MambaHSI_Demo.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MambaHSI Demo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import time\n",
    "import math\n",
    "import copy\n",
    "import torch\n",
    "import random\n",
    "import argparse\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import metrics\n",
    "from torchvision import models,transforms\n",
    "from sklearn.preprocessing import MinMaxScaler,StandardScaler\n",
    "from thop import profile\n",
    "\n",
    "import utils.data_load_operate as data_load_operate\n",
    "\n",
    "from utils.Loss import head_loss,resize\n",
    "from utils.Ievaluation import Evaluator\n",
    "from utils.HSICommonUtils import normlize3D, ImageStretching\n",
    "\n",
    "from utils.setup_logger import setup_logger\n",
    "from utils.visual_predict import visualize_predict\n",
    "from PIL import Image\n",
    "from Imodel.CommonTemplate import CommonTemplate\n",
    "\n",
    "\n",
    "torch.autograd.set_detect_anomaly(True)\n",
    "\n",
    "time_current = time.strftime(\"%y-%m-%d-%H.%M\", time.localtime())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def vis_a_image(gt_vis,pred_vis,save_single_predict_path,save_single_gt_path,only_vis_label=False):\n",
    "    visualize_predict(gt_vis,pred_vis,save_single_predict_path,save_single_gt_path,only_vis_label=only_vis_label)\n",
    "    visualize_predict(gt_vis,pred_vis,save_single_predict_path.replace('.png','_mask.png'),save_single_gt_path,only_vis_label=True)\n",
    "\n",
    "\n",
    "# random seed setting\n",
    "def setup_seed(seed):\n",
    "    torch.manual_seed(seed)\n",
    "    torch.cuda.manual_seed_all(seed)\n",
    "    os.environ['PYTHONHASHSEED'] = str(seed)\n",
    "    np.random.seed(seed)\n",
    "    random.seed(seed)\n",
    "    torch.backends.cudnn.deterministic = True\n",
    "    torch.backends.cudnn.benchmark = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### config parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_parser():\n",
    "    parser = argparse.ArgumentParser()\n",
    "    parser.add_argument('--model_flag', type=int, default=0)\n",
    "    parser.add_argument('--dataset_index', type=int,default=0) \n",
    "    # parser.add_argument('--eval_mode', action='store_true', default=False)\n",
    "    parser.add_argument('--mamba_type', type=str, default='both')\n",
    "    parser.add_argument('--use_residual', action='store_true', default=True)\n",
    "    parser.add_argument('--use_att', action='store_true', default=True)\n",
    "    parser.add_argument('--token_num', type=int, default=4)\n",
    "    parser.add_argument('--hidden_dim', type=int, default=128)\n",
    "\n",
    "    parser.add_argument('--lr', type=float, default=0.0003)\n",
    "    parser.add_argument('--max_epoch', type=int, default=200)  # HSIUNet\n",
    "  \n",
    "    parser.add_argument('--train_samples', type=int, default=30)\n",
    "    parser.add_argument('--val_samples', type=int, default=10)\n",
    "    parser.add_argument('--exp_name',type=str,default='Demo')\n",
    "    parser.add_argument('--level',type=str,default='image')\n",
    "    parser.add_argument('--patch_size', type=int, default=1)\n",
    "    parser.add_argument('--use_transformer', action='store_true',default=False)\n",
    "    args = parser.parse_known_args()[0]\n",
    "    return args\n",
    "\n",
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
    "args = get_parser()\n",
    "\n",
    "\n",
    "model_list = ['MambaHSI']\n",
    "\n",
    "record_computecost=True\n",
    "\n",
    "exp_name = args.exp_name\n",
    "\n",
    "seed_list = [6]  \n",
    "\n",
    "num_list = [args.train_samples, args.val_samples]\n",
    "\n",
    "model_flag = args.model_flag\n",
    "dataset_index = args.dataset_index\n",
    "\n",
    "max_epoch = args.max_epoch\n",
    "learning_rate = args.lr\n",
    "\n",
    "net_name = model_list[model_flag]\n",
    "\n",
    "paras_dict = {'model_flag':model_flag,'dataset_index':dataset_index,'net_name':net_name,'num_list':num_list,\n",
    "              'lr':learning_rate,'seed_list':seed_list,'mamba_type':args.mamba_type,'token_num':args.token_num,'use_residual':args.use_residual}\n",
    "\n",
    "\n",
    "\n",
    "data_set_name_list = ['UP', 'HanChuan', 'HongHu', 'Houston']\n",
    "data_set_name = data_set_name_list[dataset_index]\n",
    "\n",
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "])\n",
    "\n",
    "data_set_path = \"./data\"\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./work_dir/Demo/MambaHSI/tr30val10_lr0.0003_both_token4_HD128/UP\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/mnt/lssd/liyapeng/research/HSI/codes/MambaHSI/utils/Loss.py:20: UserWarning: When align_corners=True, the output would more aligned if input size (152, 85) is `x+1` and out size (610, 340) is `nx+1`\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter:0|loss:2.2594406604766846|loss_dict:{}\n",
      "Evaluate 0|OA:0.18888888888888888|MACC:0.1888888888888889|Kappa:0.08750000000000001|MIOU:0.05743525480367585|IOU:[0.         0.10714286 0.         0.         0.35714286 0.\n",
      " 0.05263158 0.         0.        ]|ACC:[0.  0.6 0.  0.  1.  0.  0.1 0.  0. ]\n",
      "Iter:1|loss:2.101924419403076|loss_dict:{}\n",
      "Evaluate 1|OA:0.12222222222222222|MACC:0.12222222222222223|Kappa:0.012500000000000002|MIOU:0.024346024346024345|IOU:[0.         0.09090909 0.         0.         0.12820513 0.\n",
      " 0.         0.         0.        ]|ACC:[0.  0.1 0.  0.  1.  0.  0.  0.  0. ]\n",
      "Iter:2|loss:2.0082926750183105|loss_dict:{}\n",
      "Evaluate 2|OA:0.13333333333333333|MACC:0.13333333333333336|Kappa:0.025000000000000005|MIOU:0.034599002068881585|IOU:[0.         0.09090909 0.         0.         0.12048193 0.\n",
      " 0.         0.         0.1       ]|ACC:[0.  0.1 0.  0.  1.  0.  0.  0.  0.1]\n",
      "Iter:3|loss:1.955693006515503|loss_dict:{}\n",
      "Evaluate 3|OA:0.14444444444444443|MACC:0.14444444444444446|Kappa:0.03749999999999999|MIOU:0.03961510530137981|IOU:[0.         0.         0.         0.         0.13888889 0.11764706\n",
      " 0.         0.         0.1       ]|ACC:[0.  0.  0.  0.  1.  0.2 0.  0.  0.1]\n",
      "Iter:4|loss:1.9179352521896362|loss_dict:{}\n",
      "Evaluate 4|OA:0.2|MACC:0.20000000000000004|Kappa:0.10000000000000002|MIOU:0.06987804878048781|IOU:[0.         0.         0.         0.         0.24390244 0.125\n",
      " 0.         0.16       0.1       ]|ACC:[0.  0.  0.  0.  1.  0.3 0.  0.4 0.1]\n",
      "Iter:5|loss:1.8831706047058105|loss_dict:{}\n",
      "Evaluate 5|OA:0.25555555555555554|MACC:0.2555555555555556|Kappa:0.1625|MIOU:0.10114607614607615|IOU:[0.         0.         0.         0.         0.45454545 0.125\n",
      " 0.         0.23076923 0.1       ]|ACC:[0.  0.  0.  0.  1.  0.3 0.  0.9 0.1]\n",
      "Iter:6|loss:1.8492618799209595|loss_dict:{}\n",
      "Evaluate 6|OA:0.26666666666666666|MACC:0.26666666666666666|Kappa:0.17500000000000002|MIOU:0.12376543209876543|IOU:[0.         0.         0.         0.         0.66666667 0.125\n",
      " 0.         0.22222222 0.1       ]|ACC:[0.  0.  0.  0.  1.  0.3 0.  1.  0.1]\n",
      "Iter:7|loss:1.8170024156570435|loss_dict:{}\n",
      "Evaluate 7|OA:0.26666666666666666|MACC:0.26666666666666666|Kappa:0.17500000000000002|MIOU:0.13522853957636569|IOU:[0.         0.         0.         0.         0.76923077 0.13043478\n",
      " 0.         0.2173913  0.1       ]|ACC:[0.  0.  0.  0.  1.  0.3 0.  1.  0.1]\n",
      "Iter:8|loss:1.7870796918869019|loss_dict:{}\n",
      "Evaluate 8|OA:0.3|MACC:0.3|Kappa:0.2125|MIOU:0.15701029575709885|IOU:[0.         0.17647059 0.         0.         0.76923077 0.15\n",
      " 0.         0.2173913  0.1       ]|ACC:[0.  0.3 0.  0.  1.  0.3 0.  1.  0.1]\n",
      "Iter:9|loss:1.7587295770645142|loss_dict:{}\n",
      "Evaluate 9|OA:0.3|MACC:0.3|Kappa:0.2125|MIOU:0.1567981641208186|IOU:[0.         0.15789474 0.         0.         0.76923077 0.16666667\n",
      " 0.         0.2173913  0.1       ]|ACC:[0.  0.3 0.  0.  1.  0.3 0.  1.  0.1]\n",
      "Iter:10|loss:1.7305790185928345|loss_dict:{}\n",
      "Evaluate 10|OA:0.3333333333333333|MACC:0.3333333333333333|Kappa:0.25|MIOU:0.18044363044363046|IOU:[0.         0.25       0.         0.         0.76923077 0.26666667\n",
      " 0.         0.23809524 0.1       ]|ACC:[0.  0.5 0.  0.  1.  0.4 0.  1.  0.1]\n",
      "Iter:11|loss:1.701623558998108|loss_dict:{}\n",
      "Evaluate 11|OA:0.3111111111111111|MACC:0.3111111111111111|Kappa:0.22500000000000003|MIOU:0.1637488076512467|IOU:[0.         0.22727273 0.         0.         0.76923077 0.13333333\n",
      " 0.         0.24390244 0.1       ]|ACC:[0.  0.5 0.  0.  1.  0.2 0.  1.  0.1]\n",
      "Iter:12|loss:1.6717647314071655|loss_dict:{}\n",
      "Evaluate 12|OA:0.36666666666666664|MACC:0.3666666666666667|Kappa:0.2875|MIOU:0.2094239094239094|IOU:[0.         0.27272727 0.         0.         0.76923077 0.28571429\n",
      " 0.1        0.25714286 0.2       ]|ACC:[0.  0.6 0.  0.  1.  0.4 0.2 0.9 0.2]\n",
      "Iter:13|loss:1.6419622898101807|loss_dict:{}\n",
      "Evaluate 13|OA:0.43333333333333335|MACC:0.43333333333333335|Kappa:0.36250000000000004|MIOU:0.246376956895891|IOU:[0.         0.36363636 0.         0.         0.76923077 0.33333333\n",
      " 0.26086957 0.29032258 0.2       ]|ACC:[0.  0.8 0.  0.  1.  0.4 0.6 0.9 0.2]\n",
      "Iter:14|loss:1.6123216152191162|loss_dict:{}\n",
      "Evaluate 14|OA:0.45555555555555555|MACC:0.4555555555555555|Kappa:0.3875|MIOU:0.2604410595789906|IOU:[0.         0.40909091 0.         0.         0.76923077 0.36363636\n",
      " 0.29166667 0.31034483 0.2       ]|ACC:[0.  0.9 0.  0.  1.  0.4 0.7 0.9 0.2]\n",
      "Iter:15|loss:1.582429051399231|loss_dict:{}\n",
      "Evaluate 15|OA:0.5|MACC:0.5000000000000001|Kappa:0.43750000000000006|MIOU:0.3003626003626004|IOU:[0.         0.45454545 0.         0.         0.76923077 0.6\n",
      " 0.33333333 0.34615385 0.2       ]|ACC:[0.  1.  0.  0.  1.  0.6 0.8 0.9 0.2]\n",
      "Iter:16|loss:1.552247166633606|loss_dict:{}\n",
      "Evaluate 16|OA:0.5|MACC:0.5000000000000001|Kappa:0.43750000000000006|MIOU:0.30071872571872577|IOU:[0.         0.45454545 0.         0.         0.76923077 0.6\n",
      " 0.30769231 0.375      0.2       ]|ACC:[0.  1.  0.  0.  1.  0.6 0.8 0.9 0.2]\n",
      "Iter:17|loss:1.5222080945968628|loss_dict:{}\n",
      "Evaluate 17|OA:0.5|MACC:0.5000000000000001|Kappa:0.43750000000000006|MIOU:0.30071872571872577|IOU:[0.         0.45454545 0.         0.         0.76923077 0.6\n",
      " 0.30769231 0.375      0.2       ]|ACC:[0.  1.  0.  0.  1.  0.6 0.8 0.9 0.2]\n",
      "Iter:18|loss:1.4929414987564087|loss_dict:{}\n",
      "Evaluate 18|OA:0.4888888888888889|MACC:0.48888888888888893|Kappa:0.425|MIOU:0.2874117428465255|IOU:[0.         0.43478261 0.         0.         0.76923077 0.5\n",
      " 0.30769231 0.375      0.2       ]|ACC:[0.  1.  0.  0.  1.  0.5 0.8 0.9 0.2]\n",
      "Iter:19|loss:1.46479070186615|loss_dict:{}\n",
      "Evaluate 19|OA:0.5|MACC:0.5|Kappa:0.43750000000000006|MIOU:0.3124709250313115|IOU:[0.         0.43478261 0.         0.         0.83333333 0.6\n",
      " 0.2962963  0.34782609 0.3       ]|ACC:[0.  1.  0.  0.  1.  0.6 0.8 0.8 0.3]\n",
      "Iter:20|loss:1.4370747804641724|loss_dict:{}\n",
      "Evaluate 20|OA:0.5|MACC:0.5|Kappa:0.43750000000000006|MIOU:0.3124709250313115|IOU:[0.         0.43478261 0.         0.         0.83333333 0.6\n",
      " 0.2962963  0.34782609 0.3       ]|ACC:[0.  1.  0.  0.  1.  0.6 0.8 0.8 0.3]\n",
      "Iter:21|loss:1.4093869924545288|loss_dict:{}\n",
      "Evaluate 21|OA:0.5111111111111111|MACC:0.5111111111111111|Kappa:0.45|MIOU:0.32533873355129395|IOU:[0.         0.43478261 0.         0.         0.83333333 0.7\n",
      " 0.2962963  0.36363636 0.3       ]|ACC:[0.  1.  0.  0.  1.  0.7 0.8 0.8 0.3]\n",
      "Iter:22|loss:1.3818962574005127|loss_dict:{}\n",
      "Evaluate 22|OA:0.5222222222222223|MACC:0.5222222222222223|Kappa:0.4625000000000001|MIOU:0.33771606815085076|IOU:[0.         0.43478261 0.         0.         0.83333333 0.7\n",
      " 0.30769231 0.36363636 0.4       ]|ACC:[0.  1.  0.  0.  1.  0.7 0.8 0.8 0.4]\n",
      "Iter:23|loss:1.3547626733779907|loss_dict:{}\n",
      "Evaluate 23|OA:0.5333333333333333|MACC:0.5333333333333334|Kappa:0.47500000000000003|MIOU:0.352947052947053|IOU:[0.         0.45454545 0.         0.1        0.83333333 0.7\n",
      " 0.30769231 0.38095238 0.4       ]|ACC:[0.  1.  0.  0.1 1.  0.7 0.8 0.8 0.4]\n",
      "Iter:24|loss:1.3278558254241943|loss_dict:{}\n",
      "Evaluate 24|OA:0.5333333333333333|MACC:0.5333333333333334|Kappa:0.47500000000000003|MIOU:0.352947052947053|IOU:[0.         0.45454545 0.         0.1        0.83333333 0.7\n",
      " 0.30769231 0.38095238 0.4       ]|ACC:[0.  1.  0.  0.1 1.  0.7 0.8 0.8 0.4]\n",
      "Iter:25|loss:1.3010090589523315|loss_dict:{}\n",
      "Evaluate 25|OA:0.6111111111111112|MACC:0.6111111111111112|Kappa:0.5625|MIOU:0.44842826019296605|IOU:[0.         0.58823529 0.         0.6        0.83333333 0.7\n",
      " 0.33333333 0.38095238 0.6       ]|ACC:[0.  1.  0.  0.6 1.  0.7 0.8 0.8 0.6]\n",
      "Iter:26|loss:1.2742546796798706|loss_dict:{}\n",
      "Evaluate 26|OA:0.6555555555555556|MACC:0.6555555555555554|Kappa:0.6125000000000002|MIOU:0.5159307359307359|IOU:[0.         0.71428571 0.         0.9        0.90909091 0.7\n",
      " 0.32       0.4        0.7       ]|ACC:[0.  1.  0.  0.9 1.  0.7 0.8 0.8 0.7]\n",
      "Iter:27|loss:1.2476284503936768|loss_dict:{}\n",
      "Evaluate 27|OA:0.6777777777777778|MACC:0.6777777777777778|Kappa:0.6375000000000001|MIOU:0.5566051881841356|IOU:[0.         0.83333333 0.09090909 1.         0.90909091 0.8\n",
      " 0.30769231 0.36842105 0.7       ]|ACC:[0.  1.  0.1 1.  1.  0.8 0.8 0.7 0.7]\n",
      "Iter:28|loss:1.2209721803665161|loss_dict:{}\n",
      "Evaluate 28|OA:0.7333333333333333|MACC:0.7333333333333334|Kappa:0.7000000000000001|MIOU:0.6140720934838583|IOU:[0.         0.83333333 0.25       1.         0.90909091 0.8\n",
      " 0.36363636 0.47058824 0.9       ]|ACC:[0.  1.  0.3 1.  1.  0.8 0.8 0.8 0.9]\n",
      "Iter:29|loss:1.1941949129104614|loss_dict:{}\n",
      "Evaluate 29|OA:0.7333333333333333|MACC:0.7333333333333334|Kappa:0.7000000000000001|MIOU:0.61592394533571|IOU:[0.         0.75       0.25       0.90909091 1.         0.8\n",
      " 0.36363636 0.47058824 1.        ]|ACC:[0.  0.9 0.3 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:30|loss:1.1672435998916626|loss_dict:{}\n",
      "Evaluate 30|OA:0.7444444444444445|MACC:0.7444444444444445|Kappa:0.7125|MIOU:0.6271072065189712|IOU:[0.         0.75       0.33333333 0.90909091 1.         0.8\n",
      " 0.38095238 0.47058824 1.        ]|ACC:[0.  0.9 0.4 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:31|loss:1.1400024890899658|loss_dict:{}\n",
      "Evaluate 31|OA:0.7444444444444445|MACC:0.7444444444444445|Kappa:0.7125|MIOU:0.6271072065189712|IOU:[0.         0.75       0.33333333 0.90909091 1.         0.8\n",
      " 0.38095238 0.47058824 1.        ]|ACC:[0.  0.9 0.4 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:32|loss:1.1123878955841064|loss_dict:{}\n",
      "Evaluate 32|OA:0.7444444444444445|MACC:0.7444444444444445|Kappa:0.7125|MIOU:0.6191225985343632|IOU:[0.         0.75       0.33333333 0.90909091 1.         0.8\n",
      " 0.4        0.47058824 0.90909091]|ACC:[0.  0.9 0.4 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:33|loss:1.084408164024353|loss_dict:{}\n",
      "Evaluate 33|OA:0.7222222222222222|MACC:0.7222222222222222|Kappa:0.6875000000000001|MIOU:0.5904068698186345|IOU:[0.         0.66666667 0.25       0.90909091 1.         0.72727273\n",
      " 0.38095238 0.47058824 0.90909091]|ACC:[0.  0.8 0.3 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:34|loss:1.0557798147201538|loss_dict:{}\n",
      "Evaluate 34|OA:0.7333333333333333|MACC:0.7333333333333334|Kappa:0.7000000000000001|MIOU:0.6029341029341029|IOU:[0.         0.66666667 0.33333333 0.90909091 1.         0.72727273\n",
      " 0.38095238 0.5        0.90909091]|ACC:[0.  0.8 0.4 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:35|loss:1.0262315273284912|loss_dict:{}\n",
      "Evaluate 35|OA:0.7444444444444445|MACC:0.7444444444444445|Kappa:0.7125|MIOU:0.6143097643097643|IOU:[0.         0.66666667 0.41666667 0.90909091 1.         0.72727273\n",
      " 0.4        0.5        0.90909091]|ACC:[0.  0.8 0.5 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:36|loss:0.9958024621009827|loss_dict:{}\n",
      "Evaluate 36|OA:0.7444444444444445|MACC:0.7444444444444445|Kappa:0.7125|MIOU:0.6143097643097643|IOU:[0.         0.66666667 0.41666667 0.90909091 1.         0.72727273\n",
      " 0.4        0.5        0.90909091]|ACC:[0.  0.8 0.5 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:37|loss:0.9647706151008606|loss_dict:{}\n",
      "Evaluate 37|OA:0.7333333333333333|MACC:0.7333333333333334|Kappa:0.7000000000000001|MIOU:0.5983164983164984|IOU:[0.         0.58333333 0.41666667 0.90909091 1.         0.66666667\n",
      " 0.4        0.5        0.90909091]|ACC:[0.  0.7 0.5 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:38|loss:0.9335954189300537|loss_dict:{}\n",
      "Evaluate 38|OA:0.7555555555555555|MACC:0.7555555555555556|Kappa:0.725|MIOU:0.6311270600744285|IOU:[0.1        0.66666667 0.41666667 1.         1.         0.66666667\n",
      " 0.42105263 0.5        0.90909091]|ACC:[0.1 0.8 0.5 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:39|loss:0.902673602104187|loss_dict:{}\n",
      "Evaluate 39|OA:0.7555555555555555|MACC:0.7555555555555556|Kappa:0.725|MIOU:0.632716049382716|IOU:[0.09090909 0.66666667 0.41666667 1.         1.         0.66666667\n",
      " 0.44444444 0.5        0.90909091]|ACC:[0.1 0.8 0.5 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:40|loss:0.8724632859230042|loss_dict:{}\n",
      "Evaluate 40|OA:0.7777777777777778|MACC:0.7777777777777778|Kappa:0.7500000000000001|MIOU:0.6586848880966528|IOU:[0.18181818 0.66666667 0.5        1.         1.         0.66666667\n",
      " 0.47058824 0.53333333 0.90909091]|ACC:[0.2 0.8 0.6 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:41|loss:0.8434661626815796|loss_dict:{}\n",
      "Evaluate 41|OA:0.7777777777777778|MACC:0.7777777777777778|Kappa:0.7500000000000001|MIOU:0.6602693602693602|IOU:[0.16666667 0.66666667 0.5        1.         1.         0.66666667\n",
      " 0.5        0.53333333 0.90909091]|ACC:[0.2 0.8 0.6 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:42|loss:0.8159757852554321|loss_dict:{}\n",
      "Evaluate 42|OA:0.7777777777777778|MACC:0.7777777777777778|Kappa:0.7500000000000001|MIOU:0.6602693602693602|IOU:[0.16666667 0.66666667 0.5        1.         1.         0.66666667\n",
      " 0.5        0.53333333 0.90909091]|ACC:[0.2 0.8 0.6 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:43|loss:0.789756178855896|loss_dict:{}\n",
      "Evaluate 43|OA:0.8111111111111111|MACC:0.8111111111111112|Kappa:0.7875|MIOU:0.7002164502164502|IOU:[0.33333333 0.66666667 0.58333333 1.         1.         0.66666667\n",
      " 0.57142857 0.57142857 0.90909091]|ACC:[0.4 0.8 0.7 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:44|loss:0.7643908858299255|loss_dict:{}\n",
      "Evaluate 44|OA:0.8333333333333334|MACC:0.8333333333333334|Kappa:0.8125000000000001|MIOU:0.7293169793169794|IOU:[0.5        0.66666667 0.58333333 1.         1.         0.66666667\n",
      " 0.66666667 0.57142857 0.90909091]|ACC:[0.6 0.8 0.7 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:45|loss:0.7395135760307312|loss_dict:{}\n",
      "Evaluate 45|OA:0.8444444444444444|MACC:0.8444444444444446|Kappa:0.8250000000000001|MIOU:0.7486772486772487|IOU:[0.58333333 0.66666667 0.58333333 1.         1.         0.66666667\n",
      " 0.66666667 0.57142857 1.        ]|ACC:[0.7 0.8 0.7 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:46|loss:0.7149802446365356|loss_dict:{}\n",
      "Evaluate 46|OA:0.8555555555555555|MACC:0.8555555555555556|Kappa:0.8375|MIOU:0.7646705146705146|IOU:[0.66666667 0.66666667 0.58333333 1.         1.         0.66666667\n",
      " 0.72727273 0.57142857 1.        ]|ACC:[0.8 0.8 0.7 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:47|loss:0.690940260887146|loss_dict:{}\n",
      "Evaluate 47|OA:0.8555555555555555|MACC:0.8555555555555556|Kappa:0.8375|MIOU:0.7638565138565139|IOU:[0.66666667 0.58333333 0.66666667 1.         1.         0.61538462\n",
      " 0.72727273 0.61538462 1.        ]|ACC:[0.8 0.7 0.8 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:48|loss:0.6673131585121155|loss_dict:{}\n",
      "Evaluate 48|OA:0.8555555555555555|MACC:0.8555555555555556|Kappa:0.8375|MIOU:0.7638565138565139|IOU:[0.66666667 0.58333333 0.66666667 1.         1.         0.61538462\n",
      " 0.72727273 0.61538462 1.        ]|ACC:[0.8 0.7 0.8 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:49|loss:0.6439520716667175|loss_dict:{}\n",
      "Evaluate 49|OA:0.8555555555555555|MACC:0.8555555555555556|Kappa:0.8375|MIOU:0.7638565138565139|IOU:[0.66666667 0.58333333 0.66666667 1.         1.         0.61538462\n",
      " 0.72727273 0.61538462 1.        ]|ACC:[0.8 0.7 0.8 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:50|loss:0.6208600401878357|loss_dict:{}\n",
      "Evaluate 50|OA:0.8555555555555555|MACC:0.8555555555555556|Kappa:0.8375|MIOU:0.7638565138565139|IOU:[0.66666667 0.58333333 0.66666667 1.         1.         0.61538462\n",
      " 0.72727273 0.61538462 1.        ]|ACC:[0.8 0.7 0.8 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:51|loss:0.5979958772659302|loss_dict:{}\n",
      "Evaluate 51|OA:0.8555555555555555|MACC:0.8555555555555556|Kappa:0.8375|MIOU:0.7665482665482666|IOU:[0.66666667 0.5        0.72727273 1.         1.         0.57142857\n",
      " 0.81818182 0.61538462 1.        ]|ACC:[0.8 0.6 0.8 1.  1.  0.8 0.9 0.8 1. ]\n",
      "Iter:52|loss:0.5755882263183594|loss_dict:{}\n",
      "Evaluate 52|OA:0.8666666666666667|MACC:0.8666666666666667|Kappa:0.85|MIOU:0.7823472823472825|IOU:[0.66666667 0.5        0.81818182 1.         1.         0.57142857\n",
      " 0.81818182 0.66666667 1.        ]|ACC:[0.8 0.6 0.9 1.  1.  0.8 0.9 0.8 1. ]\n",
      "Iter:53|loss:0.5535850524902344|loss_dict:{}\n",
      "Evaluate 53|OA:0.8666666666666667|MACC:0.8666666666666667|Kappa:0.85|MIOU:0.7823472823472825|IOU:[0.66666667 0.5        0.81818182 1.         1.         0.57142857\n",
      " 0.81818182 0.66666667 1.        ]|ACC:[0.8 0.6 0.9 1.  1.  0.8 0.9 0.8 1. ]\n",
      "Iter:54|loss:0.5316999554634094|loss_dict:{}\n",
      "Evaluate 54|OA:0.8777777777777778|MACC:0.8777777777777778|Kappa:0.8624999999999999|MIOU:0.7964905464905465|IOU:[0.66666667 0.58333333 0.81818182 1.         1.         0.61538462\n",
      " 0.81818182 0.66666667 1.        ]|ACC:[0.8 0.7 0.9 1.  1.  0.8 0.9 0.8 1. ]\n",
      "Iter:55|loss:0.5098071098327637|loss_dict:{}\n",
      "Evaluate 55|OA:0.8777777777777778|MACC:0.8777777777777778|Kappa:0.8624999999999999|MIOU:0.7973193473193473|IOU:[0.69230769 0.58333333 0.81818182 1.         1.         0.61538462\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.9 0.7 0.9 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:56|loss:0.4879635274410248|loss_dict:{}\n",
      "Evaluate 56|OA:0.8777777777777778|MACC:0.8777777777777778|Kappa:0.8624999999999999|MIOU:0.7973193473193473|IOU:[0.69230769 0.58333333 0.81818182 1.         1.         0.61538462\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.9 0.7 0.9 1.  1.  0.8 0.8 0.8 1. ]\n",
      "Iter:57|loss:0.46632689237594604|loss_dict:{}\n",
      "Evaluate 57|OA:0.8888888888888888|MACC:0.8888888888888888|Kappa:0.8749999999999999|MIOU:0.8117586117586119|IOU:[0.69230769 0.63636364 0.81818182 1.         1.         0.69230769\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.9 0.7 0.9 1.  1.  0.9 0.8 0.8 1. ]\n",
      "Iter:58|loss:0.4449937045574188|loss_dict:{}\n",
      "Evaluate 58|OA:0.8888888888888888|MACC:0.8888888888888888|Kappa:0.8749999999999999|MIOU:0.8117586117586119|IOU:[0.69230769 0.63636364 0.81818182 1.         1.         0.69230769\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.9 0.7 0.9 1.  1.  0.9 0.8 0.8 1. ]\n",
      "Iter:59|loss:0.4239959716796875|loss_dict:{}\n",
      "Evaluate 59|OA:0.8888888888888888|MACC:0.8888888888888888|Kappa:0.8749999999999999|MIOU:0.8117586117586119|IOU:[0.69230769 0.63636364 0.81818182 1.         1.         0.69230769\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.9 0.7 0.9 1.  1.  0.9 0.8 0.8 1. ]\n",
      "Iter:60|loss:0.4037463068962097|loss_dict:{}\n",
      "Evaluate 60|OA:0.8777777777777778|MACC:0.8777777777777778|Kappa:0.8624999999999999|MIOU:0.7956358456358457|IOU:[0.61538462 0.63636364 0.75       1.         1.         0.69230769\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.8 0.7 0.9 1.  1.  0.9 0.8 0.8 1. ]\n",
      "Iter:61|loss:0.3881341516971588|loss_dict:{}\n",
      "Evaluate 61|OA:0.8777777777777778|MACC:0.8777777777777778|Kappa:0.8624999999999999|MIOU:0.7977596477596478|IOU:[0.76923077 0.72727273 0.5        1.         1.         0.75\n",
      " 0.9        0.53333333 1.        ]|ACC:[1.  0.8 0.5 1.  1.  0.9 0.9 0.8 1. ]\n",
      "Iter:62|loss:0.4043838381767273|loss_dict:{}\n",
      "Evaluate 62|OA:0.8666666666666667|MACC:0.8666666666666667|Kappa:0.85|MIOU:0.7781995781995783|IOU:[0.61538462 0.63636364 0.71428571 1.         1.         0.69230769\n",
      " 0.8        0.54545455 1.        ]|ACC:[0.8 0.7 1.  1.  1.  0.9 0.8 0.6 1. ]\n",
      "Iter:63|loss:0.38243481516838074|loss_dict:{}\n",
      "Evaluate 63|OA:0.8888888888888888|MACC:0.8888888888888888|Kappa:0.8749999999999999|MIOU:0.8105931105931106|IOU:[0.66666667 0.63636364 0.75       1.         1.         0.69230769\n",
      " 0.8        0.75       1.        ]|ACC:[0.8 0.7 0.9 1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:64|loss:0.3427620828151703|loss_dict:{}\n",
      "Evaluate 64|OA:0.9|MACC:0.8999999999999999|Kappa:0.8875000000000002|MIOU:0.8316627816627817|IOU:[0.69230769 0.72727273 0.8        1.         1.         0.75\n",
      " 0.9        0.61538462 1.        ]|ACC:[0.9 0.8 0.8 1.  1.  0.9 0.9 0.8 1. ]\n",
      "Iter:65|loss:0.34848564863204956|loss_dict:{}\n",
      "Evaluate 65|OA:0.9|MACC:0.8999999999999999|Kappa:0.8875000000000002|MIOU:0.8282698782698783|IOU:[0.69230769 0.72727273 0.81818182 1.         1.         0.75\n",
      " 0.8        0.66666667 1.        ]|ACC:[0.9 0.8 0.9 1.  1.  0.9 0.8 0.8 1. ]\n",
      "Iter:66|loss:0.31577780842781067|loss_dict:{}\n",
      "Evaluate 66|OA:0.9111111111111111|MACC:0.911111111111111|Kappa:0.9000000000000001|MIOU:0.843939393939394|IOU:[0.66666667 0.72727273 0.83333333 1.         1.         0.75\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.8 0.8 1.  1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:67|loss:0.32051390409469604|loss_dict:{}\n",
      "Evaluate 67|OA:0.9222222222222223|MACC:0.9222222222222223|Kappa:0.9125000000000001|MIOU:0.8616161616161616|IOU:[0.75       0.72727273 0.90909091 1.         1.         0.75\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 0.8 1.  1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:68|loss:0.2900325059890747|loss_dict:{}\n",
      "Evaluate 68|OA:0.9333333333333333|MACC:0.9333333333333333|Kappa:0.925|MIOU:0.87996632996633|IOU:[0.83333333 0.81818182 0.8        1.         1.         0.81818182\n",
      " 0.9        0.75       1.        ]|ACC:[1.  0.9 0.8 1.  1.  0.9 0.9 0.9 1. ]\n",
      "Iter:69|loss:0.29254579544067383|loss_dict:{}\n",
      "Evaluate 69|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9168350168350169|IOU:[0.83333333 0.90909091 0.90909091 1.         1.         0.9\n",
      " 0.8        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:70|loss:0.26959922909736633|loss_dict:{}\n",
      "Evaluate 70|OA:0.9333333333333333|MACC:0.9333333333333333|Kappa:0.925|MIOU:0.8808080808080808|IOU:[0.66666667 0.90909091 0.83333333 1.         1.         0.9\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.8 1.  1.  1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:71|loss:0.2658061981201172|loss_dict:{}\n",
      "Evaluate 71|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:72|loss:0.2489771544933319|loss_dict:{}\n",
      "Evaluate 72|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9222222222222223|IOU:[0.75 1.   0.9  1.   1.   1.   0.9  0.75 1.  ]|ACC:[0.9 1.  0.9 1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:73|loss:0.24377916753292084|loss_dict:{}\n",
      "Evaluate 73|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9168350168350169|IOU:[0.83333333 0.90909091 0.90909091 1.         1.         0.9\n",
      " 0.8        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:74|loss:0.22890529036521912|loss_dict:{}\n",
      "Evaluate 74|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9168350168350169|IOU:[0.83333333 0.90909091 0.90909091 1.         1.         0.9\n",
      " 0.8        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  0.9 0.8 0.9 1. ]\n",
      "Iter:75|loss:0.22358091175556183|loss_dict:{}\n",
      "Evaluate 75|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:76|loss:0.20968970656394958|loss_dict:{}\n",
      "Evaluate 76|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9222222222222223|IOU:[0.75 1.   0.9  1.   1.   1.   0.9  0.75 1.  ]|ACC:[0.9 1.  0.9 1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:77|loss:0.20562490820884705|loss_dict:{}\n",
      "Evaluate 77|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9380471380471381|IOU:[0.83333333 1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:78|loss:0.1911061853170395|loss_dict:{}\n",
      "Evaluate 78|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9380471380471381|IOU:[0.83333333 1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:79|loss:0.18913333117961884|loss_dict:{}\n",
      "Evaluate 79|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:80|loss:0.17517106235027313|loss_dict:{}\n",
      "Evaluate 80|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:81|loss:0.17277365922927856|loss_dict:{}\n",
      "Evaluate 81|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:82|loss:0.16155435144901276|loss_dict:{}\n",
      "Evaluate 82|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9380471380471381|IOU:[0.83333333 1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:83|loss:0.15743936598300934|loss_dict:{}\n",
      "Evaluate 83|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9592592592592593|IOU:[0.83333333 1.         1.         1.         1.         1.\n",
      " 0.9        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:84|loss:0.15015800297260284|loss_dict:{}\n",
      "Evaluate 84|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:85|loss:0.14365345239639282|loss_dict:{}\n",
      "Evaluate 85|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:86|loss:0.1380869597196579|loss_dict:{}\n",
      "Evaluate 86|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:87|loss:0.13136214017868042|loss_dict:{}\n",
      "Evaluate 87|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9592592592592593|IOU:[0.83333333 1.         1.         1.         1.         1.\n",
      " 0.9        0.9        1.        ]|ACC:[1.  1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:88|loss:0.12659430503845215|loss_dict:{}\n",
      "Evaluate 88|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:89|loss:0.12150336056947708|loss_dict:{}\n",
      "Evaluate 89|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:90|loss:0.11565794795751572|loss_dict:{}\n",
      "Evaluate 90|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9196969696969697|IOU:[0.75       1.         0.90909091 1.         1.         1.\n",
      " 0.8        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.8 0.9 1. ]\n",
      "Iter:91|loss:0.11122295260429382|loss_dict:{}\n",
      "Evaluate 91|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:92|loss:0.10612944513559341|loss_dict:{}\n",
      "Evaluate 92|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:93|loss:0.10132408142089844|loss_dict:{}\n",
      "Evaluate 93|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:94|loss:0.09783864766359329|loss_dict:{}\n",
      "Evaluate 94|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:95|loss:0.09266966581344604|loss_dict:{}\n",
      "Evaluate 95|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:96|loss:0.0893898457288742|loss_dict:{}\n",
      "Evaluate 96|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:97|loss:0.0850927010178566|loss_dict:{}\n",
      "Evaluate 97|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:98|loss:0.08127688616514206|loss_dict:{}\n",
      "Evaluate 98|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:99|loss:0.07802437990903854|loss_dict:{}\n",
      "Evaluate 99|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:100|loss:0.07412628084421158|loss_dict:{}\n",
      "Evaluate 100|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:101|loss:0.0712280347943306|loss_dict:{}\n",
      "Evaluate 101|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:102|loss:0.0687665194272995|loss_dict:{}\n",
      "Evaluate 102|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:103|loss:0.06605313718318939|loss_dict:{}\n",
      "Evaluate 103|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:104|loss:0.06512615084648132|loss_dict:{}\n",
      "Evaluate 104|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:105|loss:0.063600555062294|loss_dict:{}\n",
      "Evaluate 105|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:106|loss:0.0632920041680336|loss_dict:{}\n",
      "Evaluate 106|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9222222222222223|IOU:[0.75 1.   0.9  1.   1.   1.   0.9  0.75 1.  ]|ACC:[0.9 1.  0.9 1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:107|loss:0.05653725191950798|loss_dict:{}\n",
      "Evaluate 107|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:108|loss:0.05173197761178017|loss_dict:{}\n",
      "Evaluate 108|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:109|loss:0.05112740397453308|loss_dict:{}\n",
      "Evaluate 109|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:110|loss:0.048654112964868546|loss_dict:{}\n",
      "Evaluate 110|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:111|loss:0.045389898121356964|loss_dict:{}\n",
      "Evaluate 111|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:112|loss:0.043874118477106094|loss_dict:{}\n",
      "Evaluate 112|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9222222222222223|IOU:[0.75 1.   0.9  1.   1.   1.   0.9  0.75 1.  ]|ACC:[0.9 1.  0.9 1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:113|loss:0.04201069846749306|loss_dict:{}\n",
      "Evaluate 113|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:114|loss:0.03935655206441879|loss_dict:{}\n",
      "Evaluate 114|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:115|loss:0.03839769959449768|loss_dict:{}\n",
      "Evaluate 115|OA:0.9555555555555556|MACC:0.9555555555555555|Kappa:0.9500000000000002|MIOU:0.9222222222222223|IOU:[0.75 1.   0.9  1.   1.   1.   0.9  0.75 1.  ]|ACC:[0.9 1.  0.9 1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:116|loss:0.036194976419210434|loss_dict:{}\n",
      "Evaluate 116|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:117|loss:0.034793831408023834|loss_dict:{}\n",
      "Evaluate 117|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:118|loss:0.033440567553043365|loss_dict:{}\n",
      "Evaluate 118|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:119|loss:0.031814172863960266|loss_dict:{}\n",
      "Evaluate 119|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.75       1.         1.         1.         1.         1.\n",
      " 0.9        0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  0.9 0.9 1. ]\n",
      "Iter:120|loss:0.030811049044132233|loss_dict:{}\n",
      "Evaluate 120|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:121|loss:0.029166312888264656|loss_dict:{}\n",
      "Evaluate 121|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:122|loss:0.028271855786442757|loss_dict:{}\n",
      "Evaluate 122|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:123|loss:0.026849696412682533|loss_dict:{}\n",
      "Evaluate 123|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:124|loss:0.025948457419872284|loss_dict:{}\n",
      "Evaluate 124|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:125|loss:0.02488931640982628|loss_dict:{}\n",
      "Evaluate 125|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:126|loss:0.023836802691221237|loss_dict:{}\n",
      "Evaluate 126|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:127|loss:0.023006878793239594|loss_dict:{}\n",
      "Evaluate 127|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:128|loss:0.022035952657461166|loss_dict:{}\n",
      "Evaluate 128|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:129|loss:0.02123056724667549|loss_dict:{}\n",
      "Evaluate 129|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:130|loss:0.02039998769760132|loss_dict:{}\n",
      "Evaluate 130|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:131|loss:0.019722256809473038|loss_dict:{}\n",
      "Evaluate 131|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:132|loss:0.018875950947403908|loss_dict:{}\n",
      "Evaluate 132|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:133|loss:0.018292536959052086|loss_dict:{}\n",
      "Evaluate 133|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:134|loss:0.017573373392224312|loss_dict:{}\n",
      "Evaluate 134|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:135|loss:0.016996517777442932|loss_dict:{}\n",
      "Evaluate 135|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:136|loss:0.016372429206967354|loss_dict:{}\n",
      "Evaluate 136|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:137|loss:0.01587037369608879|loss_dict:{}\n",
      "Evaluate 137|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:138|loss:0.01525703351944685|loss_dict:{}\n",
      "Evaluate 138|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:139|loss:0.014823159202933311|loss_dict:{}\n",
      "Evaluate 139|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:140|loss:0.014303331263363361|loss_dict:{}\n",
      "Evaluate 140|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:141|loss:0.013848109170794487|loss_dict:{}\n",
      "Evaluate 141|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:142|loss:0.013409492559731007|loss_dict:{}\n",
      "Evaluate 142|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:143|loss:0.013008804991841316|loss_dict:{}\n",
      "Evaluate 143|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:144|loss:0.012572938576340675|loss_dict:{}\n",
      "Evaluate 144|OA:0.9777777777777777|MACC:0.9777777777777779|Kappa:0.9750000000000001|MIOU:0.9595959595959597|IOU:[0.81818182 1.         1.         1.         1.         1.\n",
      " 1.         0.81818182 1.        ]|ACC:[0.9 1.  1.  1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:145|loss:0.012239702045917511|loss_dict:{}\n",
      "Evaluate 145|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:146|loss:0.011836014688014984|loss_dict:{}\n",
      "Evaluate 146|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:147|loss:0.011505299247801304|loss_dict:{}\n",
      "Evaluate 147|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:148|loss:0.011177255772054195|loss_dict:{}\n",
      "Evaluate 148|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:149|loss:0.010850475169718266|loss_dict:{}\n",
      "Evaluate 149|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:150|loss:0.01054567564278841|loss_dict:{}\n",
      "Evaluate 150|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:151|loss:0.010267998091876507|loss_dict:{}\n",
      "Evaluate 151|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:152|loss:0.009972001425921917|loss_dict:{}\n",
      "Evaluate 152|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:153|loss:0.009716905653476715|loss_dict:{}\n",
      "Evaluate 153|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:154|loss:0.009458338841795921|loss_dict:{}\n",
      "Evaluate 154|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:155|loss:0.00920859444886446|loss_dict:{}\n",
      "Evaluate 155|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:156|loss:0.008982585743069649|loss_dict:{}\n",
      "Evaluate 156|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:157|loss:0.008744160644710064|loss_dict:{}\n",
      "Evaluate 157|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:158|loss:0.008535750210285187|loss_dict:{}\n",
      "Evaluate 158|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:159|loss:0.008321343921124935|loss_dict:{}\n",
      "Evaluate 159|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:160|loss:0.008120204322040081|loss_dict:{}\n",
      "Evaluate 160|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:161|loss:0.00792811531573534|loss_dict:{}\n",
      "Evaluate 161|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:162|loss:0.00774082588031888|loss_dict:{}\n",
      "Evaluate 162|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:163|loss:0.007560312282294035|loss_dict:{}\n",
      "Evaluate 163|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:164|loss:0.007388085126876831|loss_dict:{}\n",
      "Evaluate 164|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:165|loss:0.007218585815280676|loss_dict:{}\n",
      "Evaluate 165|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:166|loss:0.007058875169605017|loss_dict:{}\n",
      "Evaluate 166|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:167|loss:0.0069033121690154076|loss_dict:{}\n",
      "Evaluate 167|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:168|loss:0.006751523818820715|loss_dict:{}\n",
      "Evaluate 168|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:169|loss:0.006608397699892521|loss_dict:{}\n",
      "Evaluate 169|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:170|loss:0.006466197781264782|loss_dict:{}\n",
      "Evaluate 170|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:171|loss:0.00633156206458807|loss_dict:{}\n",
      "Evaluate 171|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:172|loss:0.006199011579155922|loss_dict:{}\n",
      "Evaluate 172|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:173|loss:0.006072331219911575|loss_dict:{}\n",
      "Evaluate 173|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:174|loss:0.005950074642896652|loss_dict:{}\n",
      "Evaluate 174|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:175|loss:0.005829154048115015|loss_dict:{}\n",
      "Evaluate 175|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:176|loss:0.005715778097510338|loss_dict:{}\n",
      "Evaluate 176|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:177|loss:0.005602535326033831|loss_dict:{}\n",
      "Evaluate 177|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:178|loss:0.0054946523159742355|loss_dict:{}\n",
      "Evaluate 178|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:179|loss:0.005388266406953335|loss_dict:{}\n",
      "Evaluate 179|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:180|loss:0.005286399740725756|loss_dict:{}\n",
      "Evaluate 180|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:181|loss:0.0051877438090741634|loss_dict:{}\n",
      "Evaluate 181|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:182|loss:0.005090590100735426|loss_dict:{}\n",
      "Evaluate 182|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:183|loss:0.004998682998120785|loss_dict:{}\n",
      "Evaluate 183|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:184|loss:0.004906889982521534|loss_dict:{}\n",
      "Evaluate 184|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:185|loss:0.004818541929125786|loss_dict:{}\n",
      "Evaluate 185|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:186|loss:0.004733365029096603|loss_dict:{}\n",
      "Evaluate 186|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:187|loss:0.004649762529879808|loss_dict:{}\n",
      "Evaluate 187|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:188|loss:0.004568260163068771|loss_dict:{}\n",
      "Evaluate 188|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:189|loss:0.004489860963076353|loss_dict:{}\n",
      "Evaluate 189|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:190|loss:0.004412888549268246|loss_dict:{}\n",
      "Evaluate 190|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:191|loss:0.004338120110332966|loss_dict:{}\n",
      "Evaluate 191|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:192|loss:0.004265356808900833|loss_dict:{}\n",
      "Evaluate 192|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:193|loss:0.00419461727142334|loss_dict:{}\n",
      "Evaluate 193|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:194|loss:0.004125487525016069|loss_dict:{}\n",
      "Evaluate 194|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:195|loss:0.004058132879436016|loss_dict:{}\n",
      "Evaluate 195|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:196|loss:0.003992859739810228|loss_dict:{}\n",
      "Evaluate 196|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:197|loss:0.003928661346435547|loss_dict:{}\n",
      "Evaluate 197|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:198|loss:0.0038665682077407837|loss_dict:{}\n",
      "Evaluate 198|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "Iter:199|loss:0.0038053402677178383|loss_dict:{}\n",
      "Evaluate 199|OA:0.9666666666666667|MACC:0.9666666666666666|Kappa:0.9625000000000001|MIOU:0.9409090909090909|IOU:[0.81818182 1.         0.9        1.         1.         1.\n",
      " 1.         0.75       1.        ]|ACC:[0.9 1.  0.9 1.  1.  1.  1.  0.9 1. ]\n",
      "\n",
      "\n",
      "====================Starting evaluation for testing set.========================\n",
      "\n",
      "Test 199|OA:0.9667342512259525|MACC:0.958418738269007|Kappa:0.9625000000000001|MIOU:0.9153193034857181|IOU:[0.91606325 0.97453949 0.88956481 0.83944374 0.96097202 0.95761564\n",
      " 0.8957012  0.86047679 0.9434968 ]|ACC:[0.94052496 0.98935999 0.92326372 0.87830688 1.         0.99178192\n",
      " 0.98527132 0.94151565 0.97574421]\n",
      "\n",
      "\n",
      "==================== testing Result on UP========================\n",
      "\n",
      "\n",
      "====================== exp_idx=0 seed=6 learning rate=0.0003 epochs=200 train ratio=0.1 val ratio=0.01 ======================\n",
      "OA=0.9667342512259525\n",
      "AA=0.958418738269007\n",
      "kpp=0.9625000000000001\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if __name__ == '__main__':\n",
    "\n",
    "    work_dir = \"./work_dir\"\n",
    "    setting_name = 'tr{}val{}'.format(str(args.train_samples),str(args.val_samples)) + '_lr{}'.format(str(learning_rate)) + '_{}'.format(args.mamba_type)\n",
    "\n",
    "    if 'spe' in args.mamba_type or args.mamba_type=='both':\n",
    "        setting_name = setting_name + '_token{}'.format(str(args.token_num))\n",
    "\n",
    "    if args.patch_size>1:\n",
    "        setting_name = setting_name + '_P{}'.format(args.patch_size)\n",
    "    setting_name = setting_name + '_HD{}'.format(args.hidden_dim)\n",
    "\n",
    "    dataset_name = data_set_name\n",
    "\n",
    "    exp_name = args.exp_name\n",
    "    if model_flag==0:\n",
    "        save_folder = os.path.join(work_dir, exp_name, net_name, setting_name, dataset_name)\n",
    "    else:\n",
    "        save_folder = os.path.join(work_dir, exp_name, net_name, dataset_name)\n",
    "\n",
    "    if not os.path.exists(save_folder):\n",
    "        os.makedirs(save_folder)\n",
    "        print(\"makedirs {}\".format(save_folder))\n",
    "\n",
    "    save_log_path = os.path.join(save_folder,'train_tr{}_val{}.log'.format(num_list[0],num_list[1]))\n",
    "\n",
    "    torch.cuda.empty_cache()\n",
    "\n",
    "    print(save_folder)\n",
    "\n",
    "    data, gt = data_load_operate.load_data(data_set_name, data_set_path)\n",
    "\n",
    "    height, width, channels = data.shape\n",
    "\n",
    "    gt_reshape = gt.reshape(-1)\n",
    "    height, width, channels = data.shape\n",
    "    img = ImageStretching(data)\n",
    "\n",
    "    class_count = max(np.unique(gt))\n",
    "\n",
    "    flag_list = [1, 0]  # ratio or num\n",
    "    ratio_list = [0.1, 0.01]  # [train_ratio,val_ratio]\n",
    "\n",
    "    loss_func = torch.nn.CrossEntropyLoss(ignore_index=-1)\n",
    "\n",
    "    OA_ALL = []\n",
    "    AA_ALL = []\n",
    "    KPP_ALL = []\n",
    "    EACH_ACC_ALL = []\n",
    "    Train_Time_ALL = []\n",
    "    Test_Time_ALL = []\n",
    "    CLASS_ACC = np.zeros([len(seed_list), class_count])\n",
    "    evaluator = Evaluator(num_class=class_count)\n",
    "\n",
    "    for exp_idx,curr_seed in enumerate(seed_list):\n",
    "        setup_seed(curr_seed)\n",
    "        single_experiment_name = 'run{}_seed{}'.format(str(exp_idx), str(curr_seed))\n",
    "        save_single_experiment_folder = os.path.join(save_folder, single_experiment_name)\n",
    "        if not os.path.exists(save_single_experiment_folder):\n",
    "            os.mkdir(save_single_experiment_folder)\n",
    "        save_vis_folder = os.path.join(save_single_experiment_folder, 'vis')\n",
    "        if not os.path.exists(save_vis_folder):\n",
    "            os.makedirs(save_vis_folder)\n",
    "            print(\"makedirs {}\".format(save_vis_folder))\n",
    "\n",
    "        save_weight_path = os.path.join(save_single_experiment_folder, \"best_tr{}_val{}.pth\".format(num_list[0], num_list[1]))\n",
    "        results_save_path = os.path.join(save_single_experiment_folder, 'result_tr{}_val{}.txt'.format(num_list[0], num_list[1]))\n",
    "        predict_save_path = os.path.join(save_single_experiment_folder, 'pred_vis_tr{}_val{}.png'.format(num_list[0], num_list[1]))\n",
    "        gt_save_path = os.path.join(save_single_experiment_folder, 'gt_vis_tr{}_val{}.png'.format(num_list[0], num_list[1]))\n",
    "\n",
    "        train_data_index, val_data_index, test_data_index, all_data_index = data_load_operate.sampling(ratio_list,\n",
    "                                                                                                       num_list,\n",
    "                                                                                                       gt_reshape,\n",
    "                                                                                                       class_count,\n",
    "                                                                                                       flag_list[0])\n",
    "        index = (train_data_index, val_data_index, test_data_index)\n",
    "        train_label, val_label, test_label = data_load_operate.generate_image_iter(data, height, width, gt_reshape, index)\n",
    "\n",
    "        # build Model\n",
    "        net = CommonTemplate(net_name=net_name,in_channels=channels,num_classes=class_count,height=height,width=width,args=args)\n",
    "\n",
    "\n",
    "        x = transform(np.array(img))\n",
    "        x = x.unsqueeze(0).float().to(device)\n",
    "\n",
    "        train_label = train_label.to(device)\n",
    "        test_label = test_label.to(device)\n",
    "        val_label = val_label.to(device)\n",
    "\n",
    "        net.to(device)\n",
    "\n",
    "        train_loss_list = [100]\n",
    "        train_acc_list = [0]\n",
    "        val_loss_list = [100]\n",
    "        val_acc_list = [0]\n",
    "\n",
    "        optimizer = torch.optim.Adam(net.parameters(),lr=learning_rate)\n",
    "        \n",
    "        best_loss = 99999\n",
    "\n",
    "        tic1 = time.perf_counter()\n",
    "        best_val_acc = 0\n",
    "      \n",
    "\n",
    "        for epoch in range(max_epoch):\n",
    "            y_train = train_label.unsqueeze(0)\n",
    "            train_acc_sum, trained_samples_counter = 0.0, 0\n",
    "            batch_counter, train_loss_sum = 0, 0\n",
    "            time_epoch = time.time()\n",
    "            loss_dict = {}\n",
    "        \n",
    "            net.train()\n",
    " \n",
    "            \n",
    "            data_dict = net(x)\n",
    "            y_pred = data_dict['logits']\n",
    "            ls = head_loss(loss_func,y_pred, y_train.long())\n",
    "\n",
    "            logits_dict = data_dict['logits_dict']\n",
    "            optimizer.zero_grad()\n",
    "            ls.backward()\n",
    "            optimizer.step()\n",
    "            print('Iter:{}|loss:{}|loss_dict:{}'.format(epoch, ls.detach().cpu().numpy(), loss_dict))\n",
    "               \n",
    "            torch.cuda.empty_cache()\n",
    "            # evaluate stage\n",
    "            net.eval()\n",
    "            with torch.no_grad():\n",
    "                evaluator.reset()\n",
    "                val_data_dict = net(x)\n",
    "                output_val = val_data_dict['logits']\n",
    "                y_val = val_label.unsqueeze(0)\n",
    "                seg_logits = resize(input=output_val,\n",
    "                                    size=y_val.shape[1:],\n",
    "                                    mode='bilinear',\n",
    "                                    align_corners=True)\n",
    "                predict = torch.argmax(seg_logits,dim=1).cpu().numpy()\n",
    "                Y_val_np = val_label.cpu().numpy()\n",
    "                Y_val_255 = np.where(Y_val_np==-1,255,Y_val_np)\n",
    "                evaluator.add_batch(np.expand_dims(Y_val_255,axis=0),predict)\n",
    "                OA = evaluator.Pixel_Accuracy()\n",
    "                mIOU, IOU = evaluator.Mean_Intersection_over_Union()\n",
    "                mAcc, Acc = evaluator.Pixel_Accuracy_Class()\n",
    "                Kappa = evaluator.Kappa()\n",
    "                print('Evaluate {}|OA:{}|MACC:{}|Kappa:{}|MIOU:{}|IOU:{}|ACC:{}'.format(epoch, OA,mAcc,Kappa,mIOU,IOU,Acc))\n",
    "                # save weight\n",
    "                if OA>=best_val_acc:\n",
    "                    best_epoch = epoch + 1\n",
    "                    best_val_acc = OA\n",
    "                    torch.save(net.state_dict(), save_weight_path)\n",
    "                if (epoch+1)%50==0:\n",
    "                    save_single_predict_path = os.path.join(save_vis_folder,'predict_{}.png'.format(str(epoch+1)))\n",
    "                    save_single_gt_path = os.path.join(save_vis_folder,'gt.png')\n",
    "                    vis_a_image(gt,predict,save_single_predict_path, save_single_gt_path)\n",
    "\n",
    "\n",
    "            torch.cuda.empty_cache()\n",
    "\n",
    "\n",
    "        print(\"\\n\\n====================Starting evaluation for testing set.========================\\n\")\n",
    "        pred_test = []\n",
    "\n",
    "        load_weight_path = save_weight_path\n",
    "        net.update_params = None\n",
    "   \n",
    "        best_net = CommonTemplate(net_name=net_name, in_channels=channels, num_classes=class_count, height=height,\n",
    "                             width=width, args=args)\n",
    "    \n",
    "        best_net.to(device)\n",
    "        best_net.load_state_dict(torch.load(load_weight_path))\n",
    "        best_net.eval()\n",
    "        test_evaluator = Evaluator(num_class=class_count)\n",
    "        with torch.no_grad():\n",
    "            test_evaluator.reset()\n",
    "            test_data_dict = best_net(x)\n",
    "            output_test = test_data_dict['logits']\n",
    "            # output_test = best_net(x)\n",
    "            y_test = test_label.unsqueeze(0)\n",
    "            seg_logits_test = resize(input=output_test,\n",
    "                                size=y_test.shape[1:],\n",
    "                                mode='bilinear',\n",
    "                                align_corners=True)\n",
    "            predict_test = torch.argmax(seg_logits_test, dim=1).cpu().numpy()\n",
    "            Y_test_np = test_label.cpu().numpy()\n",
    "            Y_test_255 = np.where(Y_test_np == -1, 255, Y_test_np)\n",
    "            test_evaluator.add_batch(np.expand_dims(Y_test_255, axis=0), predict_test)\n",
    "            OA_test = test_evaluator.Pixel_Accuracy()\n",
    "            mIOU_test, IOU_test = test_evaluator.Mean_Intersection_over_Union()\n",
    "            mAcc_test, Acc_test = test_evaluator.Pixel_Accuracy_Class()\n",
    "            Kappa_test = evaluator.Kappa()\n",
    "            print('Test {}|OA:{}|MACC:{}|Kappa:{}|MIOU:{}|IOU:{}|ACC:{}'.format(epoch, OA_test, mAcc_test, Kappa_test, mIOU_test, IOU_test,\n",
    "                                                                                    Acc_test))\n",
    "            \n",
    "        # Output infors\n",
    "        print(\"\\n\\n==================== testing Result on {}========================\\n\".format(dataset_name))\n",
    "        str_results = '\\n======================' \\\n",
    "                      + \" exp_idx=\" + str(exp_idx) \\\n",
    "                      + \" seed=\" + str(curr_seed) \\\n",
    "                      + \" learning rate=\" + str(learning_rate) \\\n",
    "                      + \" epochs=\" + str(max_epoch) \\\n",
    "                      + \" train ratio=\" + str(ratio_list[0]) \\\n",
    "                      + \" val ratio=\" + str(ratio_list[1]) \\\n",
    "                      + \" ======================\" \\\n",
    "                      + \"\\nOA=\" + str(OA_test) \\\n",
    "                      + \"\\nAA=\" + str(mAcc_test) \\\n",
    "                      + '\\nkpp=' + str(Kappa_test) + \"\\n\"\n",
    "        print(str_results)\n",
    "        vis_a_image(gt, predict_test, predict_save_path, gt_save_path)\n",
    "        predict_img = plt.imread(predict_save_path)\n",
    "        plt.imshow(predict_img)\n",
    "        torch.cuda.empty_cache()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "mamba",
   "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": "3.9.19"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}