEkJSURGBhY4H5ErtvBDbabFB7bY3tdqQn0fgkiW11Xt4ZhEP1LNEv/WIq7mztzO8+lbZW211+viIgT5fX7O1+nhLKysti2bRvdu3d32N69e3c2bdqUpz527NjBpk2buOWWWxy2p6amUq1aNSIiIrjtttvYsWPHVfvJzMwkOTnZ4SHiVEmH4ZOB8O5ttrDiEwK3zYbHvr/usAKw6NdFLP1jKQDPt3teYUVEypR8BZYTJ05gsVgIC3Mc0g4LCyMxMfGq+0ZERODl5UWLFi0YPnw4gwYNsr9Xt25dlixZwpdffsnSpUvx9vamXbt27N+//4r9RUdHExQUZH9ERuZ/loVIocjJhPUvwxst4ffPLtykcOQ2aPEwuJmv+xD/3fdf3tj5BgDjW42n9w29r7tPEZGSpECzhC5dQdMwjGuuqrlhwwZSU1P56aefGD9+PLVq1eLee+8FoE2bNrRp08betl27djRr1ozXX3+dOXPmXLa/CRMmMHr0aPvr5ORkhRYpfn9+a1v87fRB2+uqUbbTP+GNCu0QMf/E8PzPzwMwuPFg7q93f6H1LSJSUuQrsISGhmI2m3ONphw7dizXqMulatSoAUCjRo04evQoU6dOtQeWS7m5udGyZcurjrB4eXnh5eWVn/JFCs/JA7BqAuz/1vbaP9x2k8JGffO1Su21/JLwC+PWj8NqWOl7Y19GNBlRaH2LiJQk+Tol5OnpSfPmzYmJiXHYHhMTQ9u2eT+fbhgGmZmZV31/586dVKpUKT/liRS9rLOwZjrMa2MLK24e0O5JGLkVGt9dqGFl78m9PLH2CbKt2XSt2pVJrSfp/kAiUmbl+5TQ6NGjGTBgAC1atCAqKopFixYRGxvL0KFDAdupmvj4eN577z0A5s6dS9WqValbty5gW5fl5ZdfZuTIkfY+p02bRps2bahduzbJycnMmTOHnTt3Mnfu3ML4jCLXzzDg9xW2mxQmx9u21exiW/wttHahHy42OZahq4dyNvssLcNbMrPDTMyFcC2MiEhJle/A0r9/f06ePMn06dNJSEigYcOGrFy5kmrVqgGQkJBAbGysvb3VamXChAkcPHgQd3d3atasycyZMxkyZIi9zZkzZxg8eDCJiYkEBQXRtGlT1q9fT6tW1z+zQuS6Hd1jm6Z86NwCieWqQc9oqNO7UEdUzjuedpzBMYM5lXGKuiF1mdNpDl5mnf4UkbIt3+uwuCqtwyKFLv0MrJsJvywCwwLu3nDzaGj3BHj4FMkhk7OSeWTVI+w7vY/IgEje6/UeoT6hRXIsERFXkNfvb91LSORSVivs/BBWT4W0E7Zt9f4PejwP5XKv5lxYMnIyeOL7J9h3eh+hPqEs7LZQYUVE5BwFFpGLxW+Dlf+2/Rcg9Ebo9SLU7FSkh82x5jB2/Vi2Hd2Gv4c/C7ouIDJA0/RFRM5TYBEBOHvCNqKy4wPAAM8A6DgOWg0Bd88iPbRhGMz4aQZr49bi6ebJnM5zqBNSp0iPKSJS0iiwSNlmyYGtb8Pa5yHj3E0Kb7oXuk6FgOK5A/KcHXNYsX8FbiY3XrzlRVqGtyyW44qIlCQKLFJ2HfoRVo6FY7/bXoc3ht4vQ9XWxVbC+3ve563dbwHwbJtn6VK1S7EdW0SkJFFgkbInKR5iJsNvy22vfYKhy7PQ7KFCue9PXn3191e8uOVFAJ5s9iT/uvFfxXZsEZGSRoFFyo6cTNg813ajwuyztpsUNn8YOk8C35BiLWXD4Q1M/nEyAA/Ue4BHGz5arMcXESlpFFikbPjzO1g1Hk4dsL2ObAO9X4RKNxV7KbuO7+LpH54mx8ihd43e/Lvlv7XkvojINSiwSOlmyYYVj8Hvn9le+4dBtxnQuF+RrFJ7LQfOHGD4muGk56TTrko7nmv3HG6mfN3SS0SkTFJgkdJt1QRbWHFzhzaPQ4ex4O2clZATzyYyJGYISZlJNA5tzCu3vIKH2cMptYiIlDQKLFJ6bVsCW94ETNDvfajb22mlnMk4w+CYwRxNO8oNQTcwt8tcfD18nVaPiEhJo7FoKZ3+2QRfj7E97zzRqWElLTuN4WuGczDpIGG+YSzstpBy3uWcVo+ISEmkwCKlz5lYWDYArNnQ4E5oP8ZppWRbshm9bjS/nviVIK8gFnZbSLhf8SxIJyJSmiiwSOmSdRY+vs9208LwRnDHXKdcXAtgNaxM2jiJjUc24uPuw9wuc6lZrqZTahERKekUWKT0MAz4fBgk7gbfULhnKXj6OakUg5e2vMTKgytxN7nzSsdXuKlC8U+hFhEpLRRYpPTY8DLs+RzcPKD/B1DOeXc7fvu3t/lg7wcAzLh5BjdXudlptYiIlAYKLFI6/LESvn/O9vzWl6FalNNKWf7ncl7b/hoA41qO47YbbnNaLSIipYUCi5R8x/baFocDaPkYNB/otFLW/LOG6T9NB2BQo0E8UP8Bp9UiIlKaKLBIyZZ2CpbeC1mpUL099Ix2WilbErcwdv1YrIaVu2rfxRNNn3BaLSIipY0Ci5Rclhz4ZCCcPgjlqsLd74KTVo7949QfPPH9E2RZs+gc2ZnJbSbr/kAiIoVIgUVKru8mwsEfwMMP7v0Y/Mo7pYy4lDiGxgwlNTuV5mHNeaHDC7i7aRFpEZHCpMAiJdP29+HnBbbndy2EsAZOKeNE+gmGxAzhZMZJ6gTX4fXOr+Pt7u2UWkRESjMFFil5Yn+Gr56yPe84Aerd7pQyUrJSeHz148SlxFHFvwrzu84nwDPAKbWIiJR2CixSsiQdhmUP2Jbdr/d/trsvO0GmJZMn1z7JH6f+IMQ7hEXdFlHBt4JTahERKQsUWKTkyE6Hj++Hs8cgrCH0mQ9uxf9H2GK1MG79OLYkbsHPw48FXRdQNbBqsdchIlKWKLBIyWAY8MUISNgJvuXhno/Ay98JZRjM+GkGa2LX4OHmwZxOc6hXvl6x1yEiUtYosEjJsHE2/PYpuLlDv/cguJpTynhj5xss378cN5MbL3Z4kVaVWjmlDhGRskaBRVzfn9/C6mm2571egOrOuS/Ph3s/ZNGviwCY1GYSXat1dUodIiJlkQKLuLbj++DTRwEDmj8MLQc5pYyVf69k5i8zARjRZAR333i3U+oQESmrFFjEdaWfhqX3QFYKVG0LvV50Shkb4zcy8ceJANxX9z4GNx7slDpERMoyBRZxTZYc+PQROPU3BEXarltx9yz2MnYf381T654ix8ihV/VejGs1Tkvui4g4gQKLuKbVU+DA9+Dha5sR5F/8a5z8nfQ3w9YMIz0nnbaV2/L8zc/jZtL/MiIizqC/fcX17FwKm9+wPe8zDyo1LvYSEs8mMiRmCGcyz9AotBGvdnwVDyfdWFFERBRYxNUc3gr/e9L2vMNYaHBnsZeQlJnE0JihJJ5NpHpgdeZ2mYuvh2+x1yEiIhcosIjrSE6wrWRryYS6t9nuE1TM0rLTGL5mOAeSDlDRtyILuy0k2Du42OsQERFHBQos8+bNo0aNGnh7e9O8eXM2bNhwxbY//vgj7dq1o3z58vj4+FC3bl1effXVXO2WL19O/fr18fLyon79+nz22WcFKU1Kqux0+Pg+SE2ECvXgzgXFvux+tjWbp394ml3HdxHoGcjCrgup7F+5WGsQEZHLy/c3wrJlyxg1ahQTJ05kx44dtG/fnl69ehEbG3vZ9n5+fowYMYL169ezd+9eJk2axKRJk1i0aJG9zebNm+nfvz8DBgxg165dDBgwgH79+vHzzz8X/JNJyWEYttNAR7aDTzDcuxS8iveux1bDyrMbn+XH+B/xNnszt8tcagXXKtYaRETkykyGYRj52aF169Y0a9aM+fPn27fVq1ePPn36EB0dnac+7rrrLvz8/Hj//fcB6N+/P8nJyXzzzTf2Nj179iQ4OJilS5fmqc/k5GSCgoJISkoiMDAwH59InG7jHIiZDCYzDPgMbrilWA9vGAYvbX2J9/e8j9lkZk7nOXSI6FCsNYiIlFV5/f7O1whLVlYW27Zto3v37g7bu3fvzqZNm/LUx44dO9i0aRO33HLhS2nz5s25+uzRo8dV+8zMzCQ5OdnhISXQ/tW2KcwAPaOLPawAvPPbO7y/xxaeZ7SbobAiIuKC8hVYTpw4gcViISwszGF7WFgYiYmJV903IiICLy8vWrRowfDhwxk06MIS64mJifnuMzo6mqCgIPsjMjIyPx9FXMGJ/bbF4QwrNB0ArYp/BdnP9n/G7O2zARjTYgy317y92GsQEZFrK9BVjZeu9GkYxjVX/9ywYQNbt25lwYIFzJ49O9epnvz2OWHCBJKSkuyPuLi4fH4KcaqMJFh6L2QmQWQbuHUWFPMKsmtj1zJ181QAHmn4CA81eKhYjy8iInnnnp/GoaGhmM3mXCMfx44dyzVCcqkaNWoA0KhRI44ePcrUqVO59957AQgPD893n15eXnh5eeWnfHEVVovthoYn90NgBPR/H9yL93e57eg2/r3+31gNK31q9WFUs1HFenwREcmffI2weHp60rx5c2JiYhy2x8TE0LZt2zz3YxgGmZmZ9tdRUVG5+vzuu+/y1aeUIGumwV8x4O4D93wI/hWL9fD7Tu1j5JqRZFoy6RjZkSlRU3R/IBERF5evERaA0aNHM2DAAFq0aEFUVBSLFi0iNjaWoUOHArZTNfHx8bz33nsAzJ07l6pVq1K3bl3Ati7Lyy+/zMiRI+19Pvnkk3To0IEXXniBO+64gy+++ILVq1fz448/FsZnFFfy639h42u2533mQuUmxXr4wymHGbp6KCnZKTSr2IyXOryEu1u+/zcQEZFilu+/qfv378/JkyeZPn06CQkJNGzYkJUrV1KtWjUAEhISHNZksVqtTJgwgYMHD+Lu7k7NmjWZOXMmQ4YMsbdp27YtH3/8MZMmTWLy5MnUrFmTZcuW0bp160L4iOIy4rfBFyNsz28eDQ3/VayHP5l+kiExQziRfoLawbV5vcvreLt7F2sNIiJSMPleh8VVaR0WF5eSCIs6QkoC3NgT7llarCvZpmal8si3j7D31F6q+FfhvV7vUdG3eE9FiYhIbkWyDotIgWRnwLIHbGEltA7c9WaxhpUsSxaj1o5i76m9hHiHsLDbQoUVEZESRoFFipZhwNej4fAW8A6yLbvvXXwjYNmWbMatH8fPiT/j6+7LvK7zqBZYrdiOLyIihUNXG0rR+mk+7PwQTG5w9xIoX7PYDp2WncZT655i05FNeLh5MKfzHBqUb1BsxxcRkcKjwCJF58D38N1E2/Puz0PNzsV26FMZpxi+eji/nfwNH3cfXun4Cq0r6SJuEZGSSoFFisbJA/DJw7Zl95vcD20eL7ZDx6fGMzRmKIeSD1HOqxxzu8ylcYXGxXZ8EREpfAosUvgykm3L7mecgYiWcNurxbbs/r5T+3h89eMcTz9OJb9KLOi2gBuCbiiWY4uISNFRYJHCZbXAisfgxD4IqAz9Pyi2Zfe3Hd3GyDUjSclOoVa5WizouoAwv6vfMkJEREoGzRKSwvX9c/DnKjB7wT0fQEB4sRx2TewaBn832L6C7ZKeSxRWRERKEQUWKTy7P4UfX7E9v+MNqNK8WA67/M/ljF43mixrFh0jO7Kw20KCvIKK5dgiIlI8dEpICseRnReW3W/7BDTuV+SHNAyDN3e/yes7Xgfgrtp3MbnNZN0bSESkFNLf7HL9Uo/Bx/dBTjrU6gZdpxb5Ia2GlZm/zGTpH0sBeKzRY4xsOlJ3XRYRKaUUWOT65GTCsgGQHA/la0Pft8HNXKSHzLJk8cyPz/DtoW8xYWJcq3HcX+/+Ij2miIg4lwKLFJxhwNdPQ9xP4BUE935sW36/CJ3NPsuTa5/k54SfcXdz5z83/4deNXoV6TFFRMT5FFik4H5ZBDvety273/cdCK1VpIc7mX6Sx1c/zt5Te/F192V2p9lEVY4q0mOKiIhrUGCRgvl7HayaYHvedRrU7lqkh4tLiWNozFBiU2IJ8Q5hXpd5NAjVfYFERMoKBRbJv1MH4ZOBYFigcX9oO7JID/fHqT94fPXjnEg/QRX/KizstlB3XBYRKWMUWCR/MlNsy+6nn4bKzeD214p02f0tiVt44vsnSM1O5cbgG1nQdQEVfCsU2fFERMQ1aeE4yTurFVYMgeN7wT8c7vkQPHyK7HAx/8QwJGYIqdmptAhrwZKeSxRWRETKKAUWybt10bDv63PL7n8IgZWL7FD/3fdfnl73NNnWbLpW7cqCbgsI8AwosuOJiIhr0ykhyZvfP4P1L9qe3/4aRLQoksMYhsGCXQuYt2seAH1v7Muk1pMwF/HaLiIi4toUWOTaEn6Fz4fZnkeNgCb3FslhLFYL0b9Es2zfMgCG3jSUYTcN0+q1IiKiwCLXkHrctux+dhrU7GybwlwEMi2ZTNgwgZh/YjBh4pnWz3BP3XuK5FgiIlLyKLDIleVkwX8fhKQ4CKlpWxzOXPh/ZFKyUnhy7ZNsSdyCh5sHM9vPpHv17oV+HBERKbkUWOTKvhkLsZvAMwDuXQo+wYV+iBPpJ3h89eP8ceoP/Dz8eK3Ta7Su1LrQjyMiIiWbAotc3pa3YNtiwGS7oWGFOoV+iNjkWAbHDCY+NZ7y3uWZ33U+9crXK/TjiIhIyafAIrkd+hG+GWd73uVZuLFHoR/i95O/M2z1ME5lnCLCP4JF3RYRGRhZ6McREZHSQYFFHJ3+x3bdijUHGvaFm58q9ENsPrKZUWtHkZaTRr2QeszrOo9Qn9BCP46IiJQeWjhOLshMtc0ISjsJlZrAHW8U+rL7qw6tYtiaYaTlpNE6vDXv9HhHYUVERK5JIyxiY7XC54/D0d/AryLc81GhL7v/0d6PmPnLTAwMulXrxsz2M/E0exbqMUREpHRSYBGb9S/C3i/B7An9P4CgKoXWtWEYvLHzDRb9ugiAe+rcw/hW47V6rYiI5JkCi8CeL233CQK49RWoWnjTinOsOTz303Ms378cgBFNRjC48WCtXisiIvmiwFLWHf0dPhtqe956KDQbUGhdZ+RkMG79OL6P+x43kxuT20ym7419C61/EREpOxRYyrKzJ2HpPZB9FmrcAt2fL7Suk7OSGblmJNuPbcfTzZMXO7xIl2pdCq1/EREpWxRYyipLNnzyEJyJheDqcPeSQlt2/1jaMYauHsr+0/sJ8AhgTuc5tAgvmrs7i4hI2aDAUlatmgCHNoCnP9z7MfiGFEq3h5IOMSRmCEfOHiHUJ5QFXRdQJ6TwV8kVEZGypUDrsMybN48aNWrg7e1N8+bN2bBhwxXbrlixgm7dulGhQgUCAwOJiori22+/dWizZMkSTCZTrkdGRkZBypNr2bYEtrwJmOCuN6Fi4SyH/9uJ33jwmwc5cvYI1QKr8X6v9xVWRESkUOQ7sCxbtoxRo0YxceJEduzYQfv27enVqxexsbGXbb9+/Xq6devGypUr2bZtG506deL2229nx44dDu0CAwNJSEhweHh7exfsU8mV/bMJvh5je955ItTtXSjdborfxCPfPsLpzNM0KN+A93q9R0RARKH0LSIiYjIMw8jPDq1bt6ZZs2bMnz/fvq1evXr06dOH6OjoPPXRoEED+vfvz7PPPgvYRlhGjRrFmTNn8lOKg+TkZIKCgkhKSiIwMLDA/ZRqZ2JhUSdIOwEN7oS+iwtlJduv//6aST9OIsfIIapSFK92ehU/D79CKFhEREq7vH5/52uEJSsri23bttG9e3eH7d27d2fTpk156sNqtZKSkkJIiOM1E6mpqVSrVo2IiAhuu+22XCMwcp2yzp5bdv8EhDeCO+YWSlj5YM8HjN8wnhwjh141ejG3y1yFFRERKXT5uuj2xIkTWCwWwsLCHLaHhYWRmJiYpz5mzZrF2bNn6devn31b3bp1WbJkCY0aNSI5OZnXXnuNdu3asWvXLmrXrn3ZfjIzM8nMzLS/Tk5Ozs9HKVsMAz4fBom7wTcU7lkKntcXKgzD4LXtr/H2b28DcH+9+xnbcixuJt2eSkRECl+BZgldukqpYRh5Wrl06dKlTJ06lS+++IKKFSvat7dp04Y2bdrYX7dr145mzZrx+uuvM2fOnMv2FR0dzbRp0wpSftmz4WXY8zm4eUD/96Fc5HV1l2PNYdrmaXz+1+cAPNnsSR5t+KhWrxURkSKTr38Oh4aGYjabc42mHDt2LNeoy6WWLVvGo48+yn//+1+6du169aLc3GjZsiX79++/YpsJEyaQlJRkf8TFxeX9g5Qlf6yE75+zPe/9ElRre13dpeekM2rtKD7/63PcTG5MbzudQY0GKayIiEiRyldg8fT0pHnz5sTExDhsj4mJoW3bK38RLl26lIEDB/LRRx9x6623XvM4hmGwc+dOKlWqdMU2Xl5eBAYGOjzkEsf2worHbM9bDoIWD19Xd0mZSQyJGcIPh3/Ay+zF7I6zubP2nYVQqIiIyNXl+5TQ6NGjGTBgAC1atCAqKopFixYRGxvL0KG2+9FMmDCB+Ph43nvvPcAWVh588EFee+012rRpYx+d8fHxISgoCIBp06bRpk0bateuTXJyMnPmzGHnzp3MnTu3sD5n2ZN2CpbeC1mpUL099Jx5Xd0lnk3k8dWP89eZvwjwDOCNzm/QLKxZIRUrIiJydfkOLP379+fkyZNMnz6dhIQEGjZsyMqVK6lWrRoACQkJDmuyLFy4kJycHIYPH87w4cPt2x966CGWLFkCwJkzZxg8eDCJiYkEBQXRtGlT1q9fT6tWra7z45VRlhz4ZCCcPgjlqsLd74LZo8Dd/X3mb4asHkLi2UQq+lRkQbcF1A6+/MXQIiIiRSHf67C4Kq3DcpFvxsHPC8DDDwbFQFiDAnf16/FfGbZmGEmZSVQPrM7Cbgup7F+5EIsVEZGyrEjWYZESYPv7trACcNfC6worGw5vYNB3g0jKTKJRaCPe6/WewoqIiDiFAktpEvszfPWU7XnHCVDv9gJ39eWBLxn5/UjSc9JpV6Udb3V/i2Dv4EIqVEREJH90t+bSIukwLHsArNlQ7/+gw9gCd7XktyXM2jYLgNtuuI3p7abj4Vbwa2BERESulwJLaZCdDh/fD2ePQVhD6DMf3PI/eGY1rLyy9RXe3fMuAA/Wf5CnWzyt1WtFRMTpFFhKOsOAL0ZAwk7wCYF7PgQv/3x3k23NZsrGKfzv7/8BMLr5aB5ueH3rtoiIiBQWBZaSbuNs+O1TcHOHfu9BcPV8d5GWncbTPzzNj/E/YjaZmd5uOv9X8/8KvVQREZGCUmApyf78Flafu59SrxegRvt8d3Em4wzD1wzn1xO/4m32ZlbHWXSI6FDIhYqIiFwfBZaS6vg++PRRwIDmD9uW3s+nhNQEhqwewsGkgwR5BTG3y1xuqnBT4dcqIiJynRRYSqL007D0HshKgaptodeL+e7ir9N/MWT1EI6lHSPcL5yFXRdyQ7kbiqBYERGR66fAUtJYcuDTR+DU3xAUabtuxd0zX13sOLaD4WuGk5KVQs2gmizotoBwv/AiKlhEROT6ab5qSbN6Chz4Hjx84Z6PwL9CvnZfF7eOx757jJSsFJpUaMK7vd5VWBEREZenwFKS7FwKm9+wPe8zDyo1ztfun+3/jFFrR5FpyaRDRAcWdV9EkFdQERQqIiJSuHRKqKQ4vBX+96TteYd/Q4M787yrYRi8/dvbvLb9NQDuqHkHU9pO0eq1IiJSYiiwlATJR2wr2Voyoc6t0PGZPO9qNay8tOUlPtj7AQCPNHyEUc1GYTKZiqpaERGRQqfA4urOL7ufmggV6tnuwJzHZfezLdlM2jiJlQdXAvDvFv/mwQYPFmW1IiIiRUKBxZUZhu000JHt4BMM9y4Fr4A87ZqWncZT655i05FNuJvcmXHzDG674bYiLlhERKRoKLC4sk2vw6/LwGSGu9+FkBp52u1UximGrx7Obyd/w8fdh1c7vkq7Ku2KuFgREZGio8Diqvavtk1hBugZDTfckqfd4lPjGRozlEPJhyjnVY55XebRqEKjIixURESk6CmwuKIT+22LwxlWaDoAWg3O0277Tu3j8dWPczz9OJX8KrGw20JqBOVtVEZERMSVKbC4mowkWHovZCZBZGu4dRbkYUbPtqPbGLlmJCnZKdQqV4sFXRcQ5hdWDAWLiIgUPS0c50qsFtsNDU/uh8Aq0P8DcPe65m5rYtcw+LvBpGSn0KxiM5b0XKKwIiIipYoCiytZMw3+igF3n3PL7le85i7L/1zO6HWjybJm0TGyIwu7LdTqtSIiUurolJCr2LUMNtpWoqXPXKjc5KrNDcPgzd1v8vqO1wG4q/ZdTG4zGXc3/UpFRKT00bebK4jfBl+OtD2/eTQ0/NdVm1sNKzN/mcnSP5YC8FijxxjZdKRWrxURkVJLgcXZUhIvLLt/Y0/oPPmqzbMsWTzz4zN8e+hbTJgY12oc99e7v5iKFRERcQ4FFmfKzoBlD0BKAoTWgbvevOqy+6lZqYxaO4qfE3/G3c2d/9z8H3rV6FWMBYuIiDiHAouzGAZ8PRoObwHvINuy+96BV2x+Iv0Ew1YPY++pvfi6+zK702yiKkcVY8EiIiLOo8DiLD/Nh50fgskN+i6G8jWv2DQuJY4hMUOIS4kjxDuEeV3m0SC0QTEWKyIi4lwKLM5w4Hv4bqLteffnoFaXKzb949QfDI0ZysmMk1Txr8LCbgupFlitmAoVERFxDQosxe3kAfjkYduy+03uhzbDrtj0l4RfeGLtE5zNPkud4DrM7zqfCr4VirFYERER16CF44pTRrJt2f2MMxDREm579YrL7n936DuGrh7K2eyztAhrweKeixVWRESkzFJgKS5WC6x4DE7sg4DKV112/7/7/suYH8aQbc2ma9WuLOi2gADPgGIuWERExHXolFBx+f45+HMVmL3gng8gIDxXE8MwmL9rPvN3zQfg7hvvZmLriZjdzMVdrYiIiEtRYCkOuz+FH1+xPb/jDajSPFcTi9XCf37+D//9878APH7T4zx+0+NavVZERAQFlqJ3ZCd8McL2vO0T0LhfriaZlkzGrx/P6tjVmDAxsfVE+tftX7x1ioiIuLACXcMyb948atSogbe3N82bN2fDhg1XbLtixQq6detGhQoVCAwMJCoqim+//TZXu+XLl1O/fn28vLyoX78+n332WUFKcy2px+Dj+yAnHWp1g65TczVJyUrh8dWPszp2NR5uHrx8y8sKKyIiIpfId2BZtmwZo0aNYuLEiezYsYP27dvTq1cvYmNjL9t+/fr1dOvWjZUrV7Jt2zY6derE7bffzo4dO+xtNm/eTP/+/RkwYAC7du1iwIAB9OvXj59//rngn8zZcjJh2QBIjofyteFfb8El16IcTzvOw6seZkviFvw8/FjQdQHdq3d3UsEiIiKuy2QYhpGfHVq3bk2zZs2YP3++fVu9evXo06cP0dHReeqjQYMG9O/fn2effRaA/v37k5yczDfffGNv07NnT4KDg1m6dGme+kxOTiYoKIikpCQCA6+8xH2xMAzb3Zd3vA9eQfDYGgit7dAkNjmWwTGDiU+Np7x3eeZ3nU+98vWcVLCIiIhz5PX7O18jLFlZWWzbto3u3R1HAbp3786mTZvy1IfVaiUlJYWQkBD7ts2bN+fqs0ePHnnu0+X8ssgWVkxu0PedXGHl95O/M+CbAcSnxhMZEMn7vd5XWBEREbmKfF10e+LECSwWC2FhYQ7bw8LCSExMzFMfs2bN4uzZs/Trd+Hi08TExHz3mZmZSWZmpv11cnJyno5f5P5eB6sm2J53nQa1uzq8vfnIZkatHUVaThr1Quoxr+s8Qn1Ci79OERGREqRAF91eOtXWMIw8Tb9dunQpU6dOZdmyZVSsWPG6+oyOjiYoKMj+iIyMzMcnKCKnDsInA8GwQOP+0Hakw9urDq1i2JphpOWk0Tq8Ne/0eEdhRUREJA/yFVhCQ0Mxm825Rj6OHTuWa4TkUsuWLePRRx/lv//9L127Oo46hIeH57vPCRMmkJSUZH/ExcXl56MUvswU27L76aehcjO4/TWHZfc/2vsRY38YS441h+7VujOv6zz8Pf2dWLCIiEjJka/A4unpSfPmzYmJiXHYHhMTQ9u2ba+439KlSxk4cCAfffQRt956a673o6KicvX53XffXbVPLy8vAgMDHR5OY7XCiiFwfC/4h8M9H4KHD2AbKXp9x+tE/xKNgcE9de7hxQ4v4mn2dF69IiIiJUy+F44bPXo0AwYMoEWLFkRFRbFo0SJiY2MZOnQoYBv5iI+P57333gNsYeXBBx/ktddeo02bNvaRFB8fH4KCggB48skn6dChAy+88AJ33HEHX3zxBatXr+bHH38srM9ZtNZFw76vwexpu0dQYGUAcqw5PPfTcyzfvxyAEU1GMLjxYK1eKyIikk/5Diz9+/fn5MmTTJ8+nYSEBBo2bMjKlSupVq0aAAkJCQ5rsixcuJCcnByGDx/O8OHD7dsfeughlixZAkDbtm35+OOPmTRpEpMnT6ZmzZosW7aM1q1bX+fHKwa/fwbrX7Q9v/01iGwJQEZOBuPWj+P7uO9xM7kxuc1k+t7Y14mFlg5Wq8HZrBxSMs4/sm3/zbzouf2/F56bTBDo7UGgjweB3h4E+XgQ6ON+0XPb66Bz7/t6mhUsRURcSL7XYXFVTlmHJeFXeKcHZKdB1Ajo8bytlqxkRq4ZyfZj2/F08+TFDi/SpVqX4qnJheVYrJzNtJB8abDIvDhgXCV4ZOaQmplDcfyJdXcznQs37gT6eNiDTKCPuz30XNieu42Xu25YKSKSF3n9/ta9hAoq9bht2f3sNKjZ2TaFGTiWdoyhq4ey//R+AjwCmNN5Di3CWzi52OuXlWN1DBCZuUcxHEc7cm9Py7IUWj0eZhMB3h4EeLvbHl625/7etlET+/Zzzw0DkjOySUrPJjk956Ln2SRn5Nj+m27blmM1yLEanDqbxamzWQWqz9vD7ZqhJihX8LFtD/D2wOym0R0RkYspsBREThb890FIioOQmrbF4czuHEw6yNCYoRw5e4RQn1AWdF1AnZA6Ti3VMAwyc6wXjWpcCBCpGTmX3Z6SmU3quW3J57Zn5lgLrSZvDzf8vWxf4heHiouf+3tdHDxyv+/l7lYkp2wMwyA923KZUJNNUtqFcJN0btul7VLOjQBlZFvJyM7kWErmtQ96GQFe50Zy8hp2Ljq9pdNZIlIaKbAUxDdjIXYTeAbAvUvBJ5jdx3czbM0wzmSeoVpgNRZ0XUBEQMR1HcYwDM5mWc6Fh2x7eLg4YKRmng8Wlz+NkpqZQ7al8M6h+HmaHQKE/7nngefDhNeFYOF/rs3FwcPfyx1P9wIt/1MsTCYTvp7u+Hq6Ex7kne/9rVaDlMycCyHnohGd5ItGdByC0EVtzo9CpWTaRqriz6Tnuwazm4lAb/cL1+Zccs3OxUHocm10OktEXJECS35teQu2LQZM0PdtqFCHjfEbeWrdU6TnpNOgfAPmdZ1HkGcwSenZF0YzLrooNPmS0yWp9oDhGDxSM3OwFlLWMJm4ZNTCcTQj4NLgcZnRDn9vd52quAY3NxNB50Y9CuL8qberhZqkS4JPykVtsi0GFqvB6bRsTqdlF6gGL3e3Ap3KCvLx0OmsEsh67hSoxWqQbbVisVx4nWO12rZbHF/nWA1yLI6vLede2/e172PbbhhQzteD8n5ehPh5Ut7fk2BfT5f+B4y4Fl10ew37j6ZwIjWLlIxsvOM3c/PmQbgZOayNGMZ3IffxV9oP/JHzFpgseGbXxePEQFLT3UnNzCm0GtzdTPbrM85fqxHg7XhKxf+SEHJp8PD1MOOmL5JSzTAMMrKtuU9lnQ87Vwo/F53aKoy/Dfy9zoeXvISdi9r4eODn5NNZ1vNf2g5fwue+iC/6ArZ9iVsdXtu/vO1f+NaL2jq+zrn4y91ytcBwbr98hYjLH+uyn8lqFMtF7FcT4O1OeT9PyvufCzJ+noSce5T39yTEz+vc+7ZtGgEsfXTRbSEZ/d9d7I5PIsJ0nC88J+FmyuELS1ue/KsdHsEf4R3+FZggO+kmUo7cfW6vC2HF093NITz4e1163cYlweOi98/v5+1RNNdrSOliMpnw8TTj42kmLLBgp7NSsy66RucqFydfLvicP52Vem42V0GcP5116WmqAC8P3NzI9a/2S0NETq4v9/x94Tv7y9uVuLuZMLuZ8DC7YXYz2V+7u5kwm014uNm2m91MuJtNmN3c8Ljk9YU+TBgGnE7Lsl/MfupsFlYD++jyoZNpearL38v9QqCxBxuvC0HH/6Ltfl74eCrglBYKLNdQtbwv1swU3kifTXlLCrFetdlScyotjf/xR8ZXALQOuYO7mg2jnI/nJaMduh5ASg43N5Nt9MPbg4jg/O+fbbGSko9TWcn20Z3CO51VVDzM57543dwwmy/+8nY79+V8fpvjl/T5Nuffdzc7vjZfbpv5Ql8el7y+0IfjsWzbHV+bL912mX6uVJ+bKff93Qqb1WpwJj2bU2czOZlqCzAnLwozJ89mcTI10/789NkscqyGPRDHnspbwPH1NF8ycuNlH60pf8koToifpy5ad2E6JXQtVit88hDs/RL8KpIzaDVTf3+TLw58AcCTzZ7k0YaP6g+4yHWwz2a7ZAZW0kWhBnD8kj735Xrxa3uIcPiivxAsLn59pVBx8eiBu5tJp1JdhGEYJKfncPKsLcScSD0fbjIdg05qlr1NQSYceLm7Xf4UlX3kxrY99Fzo8fdy19//10mnhArL+hdtYcXNg/S+b/Pv7S/xw+EfMJvMTImawp2173R2hSIlnslkwtvDjLeHmYoFOJ0lpZ/JZCLI14MgXw9uqHDt9oZhm7F3KvXikZvMi4LO+e2Z9jaZOVYyc6wcScrgSFJGnuryNLtdcs2N7VTU+ecOp678vAj0UcApKAWWq8k6C7uWApDU8z+M2PsmO4/vxMvsxcu3vEzHyI7OrU9ERC7LZLpwirN6qN81259fRuLURSM0F4/cnDh3eur8KM6ps1mkZ1vIslhJTM4gMTlvAcfDbCLYN/dFxedfXzyKU97PkyAfD43ynaNTQtdy9iSJu95j6LEfOJB0gADPAOZ2mUvTik0L7xgiIlLipGdZOJnrGpxzp6jObTtx0SjO2QKs9m12MxF80XRwx4uKL1yTc35bOV/PEre0gE4JFZK/s5MYcmQliWcTqehbkYVdF1IruJazyxIRESfz8TQT4elLRLBvntpnZFuueFHxhVNXF7alZORgsRqcSLVds5MXJhP2EZyLr7VxGMmxBx8vgn09cDeXjLVwFFiuIi07jUe+fYSTGSepHlidRd0WUcm/krPLEhGREsjbw0zlcj5ULueTp/aZORZOn822n6K6+HTUxSM75wNOUrptLaX83getnK9H7plUl1yTc/EpK2cFHAWWq/D18OXpFk/z8b6PeaPzGwR7F2Cup4iISAF4uZsJDzLn+TYh2RYrp9OyHE9RpTpej3PxLKoz5wLOmbRszqRl8/fxs9c8xkeDWtO2Vuj1frQCUWC5httr3k7vGr0xu2k9FRERcV0eZjcqBnhTMSBvASfHYj23Fo5jkHEcubmw7XRaFuX9vYr4U1yZAkseKKyIiEhp4252I9Tfi1B/Lwi7dntrYd3croAUWEREROSanD29umRcGiwiIiJlmgKLiIiIuDwFFhEREXF5CiwiIiLi8hRYRERExOUpsIiIiIjLU2ARERERl6fAIiIiIi5PgUVERERcngKLiIiIuDwFFhEREXF5CiwiIiLi8hRYRERExOWVmrs1G4btttfJyclOrkRERETy6vz39vnv8SspNYElJSUFgMjISCdXIiIiIvmVkpJCUFDQFd83GdeKNCWE1WrlyJEjBAQEYDKZCq3f5ORkIiMjiYuLIzAwsND6leKj32HJp99hyabfX8lXlL9DwzBISUmhcuXKuLld+UqVUjPC4ubmRkRERJH1HxgYqP/RSjj9Dks+/Q5LNv3+Sr6i+h1ebWTlPF10KyIiIi5PgUVERERcngLLNXh5eTFlyhS8vLycXYoUkH6HJZ9+hyWbfn8lnyv8DkvNRbciIiJSemmERURERFyeAouIiIi4PAUWERERcXkKLCIiIuLyFFiuYd68edSoUQNvb2+aN2/Ohg0bnF2S5NH69eu5/fbbqVy5MiaTic8//9zZJUk+REdH07JlSwICAqhYsSJ9+vRh3759zi5L8mH+/Pk0btzYvthYVFQU33zzjbPLkgKKjo7GZDIxatQopxxfgeUqli1bxqhRo5g4cSI7duygffv29OrVi9jYWGeXJnlw9uxZbrrpJt544w1nlyIF8MMPPzB8+HB++uknYmJiyMnJoXv37pw9e9bZpUkeRUREMHPmTLZu3crWrVvp3Lkzd9xxB7///ruzS5N82rJlC4sWLaJx48ZOq0HTmq+idevWNGvWjPnz59u31atXjz59+hAdHe3EyiS/TCYTn332GX369HF2KVJAx48fp2LFivzwww906NDB2eVIAYWEhPDSSy/x6KOPOrsUyaPU1FSaNWvGvHnzeO6552jSpAmzZ88u9jo0wnIFWVlZbNu2je7duzts7969O5s2bXJSVSJlV1JSEmD7wpOSx2Kx8PHHH3P27FmioqKcXY7kw/Dhw7n11lvp2rWrU+soNTc/LGwnTpzAYrEQFhbmsD0sLIzExEQnVSVSNhmGwejRo7n55ptp2LChs8uRfNi9ezdRUVFkZGTg7+/PZ599Rv369Z1dluTRxx9/zPbt29myZYuzS1FguRaTyeTw2jCMXNtEpGiNGDGCX3/9lR9//NHZpUg+1alTh507d3LmzBmWL1/OQw89xA8//KDQUgLExcXx5JNP8t133+Ht7e3schRYriQ0NBSz2ZxrNOXYsWO5Rl1EpOiMHDmSL7/8kvXr1xMREeHsciSfPD09qVWrFgAtWrRgy5YtvPbaayxcuNDJlcm1bNu2jWPHjtG8eXP7NovFwvr163njjTfIzMzEbDYXWz26huUKPD09ad68OTExMQ7bY2JiaNu2rZOqEik7DMNgxIgRrFixgu+//54aNWo4uyQpBIZhkJmZ6ewyJA+6dOnC7t272blzp/3RokUL7r//fnbu3FmsYQU0wnJVo0ePZsCAAbRo0YKoqCgWLVpEbGwsQ4cOdXZpkgepqan89ddf9tcHDx5k586dhISEULVqVSdWJnkxfPhwPvroI7744gsCAgLso51BQUH4+Pg4uTrJi2eeeYZevXoRGRlJSkoKH3/8MevWrWPVqlXOLk3yICAgINc1Y35+fpQvX94p15IpsFxF//79OXnyJNOnTychIYGGDRuycuVKqlWr5uzSJA+2bt1Kp06d7K9Hjx4NwEMPPcSSJUucVJXk1fnlBDp27OiwffHixQwcOLD4C5J8O3r0KAMGDCAhIYGgoCAaN27MqlWr6Natm7NLkxJI67CIiIiIy9M1LCIiIuLyFFhERETE5SmwiIiIiMtTYBERERGXp8AiIiIiLk+BRURERFyeAouIiIi4PAUWESkxBg4cSJ8+fZxdhog4gRaOExEHAwcO5N133821vUePHk5fUj0pKQnDMChXrpxT6xCR4qel+UUkl549e7J48WKHbV5eXk6qxnaHWJPJRFBQkNNqEBHn0ikhEcnFy8uL8PBwh0dwcDDr1q3D09OTDRs22NvOmjWL0NBQEhISANu9f0aMGMGIESMoV64c5cuXZ9KkSVw8mJuVlcXYsWOpUqUKfn5+tG7dmnXr1tnfX7JkCeXKleOrr76ifv36eHl58c8//+Q6JWQYBi+++CI33HADPj4+3HTTTXz66af299etW4fJZGLNmjW0aNECX19f2rZty759+xw+75dffkmLFi3w9vYmNDSUu+66K8+1ikjxUGARkTzr2LEjo0aNYsCAASQlJbFr1y4mTpzIm2++SaVKlezt3n33Xdzd3fn555+ZM2cOr776Km+99Zb9/YcffpiNGzfy8ccf8+uvv3L33XfTs2dP9u/fb2+TlpZGdHQ0b731Fr///jsVK1bMVc+kSZNYvHgx8+fP5/fff+epp57igQce4IcffnBoN3HiRGbNmsXWrVtxd3fnkUcesb/39ddfc9ddd3HrrbeyY8cOe7jJT60iUgwMEZGLPPTQQ4bZbDb8/PwcHtOnTzcMwzAyMzONpk2bGv369TMaNGhgDBo0yGH/W265xahXr55htVrt28aNG2fUq1fPMAzD+OuvvwyTyWTEx8c77NelSxdjwoQJhmEYxuLFiw3A2LlzZ67a7rjjDsMwDCM1NdXw9vY2Nm3a5NDm0UcfNe69917DMAxj7dq1BmCsXr3a/v7XX39tAEZ6erphGIYRFRVl3H///Zf9WeSlVhEpHrqGRURy6dSpE/Pnz3fYFhISAoCnpycffPABjRs3plq1asyePTvX/m3atMFkMtlfR0VFMWvWLCwWC9u3b8cwDG688UaHfTIzMylfvrz9taenJ40bN75ijXv27CEjI4Nu3bo5bM/KyqJp06YO2y7u5/xI0LFjx6hatSo7d+7kscceu+wx8lqriBQ9BRYRycXPz49atWpd8f1NmzYBcOrUKU6dOoWfn1+e+7ZarZjNZrZt24bZbHZ4z9/f3/7cx8fHIfRcrh+wndKpUqWKw3uXXiDs4eFhf36+z/P7+/j4XHetIlL0FFhEJF8OHDjAU089xZtvvsl///tfHnzwQdasWYOb24VL4n766SeHfX766Sdq166N2WymadOmWCwWjh07Rvv27Qtcx/mLcWNjY7nlllsK3E/jxo1Zs2YNDz/8cK73CqtWEbl+CiwikktmZiaJiYkO29zd3QkODmbAgAF0796dhx9+mF69etGoUSNmzZrFv//9b3vbuLg4Ro8ezZAhQ9i+fTuvv/46s2bNAuDGG2/k/vvv58EHH2TWrFk0bdqUEydO8P3339OoUSN69+6dpxoDAgIYM2YMTz31FFarlZtvvpnk5GQ2bdqEv78/Dz30UJ76mTJlCl26dKFmzZrcc8895OTk8M033zB27NhCq1VErp8Ci4jksmrVKodZPwB16tThvvvu49ChQ/zvf/8DIDw8nLfeeot+/frRrVs3mjRpAsCDDz5Ieno6rVq1wmw2M3LkSAYPHmzva/HixTz33HM8/fTTxMfHU758eaKiovIdAGbMmEHFihWJjo7m77//ply5cjRr1oxnnnkmz3107NiRTz75hBkzZjBz5kwCAwPp0KFDodcqItdHK92KSKHq2LEjTZo0uezFuCIiBaV1WERERMTlKbCIiIiIy9MpIREREXF5GmERERERl6fAIiIiIi5PgUVERERcngKLiIiIuDwFFhEREXF5CiwiIiLi8hRYRERExOUpsIiIiIjLU2ARERERl/f/OTXKaMbJaXIAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(acc_stream_naive, label='Naive')\n",
    "plt.plot(acc_stream_er, label='ER')\n",
    "plt.plot(acc_stream_der, label='DER')\n",
    "plt.xticks(range(len(acc_stream_naive)))\n",
    "plt.xlabel('Experience')\n",
    "plt.title('Accuracy on stream')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(acc_trained_exps_naive, label='Naive')\n",
    "plt.plot(acc_trained_exps_er, label='ER')\n",
    "plt.plot(acc_trained_exps_der, label='DER')\n",
    "plt.xticks(range(len(acc_stream_naive)))\n",
    "plt.xlabel('Experience')\n",
    "plt.title('Accuracy on trained experiences')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(loss_stream_naive, label='Naive')  \n",
    "plt.plot(loss_stream_er, label='ER')\n",
    "plt.plot(loss_stream_der, label='DER')\n",
    "plt.xticks(range(len(acc_stream_naive)))\n",
    "plt.xlabel('Experience')\n",
    "plt.title('Loss on stream')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "continual",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.9"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}