Training_Yes_No_Torch.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8d9527d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "import IPython as ipd\n",
    "import librosa\n",
    "from utils import *\n",
    "\n",
    "import torch\n",
    "import torch.nn\n",
    "import torch.optim\n",
    "from torch.utils.data import IterableDataset\n",
    "\n",
    "import os\n",
    "import timeit\n",
    "from time import time\n",
    "\n",
    "import torchaudio\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d8514309",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "USING DEVICE cuda\n"
     ]
    }
   ],
   "source": [
    "DEVICE = \"cpu\"\n",
    "if torch.cuda.is_available():\n",
    "    DEVICE = \"cuda\"\n",
    "\n",
    "print(f\"USING DEVICE {DEVICE}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e5298d22",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  Training Set: 84843\n",
      "   Testing Set: 11005\n",
      "Validation Set: 9981\n",
      "   Data Labels: 35\n"
     ]
    }
   ],
   "source": [
    "# Define source paths. Place data under working directory and name folder \"voicedata\"\n",
    "DATA_DIR = \"./commands\"\n",
    "TARGET_RATE = 8000\n",
    "INPUT_SIZE = 64\n",
    "EPOCHS = 10\n",
    "os.makedirs(DATA_DIR, exist_ok=True)\n",
    "train_set = SpeechCommandsSubset(root=DATA_DIR, download=True, rate=TARGET_RATE)\n",
    "test_set = SpeechCommandsSubset(root=DATA_DIR, download=True, subset=\"testing\", rate=TARGET_RATE)\n",
    "validate_set = SpeechCommandsSubset(root=DATA_DIR, download=True, subset=\"validation\", rate=TARGET_RATE)\n",
    "\n",
    "# labels = sorted(list(set(datapoint[2] for datapoint in train_set)))\n",
    "labels = train_set.labels\n",
    "\n",
    "print(f'  Training Set: {len(train_set)}')\n",
    "print(f'   Testing Set: {len(test_set)}')\n",
    "print(f'Validation Set: {len(validate_set)}')\n",
    "\n",
    "print(f'   Data Labels: {len(labels)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b6705c8",
   "metadata": {},
   "source": [
    "# Preview sample 52000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1f3c0731",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label: right (22)\n",
      "8000\n",
      "torch.Size([1, 8000])\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "                <audio  controls=\"controls\" >\n",
       "                    <source src=\"data:audio/wav;base64,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\" type=\"audio/wav\" />\n",
       "                    Your browser does not support the audio element.\n",
       "                </audio>\n",
       "              "
      ],
      "text/plain": [
       "<IPython.lib.display.Audio object>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "waveform, rate, label, speaker, id = train_set[52000]\n",
    "print(f\"Label: {label} ({label_to_index(label, labels)})\")\n",
    "print(rate)\n",
    "print(waveform.size())\n",
    "ipd.display.Audio(data=waveform, rate=rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "da62df26",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y = waveform.view(-1).numpy()\n",
    "times = np.linspace(0, len(Y) / rate, len(Y))\n",
    "plot(times, Y, title=\"Sample 52000\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "23eb5f2c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "mel = torchaudio.transforms.MelSpectrogram(sample_rate=TARGET_RATE, n_mels=INPUT_SIZE)\n",
    "melspec = mel(waveform).squeeze()\n",
    "plot_mel(melspec, title=\"sample 52000 MelSpectrogram\", figsize=(8,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "dc08c633",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([1, 64, 32])\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "mfcc = torchaudio.transforms.MFCC(\n",
    "    sample_rate=TARGET_RATE, \n",
    "    n_mfcc=INPUT_SIZE,\n",
    "    melkwargs={\n",
    "        \"n_fft\": 2048,\n",
    "        \"n_mels\": 128,\n",
    "        \"hop_length\": 256,\n",
    "        \"mel_scale\": \"htk\",\n",
    "    },\n",
    ")\n",
    "mfccspec = mfcc(waveform)\n",
    "plot_mel(mfccspec.squeeze(), title=\"sample 52000 MFCC\", figsize=(8,6))\n",
    "print(mfccspec.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "62ff510c",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_set = SpeechCommandsSubset(\n",
    "    root=DATA_DIR,\n",
    "    download=True,\n",
    "    rate=TARGET_RATE,\n",
    "    feature=mfcc\n",
    ")\n",
    "test_set = SpeechCommandsSubset(\n",
    "    root=DATA_DIR,\n",
    "    download=True,\n",
    "    subset=\"testing\",\n",
    "    rate=TARGET_RATE,\n",
    "    feature=mfcc\n",
    ")\n",
    "validate_set = SpeechCommandsSubset(\n",
    "    root=DATA_DIR,\n",
    "    download=True,\n",
    "    subset=\"validation\",\n",
    "    rate=TARGET_RATE,\n",
    "    feature=mfcc\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "76bedb3c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sample, rate, label, speaker, id = train_set[52000]\n",
    "plot_mel(sample.squeeze(), title=\"sample 52000\", figsize=(8,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5594b42d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sample, rate, label, speaker, id = test_set[1000]\n",
    "plot_mel(sample.squeeze(), title=\"sample 52000\", figsize=(8,6))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec55e773",
   "metadata": {},
   "source": [
    "# Define and run the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "e51e199d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_results(CATEGORY, cat_title, time_data, loss_data, accuracy_data):\n",
    "    fig, ax = plt.subplots(3, 1, figsize=(10,16))\n",
    "    plot(CATEGORY, time_data, ax=ax[0])\n",
    "    ax[0].set_xlabel(cat_title)\n",
    "    ax[0].set_xticks(CATEGORY)\n",
    "    ax[0].grid(True)\n",
    "    ax[0].set_ylabel(\"time to run\")\n",
    "    ax[0].set_title(f\"time to run for each {cat_title}\")\n",
    "\n",
    "    for l in loss_data:\n",
    "        plot(list(range(DEF_EPOCHS)), l, ax=ax[1])\n",
    "    ax[1].legend(CATEGORY)\n",
    "    ax[1].set_xlabel(\"epoch\")\n",
    "    ax[1].grid(True)\n",
    "    ax[1].set_ylabel(\"loss\")\n",
    "    ax[1].set_title(f\"Loss curves for {cat_title}\")\n",
    "\n",
    "    plot(CATEGORY, accuracy_data, ax=ax[2])\n",
    "    ax[2].set_xlabel(cat_title)\n",
    "    ax[2].set_xticks(CATEGORY)\n",
    "    ax[2].grid(True)\n",
    "    ax[2].set_ylabel(\"accuracy\")\n",
    "    ax[2].set_title(f\"Accuracy over {cat_title}\")\n",
    "\n",
    "    pass\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "de274aff",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ImprovedRNNCell(RNNCell):\n",
    "    '''\n",
    "    Simple RNN with weighted residual connection.\n",
    "    '''\n",
    "\n",
    "    def __init__(self, input_size, hidden_size,\n",
    "                 update_nonlinearity=\"tanh\",\n",
    "                 name=\"ImprovedRNN\"):\n",
    "        super(ImprovedRNNCell, self).__init__(input_size, hidden_size,\n",
    "                                           None, update_nonlinearity,\n",
    "                                           1, 1, 1, 0, 0, 0, 0)\n",
    "        # TODO: Initialize all the required variables\n",
    "        \n",
    "        # Reduce W and U starting values to a small fraction.\n",
    "        self.rnn = torch.nn.RNN(\n",
    "            input_size, \n",
    "            hidden_size, \n",
    "            nonlinearity=update_nonlinearity,\n",
    "            batch_first=True\n",
    "        )\n",
    "        self._name = name\n",
    "\n",
    "    @property\n",
    "    def name(self):\n",
    "        return self._name\n",
    "\n",
    "    @property\n",
    "    def cellType(self):\n",
    "        return \"RNN\"\n",
    "\n",
    "    def forward(self, input, state):  \n",
    "        return self.rnn(input, state)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b21015a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "BATCH_SIZE = 24\n",
    "LEARNING_RATE = 0.005\n",
    "NON_LINEARITY = \"tanh\"\n",
    "PERFORMANCE_MODE = False\n",
    "DEBUG_MODE = False\n",
    "\n",
    "def initialize_model(input_size, hidden_size, non_linearity=NON_LINEARITY, fn=None, learning_rate=LEARNING_RATE):\n",
    "    model = ImprovedRNNCell(input_size, hidden_size, update_nonlinearity=non_linearity)\n",
    "    model = model.to(DEVICE)\n",
    "    optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
    "    if not fn:\n",
    "        loss_fn = torch.nn.CrossEntropyLoss()\n",
    "    else:\n",
    "        loss_fn = fn\n",
    "\n",
    "    return model, optimizer, loss_fn\n",
    "\n",
    "def train_model(model, optimizer, loss_fn, train_data, labels, batch_size=BATCH_SIZE, epochs=1, cap=None):\n",
    "    # Initialize model\n",
    "    model.train()\n",
    "\n",
    "    hidden = torch.zeros(1, batch_size, len(labels))\n",
    "    hidden = hidden.to(DEVICE)\n",
    "    loss_summary = 0\n",
    "    data_size = 0\n",
    "    losses = []\n",
    "    micro_times = {\n",
    "        \"init\": 0,\n",
    "        \"model\": 0,\n",
    "        \"lossfn\": 0,\n",
    "        \"backward\": 0,\n",
    "        \"optimizer\": 0,\n",
    "        \"total\": 0\n",
    "    }\n",
    "\n",
    "    num_workers = 0 if DEVICE==\"cpu\" else 4\n",
    "    pin_memory = False if DEVICE==\"cpu\" else True\n",
    "\n",
    "    data_loader = torch.utils.data.DataLoader(\n",
    "        train_data,\n",
    "        batch_size=batch_size,\n",
    "        shuffle=True,\n",
    "        num_workers=num_workers,\n",
    "        pin_memory=pin_memory\n",
    "    )\n",
    "\n",
    "    for e in range(epochs):\n",
    "        print(f\"=> [{e}] \", end=\"\")\n",
    "\n",
    "        diverged = False\n",
    "        count = 0.0\n",
    "        for waveform, rate, label, speaker, id in data_loader:\n",
    "            beggining = time()\n",
    "\n",
    "            if cap and count > cap:\n",
    "                break\n",
    "            count += 1\n",
    "            if waveform.shape[0] != batch_size:\n",
    "                continue\n",
    "\n",
    "            optimizer.zero_grad()\n",
    "            target = torch.tensor([label_to_index(x, labels) for x in label])\n",
    "\n",
    "            waveform = waveform.squeeze(1)\n",
    "            waveform = waveform.to(DEVICE)\n",
    "            target = target.squeeze(0)\n",
    "            target = target.to(DEVICE)\n",
    "            batch_len = waveform.shape[0]\n",
    "\n",
    "            starting = time()\n",
    "\n",
    "            output, hidden_next = model(waveform, hidden)\n",
    "            output = output.sum(dim=1)\n",
    "\n",
    "            model_done_time = time()\n",
    "            loss = loss_fn(output.view(batch_len, -1), target)\n",
    "            loss_summary += loss.item() * batch_len\n",
    "            data_size += batch_len\n",
    "\n",
    "            if not diverged and loss.item() != loss.item():\n",
    "                print(\"DIVERGED at \", count)\n",
    "                diverged = True\n",
    "                if DEBUG_MODE:\n",
    "                    print(\"output:\", torch.max(output).item())\n",
    "                    print(\"target:\", target)\n",
    "                    print(\"hidden:\", hidden)\n",
    "                    print(\"hidden:\", hidden_next)\n",
    "                    print(\"  loss:\", loss.item)\n",
    "                break\n",
    "                \n",
    "            loss_done_time = time()\n",
    "            loss.backward()\n",
    "            backward_done_time = time()\n",
    "            optimizer.step()\n",
    "            optimizer_done_time = time()\n",
    "            hidden = hidden_next.detach()\n",
    "            #hidden = torch.zeros(1, len(labels)).to(DEVICE)\n",
    "\n",
    "            final_time = time()\n",
    "            \n",
    "            micro_times[\"init\"] += starting - beggining\n",
    "            micro_times[\"model\"] += model_done_time - starting\n",
    "            micro_times[\"lossfn\"] += loss_done_time - model_done_time\n",
    "            micro_times[\"backward\"] += backward_done_time - loss_done_time\n",
    "            micro_times[\"optimizer\"] += optimizer_done_time - backward_done_time\n",
    "            micro_times[\"total\"] += final_time - beggining\n",
    "\n",
    "        # print(\"loss:\", loss_summary)\n",
    "        losses.append(loss_summary / data_size)\n",
    "    print(\"finished...\")\n",
    "    if PERFORMANCE_MODE:\n",
    "        print(f\"init: {micro_times['init']:.3f} seconds\")\n",
    "        print(f\"model: {micro_times['model']:.3f} seconds\")\n",
    "        print(f\"lossfn: {micro_times['lossfn']:.3f} seconds\")\n",
    "        print(f\"backward: {micro_times['backward']:.3f} seconds\")\n",
    "        print(f\"optimizer: {micro_times['optimizer']:.3f} seconds\")\n",
    "        print(f\"train time: {micro_times['total']:.3f} seconds\")\n",
    "    return losses\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4e6f27e6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=> [0] => [1] => [2] => [3] => [4] finished...\n",
      "init: 0.039 seconds\n",
      "model: 0.061 seconds\n",
      "lossfn: 0.014 seconds\n",
      "backward: 0.078 seconds\n",
      "optimizer: 0.039 seconds\n",
      "train time: 0.233 seconds\n",
      "[19.99553074155535, 15.115434987204415, 12.384351976334102, 10.490512019111996, 9.36503519330706]\n",
      "time result: 1.5372117970255204 seconds\n"
     ]
    }
   ],
   "source": [
    "m, o, l = initialize_model(INPUT_SIZE, len(labels), learning_rate=0.05)\n",
    "def test_model(model, opt, lossfn):\n",
    "    loss = train_model(model, opt, lossfn, train_set, labels, epochs=5, cap=20)\n",
    "    print(loss)\n",
    "    # train_model(model, opt, lossfn, train_set, labels, EPOCHS)\n",
    "\n",
    "PERFORMANCE_MODE = True\n",
    "result = timeit.timeit(stmt='test_model(m, o, l)', globals=globals(), number = 1)\n",
    "PERFORMANCE_MODE = False\n",
    "print(f\"time result: {result} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "3aa9b11f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=> [0] => [1] finished...\n",
      "init: 2.426 seconds\n",
      "model: 3.787 seconds\n",
      "lossfn: 0.897 seconds\n",
      "backward: 5.099 seconds\n",
      "optimizer: 3.111 seconds\n",
      "train time: 15.388 seconds\n",
      "[4.278400248955129, 4.087265914096859]\n",
      "time result: 56.0996488549863 seconds\n"
     ]
    }
   ],
   "source": [
    "m, o, l = initialize_model(INPUT_SIZE, len(labels), learning_rate=0.05)\n",
    "def test_model(model, opt, lossfn):\n",
    "    loss = train_model(model, opt, lossfn, train_set, labels, epochs=2, cap=None)\n",
    "    print(loss)\n",
    "    # train_model(model, opt, lossfn, train_set, labels, EPOCHS)\n",
    "\n",
    "PERFORMANCE_MODE = True\n",
    "result = timeit.timeit(stmt='test_model(m, o, l)', globals=globals(), number = 1)\n",
    "PERFORMANCE_MODE = False\n",
    "print(f\"time result: {result} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "d29a5eda",
   "metadata": {},
   "outputs": [],
   "source": [
    "def eval_model(model, loss_fn, test_data, labels, cap=3, batch_size = 1):\n",
    "    with torch.no_grad():\n",
    "        model.eval()\n",
    "        loss_summary = 0\n",
    "        data_size = 0\n",
    "        attempts = 0\n",
    "        hits = 0\n",
    "\n",
    "        hidden = torch.zeros(1, 1, len(labels))\n",
    "        hidden = hidden.to(DEVICE)\n",
    "\n",
    "        num_workers = 0 if DEVICE==\"cpu\" else 4\n",
    "        pin_memory = False if DEVICE==\"cpu\" else True\n",
    "        \n",
    "        data_loader = torch.utils.data.DataLoader(\n",
    "            test_data,\n",
    "            batch_size=batch_size,\n",
    "            shuffle=True,\n",
    "            num_workers=num_workers,\n",
    "            pin_memory=pin_memory\n",
    "        )\n",
    "\n",
    "        count = 0\n",
    "        for waveform, rate, label, speaker, id in data_loader:\n",
    "            if cap and count > cap:\n",
    "                break\n",
    "            count += 1\n",
    "            if waveform.shape[0] != batch_size:\n",
    "                continue\n",
    "\n",
    "            # send data to target device\n",
    "            waveform = waveform.squeeze(1)\n",
    "            waveform = waveform.to(DEVICE)\n",
    "            target = torch.tensor([label_to_index(x, labels) for x in label])\n",
    "            target = target.to(DEVICE)\n",
    "            batch_len = waveform.shape[0]\n",
    "\n",
    "            # Predict, calculate loss, calculate gradients\n",
    "            output, hidden_next = model(waveform, hidden)\n",
    "\n",
    "            loss = loss_fn(output.view(batch_len, -1), target)\n",
    "            loss_summary += loss.item()\n",
    "            data_size += 1\n",
    "\n",
    "            hidden = hidden_next.detach()\n",
    "            \n",
    "            prediction = torch.argmax(output.view(batch_len, -1), dim=1)\n",
    "            hits += prediction[prediction == target].shape[0]\n",
    "            attempts += 1\n",
    "            # print(prediction == target)\n",
    "            # print(prediction[prediction == target])\n",
    "        print(data_size)\n",
    "        loss_summary = loss_summary / data_size\n",
    "        return model, loss_summary, attempts, hits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "5a47c062",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=> [0] => [1] => [2] => [3] => [4] finished...\n",
      "init: 6.614 seconds\n",
      "model: 10.480 seconds\n",
      "lossfn: 2.432 seconds\n",
      "backward: 13.785 seconds\n",
      "optimizer: 8.075 seconds\n",
      "train time: 41.563 seconds\n",
      "11005\n",
      "[[4.0629006328474855, 3.7841235638339246, 3.686606392395041, 3.630905904628263, 3.5965718058109957]]\n",
      "[143.28029108047485]\n",
      "[0.015174920490686052]\n",
      "[167]\n"
     ]
    }
   ],
   "source": [
    "# Run a 5 epoch test on the whole data set to collect performance numbers on model\n",
    "losses = []\n",
    "times = []\n",
    "accuracies = []\n",
    "correct = []\n",
    "\n",
    "def timed_test(rate=INPUT_SIZE, nl=NON_LINEARITY, fn=None, lr=0.005, cap=5, ep=5, batch=16, ts=train_set, tst=test_set):\n",
    "    m, o, l = initialize_model(rate, len(labels), non_linearity=nl, fn=fn, learning_rate=lr)\n",
    "    start_time = time()\n",
    "    losses.append(train_model(m, o, l, ts, labels, epochs=ep, cap=cap, batch_size=24))\n",
    "    result = time() - start_time\n",
    "    times.append(result)\n",
    "    \n",
    "    m, loss, attempts, hits = eval_model(m, l, tst, labels, cap=cap, batch_size=1)\n",
    "    accuracies.append(hits/attempts)\n",
    "    correct.append(hits)\n",
    "    \n",
    "PERFORMANCE_MODE = True\n",
    "timed_test(nl=\"tanh\", fn=torch.nn.CrossEntropyLoss(), lr=0.005, cap=None, ep=5, batch=24)\n",
    "PERFORMANCE_MODE = False\n",
    "print(losses)\n",
    "print(times)\n",
    "print(accuracies)\n",
    "print(correct)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "547faed8",
   "metadata": {},
   "outputs": [],
   "source": [
    "DEF_CAP=None\n",
    "DEF_EPOCHS=10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "d61df8eb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "== Training with 1 samples per batch\n",
      "=> [0] => [1] => [2] => [3] => [4] => [5] => [6] => [7] => [8] => [9] finished...\n",
      "11005\n",
      "== Training with 8 samples per batch\n",
      "=> [0] => [1] => [2] => [3] => [4] => [5] => [6] => [7] => [8] => [9] finished...\n",
      "11005\n",
      "== Training with 16 samples per batch\n",
      "=> [0] => [1] => [2] => [3] => [4] => [5] => [6] => [7] => [8] => [9] finished...\n",
      "11005\n",
      "== Training with 24 samples per batch\n",
      "=> [0] => [1] => [2] => [3] => [4] => [5] => [6] => [7] => [8] => [9] finished...\n",
      "11005\n",
      "== Training with 48 samples per batch\n",
      "=> [0] => [1] => [2] => [3] => [4] => [5] => [6] => [7] => [8] => [9] finished...\n",
      "11005\n"
     ]
    }
   ],
   "source": [
    "losses = []\n",
    "times = []\n",
    "accuracies = []\n",
    "correct = []\n",
    "\n",
    "BATCHES = [1, 8, 16, 24, 48]\n",
    "for b in BATCHES:\n",
    "    print(f\"== Training with {b} samples per batch\")\n",
    "    timed_test(nl=\"tanh\", fn=torch.nn.CrossEntropyLoss(), lr=0.005, cap=DEF_CAP, ep=DEF_EPOCHS, batch=b)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "fc51398d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 batch: time taken = 292.495 seconds\n",
      "8 batch: time taken = 288.462 seconds\n",
      "16 batch: time taken = 289.989 seconds\n",
      "24 batch: time taken = 287.220 seconds\n",
      "48 batch: time taken = 285.421 seconds\n",
      "1 batch: final loss = 3.500224050072666\n",
      "8 batch: final loss = 3.5044865201925792\n",
      "16 batch: final loss = 3.5208801197769617\n",
      "24 batch: final loss = 3.4984566583323176\n",
      "48 batch: final loss = 3.506333531389142\n",
      "1 batch: accuracy = 1.517%\n",
      "8 batch: accuracy = 1.508%\n",
      "16 batch: accuracy = 1.545%\n",
      "24 batch: accuracy = 1.517%\n",
      "48 batch: accuracy = 1.490%\n"
     ]
    }
   ],
   "source": [
    "for t, b in zip (times, BATCHES):\n",
    "    print(f\"{b} batch: time taken = {t:.3f} seconds\")\n",
    "\n",
    "for l, b in zip (losses, BATCHES):\n",
    "    print(f\"{b} batch: final loss = {l[-1]}\")\n",
    "\n",
    "for a, b in zip (accuracies, BATCHES):\n",
    "    print(f\"{b} batch: accuracy = {a*100:.3f}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "c2cb0d72",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1152 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_results(BATCHES, \"batch sizes\", times, losses, accuracies)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b498fd18",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}