resnet101_era_prediction.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from models import model, train_model\n",
    "from torchvision.models import resnet101, ResNet101_Weights\n",
    "from torchvision.datasets import ImageFolder\n",
    "from sklearn.model_selection import train_test_split\n",
    "from torch.utils.data import Subset\n",
    "import numpy as np\n",
    "from torchvision.transforms import RandomErasing,Compose\n",
    "np.random.seed(0)\n",
    "from torchmetrics import Accuracy\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sn\n",
    "import torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\Anaconda\\envs\\Babylon\\Lib\\site-packages\\torchvision\\models\\_utils.py:135: UserWarning: Using 'weights' as positional parameter(s) is deprecated since 0.13 and may be removed in the future. Please use keyword parameter(s) instead.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "resnet101 = model.Model(resnet101(ResNet101_Weights.IMAGENET1K_V2).to('cuda'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = ImageFolder('all_signs', ResNet101_Weights.IMAGENET1K_V2.transforms())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "targets = data.targets\n",
    "train_indices, test_val_indices = train_test_split(np.arange(len(targets)), stratify=targets, train_size=0.65, random_state=21)\n",
    "train_data = Subset(data, indices=train_indices)\n",
    "val_test_data = Subset(data, indices=test_val_indices)\n",
    "targets = np.array(targets)\n",
    "test_val_targets = targets[test_val_indices]\n",
    "test_indices, val_indices = train_test_split(test_val_indices, stratify=test_val_targets, train_size=0.57, random_state=21)\n",
    "val_data, test_data = Subset(data, indices=val_indices), Subset(data, indices=test_indices)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "idx_to_class = {v: k for k, v in test_data.dataset.class_to_idx.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of train set: 24547\n",
      "Length of val set: 5684\n",
      "Length of test set: 7534\n"
     ]
    }
   ],
   "source": [
    "print(\"Length of train set:\", len(train_data))\n",
    "print(\"Length of val set:\", len(val_data))\n",
    "print(\"Length of test set:\", len(test_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/100 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1 \t\t Training Loss: 0.2380041013826716 \t\t Validation Loss: 0.13726379774155073\n",
      "Validation Loss Decreased(inf--->48.865912) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  2%|▏         | 2/100 [20:34<17:03:10, 626.44s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2 \t\t Training Loss: 0.1393652757961934 \t\t Validation Loss: 0.14180001849439414\n",
      "Epoch 3 \t\t Training Loss: 0.10562620467645896 \t\t Validation Loss: 0.09016515244272004\n",
      "Validation Loss Decreased(48.865912--->32.098794) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  3%|▎         | 3/100 [31:38<17:20:19, 643.50s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4 \t\t Training Loss: 0.09247395708786887 \t\t Validation Loss: 0.08079409181484377\n",
      "Validation Loss Decreased(32.098794--->28.762697) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  5%|▌         | 5/100 [53:54<17:22:35, 658.48s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5 \t\t Training Loss: 0.08175341619794099 \t\t Validation Loss: 0.09394797687667286\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  6%|▌         | 6/100 [1:05:12<17:22:00, 665.11s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 6 \t\t Training Loss: 0.07224017821291404 \t\t Validation Loss: 0.11974252717506685\n",
      "Epoch 7 \t\t Training Loss: 0.065293726098453 \t\t Validation Loss: 0.0755803684116697\n",
      "Validation Loss Decreased(28.762697--->26.906611) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  7%|▋         | 7/100 [1:16:40<17:22:31, 672.60s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8 \t\t Training Loss: 0.055609598811413426 \t\t Validation Loss: 0.06853162143541729\n",
      "Validation Loss Decreased(26.906611--->24.397257) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  8%|▊         | 8/100 [1:28:09<17:19:13, 677.75s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9 \t\t Training Loss: 0.054825101974749206 \t\t Validation Loss: 0.05801498646848885\n",
      "Validation Loss Decreased(24.397257--->20.653335) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 10%|█         | 10/100 [1:51:07<17:05:20, 683.56s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10 \t\t Training Loss: 0.04424726879140105 \t\t Validation Loss: 0.08226275851059275\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 11%|█         | 11/100 [2:02:34<16:55:34, 684.66s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 11 \t\t Training Loss: 0.04354517812810447 \t\t Validation Loss: 0.05979960227733919\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 12%|█▏        | 12/100 [2:15:34<17:26:34, 713.57s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 12 \t\t Training Loss: 0.03908880457148587 \t\t Validation Loss: 0.08627784586803572\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 13%|█▎        | 13/100 [2:37:27<21:38:00, 895.18s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13 \t\t Training Loss: 0.03539086085049307 \t\t Validation Loss: 0.0785983095724476\n",
      "Epoch 14 \t\t Training Loss: 0.032838099297357225 \t\t Validation Loss: 0.057734783869593964\n",
      "Validation Loss Decreased(20.653335--->20.553583) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 14%|█▍        | 14/100 [2:58:36<24:05:07, 1008.23s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 15 \t\t Training Loss: 0.028625817910424983 \t\t Validation Loss: 0.05707966775925128\n",
      "Validation Loss Decreased(20.553583--->20.320362) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 15%|█▌        | 15/100 [3:19:20<25:28:53, 1079.22s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 16 \t\t Training Loss: 0.025076917763708004 \t\t Validation Loss: 0.04854829485128995\n",
      "Validation Loss Decreased(20.320362--->17.283193) \t Saving The Model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 17%|█▋        | 17/100 [4:01:13<26:59:15, 1170.55s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 17 \t\t Training Loss: 0.027280465085873593 \t\t Validation Loss: 0.06974971879817735\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 18%|█▊        | 18/100 [4:25:20<28:33:03, 1253.46s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 18 \t\t Training Loss: 0.02036185465328877 \t\t Validation Loss: 0.06318315832417094\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 19%|█▉        | 19/100 [4:46:37<28:21:50, 1260.62s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 19 \t\t Training Loss: 0.022398559522715755 \t\t Validation Loss: 0.06364461094681936\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 20%|██        | 20/100 [5:07:24<27:55:16, 1256.46s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20 \t\t Training Loss: 0.019448314075889622 \t\t Validation Loss: 0.06893438670798885\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 20%|██        | 20/100 [5:26:42<21:46:48, 980.10s/it] "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 21 \t\t Training Loss: 0.021064298448579582 \t\t Validation Loss: 0.06189335336275588\n",
      "Training stopped\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "train_model.train_model(train_data,val_data , resnet101, f'models/models_weights/resnet101_era.pth', epochs=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "idx_to_class = {v: k for k, v in test_data.dataset.class_to_idx.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_loader = torch.utils.data.DataLoader(test_data,\n",
    "                                             batch_size=16, shuffle=True,\n",
    "                                             num_workers=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resnet101.model.load_state_dict(torch.load('models/models_weights/resnet101_era.pth'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "correct_pred = {classname: 0 for classname in idx_to_class}\n",
    "total_pred = {classname: 0 for classname in idx_to_class}\n",
    "with torch.no_grad():\n",
    "    for data in test_loader:\n",
    "        images, labels = data\n",
    "        images = images.to('cuda')\n",
    "        labels = labels.to('cuda')\n",
    "        outputs = resnet101.model(images)\n",
    "        _, predictions = torch.max(outputs, 1)\n",
    "        for label, prediction in zip(labels, predictions):\n",
    "            if idx_to_class[int(label)] == idx_to_class[(int(prediction))]:\n",
    "                correct_pred[int(label)] += 1\n",
    "            total_pred[int(label)] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy for class: Neo-Assyrian is 96.1 %\n",
      "Accuracy for class: Neo-Babylonian is 98.7 %\n"
     ]
    }
   ],
   "source": [
    "for classname, correct_count in correct_pred.items():\n",
    "    accuracy = 100 * float(correct_count) / total_pred[classname]\n",
    "    print(f'Accuracy for class: {idx_to_class[classname]:5s} is {accuracy:.1f} %')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overall accuracy:  0.9796920626493231\n"
     ]
    }
   ],
   "source": [
    "print(\"Overall accuracy: \",sum(correct_pred.values()) / sum(total_pred.values()))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Babylon",
   "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": "3.11.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}