{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "08615cee",
   "metadata": {},
   "source": [
    "# 03 — ResNet50 Transfer Learning\n",
    "\n",
    "Fine-tune a pretrained ResNet50 (ImageNet) for cattle breed classification.\n",
    "\n",
    "### Two-Phase Training\n",
    "1. **Phase 1** (10 epochs): Freeze backbone, train only the classifier head\n",
    "2. **Phase 2** (20 epochs): Unfreeze `layer3` + `layer4`, fine-tune with lower LR\n",
    "\n",
    "### Expected Output\n",
    "- Best checkpoint → `ml/artifacts/checkpoints/resnet_best.pth`\n",
    "- Training curves, confusion matrix, per-class F1\n",
    "- Classification report and training summary JSON"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "457f79bb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Project root: /Users/ajitkumarsingh/Desktop/desktop/cattle-breed-classifier-webapp\n"
     ]
    }
   ],
   "source": [
    "import sys, os\n",
    "from pathlib import Path\n",
    "\n",
    "PROJECT_ROOT = Path(os.getcwd()).resolve()\n",
    "if 'notebooks' in str(PROJECT_ROOT):\n",
    "    PROJECT_ROOT = PROJECT_ROOT.parent.parent\n",
    "sys.path.insert(0, str(PROJECT_ROOT))\n",
    "print(f'Project root: {PROJECT_ROOT}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ae58ea25",
   "metadata": {},
   "outputs": [
    {
     "ename": "InterruptedError",
     "evalue": "[Errno 4] Interrupted system call",
     "output_type": "error",
     "traceback": [
      "\u001b[31m---------------------------------------------------------------------------\u001b[39m",
      "\u001b[31mInterruptedError\u001b[39m                          Traceback (most recent call last)",
      "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[2]\u001b[39m\u001b[32m, line 4\u001b[39m\n\u001b[32m      1\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.utils.seed \u001b[38;5;28;01mimport\u001b[39;00m set_seed\n\u001b[32m      2\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.utils.device \u001b[38;5;28;01mimport\u001b[39;00m get_device\n\u001b[32m      3\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.utils.io \u001b[38;5;28;01mimport\u001b[39;00m load_config\n\u001b[32m----> \u001b[39m\u001b[32m4\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.utils.manifests \u001b[38;5;28;01mimport\u001b[39;00m create_dataloaders\n\u001b[32m      5\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.data.transforms \u001b[38;5;28;01mimport\u001b[39;00m get_train_transforms, get_eval_transforms\n\u001b[32m      6\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.models.resnet \u001b[38;5;28;01mimport\u001b[39;00m CattleResNet\n\u001b[32m      7\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m ml.src.training.trainer \u001b[38;5;28;01mimport\u001b[39;00m Trainer\n",
      "\u001b[36mFile \u001b[39m\u001b[32m~/Desktop/desktop/cattle-breed-classifier-webapp/ml/src/utils/manifests.py:10\u001b[39m\n\u001b[32m      7\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtyping\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m Optional\n\u001b[32m      9\u001b[39m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mpandas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mpd\u001b[39;00m\n\u001b[32m---> \u001b[39m\u001b[32m10\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mPIL\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m Image\n\u001b[32m     11\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorch\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mutils\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mdata\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m Dataset, DataLoader\n\u001b[32m     12\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorchvision\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m transforms\n",
      "\u001b[36mFile \u001b[39m\u001b[32m<frozen importlib._bootstrap>:1360\u001b[39m, in \u001b[36m_find_and_load\u001b[39m\u001b[34m(name, import_)\u001b[39m\n",
      "\u001b[36mFile \u001b[39m\u001b[32m<frozen importlib._bootstrap>:1322\u001b[39m, in \u001b[36m_find_and_load_unlocked\u001b[39m\u001b[34m(name, import_)\u001b[39m\n",
      "\u001b[36mFile \u001b[39m\u001b[32m<frozen importlib._bootstrap>:1262\u001b[39m, in \u001b[36m_find_spec\u001b[39m\u001b[34m(name, path, target)\u001b[39m\n",
      "\u001b[36mFile \u001b[39m\u001b[32m<frozen importlib._bootstrap_external>:1532\u001b[39m, in \u001b[36mPathFinder.find_spec\u001b[39m\u001b[34m(cls, fullname, path, target)\u001b[39m\n",
      "\u001b[36mFile \u001b[39m\u001b[32m<frozen importlib._bootstrap_external>:1504\u001b[39m, in \u001b[36mPathFinder._get_spec\u001b[39m\u001b[34m(cls, fullname, path, target)\u001b[39m\n",
      "\u001b[36mFile \u001b[39m\u001b[32m<frozen importlib._bootstrap_external>:1484\u001b[39m, in \u001b[36mPathFinder._path_importer_cache\u001b[39m\u001b[34m(cls, path)\u001b[39m\n",
      "\u001b[31mInterruptedError\u001b[39m: [Errno 4] Interrupted system call"
     ]
    }
   ],
   "source": [
    "from ml.src.utils.seed import set_seed\n",
    "from ml.src.utils.device import get_device\n",
    "from ml.src.utils.io import load_config\n",
    "from ml.src.utils.manifests import create_dataloaders\n",
    "from ml.src.data.transforms import get_train_transforms, get_eval_transforms\n",
    "from ml.src.models.resnet import CattleResNet\n",
    "from ml.src.training.trainer import Trainer\n",
    "\n",
    "set_seed(42)\n",
    "device = get_device()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4d85602",
   "metadata": {},
   "source": [
    "## 1. Load Config & Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "44de86e3",
   "metadata": {},
   "outputs": [],
   "source": [
    "config = load_config('resnet', PROJECT_ROOT / 'ml' / 'configs')\n",
    "config['num_classes'] = 26\n",
    "\n",
    "img_size = config['image']['size']\n",
    "batch_size = config['training']['batch_size']\n",
    "arch = config['model']['architecture']\n",
    "\n",
    "print(f'Backbone: {arch[\"backbone\"]}')\n",
    "print(f'Pretrained: {arch[\"pretrained\"]}')\n",
    "print(f'Freeze backbone: {arch[\"freeze_backbone\"]}')\n",
    "print(f'Unfreeze after: {arch[\"unfreeze_after_epochs\"]} epochs')\n",
    "print(f'Unfreeze layers: {arch[\"unfreeze_layers\"]}')\n",
    "print(f'Epochs: {config[\"training\"][\"num_epochs\"]}')\n",
    "print(f'LR: {config[\"training\"][\"learning_rate\"]}  Fine-tune LR: {config[\"training\"][\"fine_tune_lr\"]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a2807dbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_transforms = get_train_transforms(img_size=img_size)\n",
    "eval_transforms = get_eval_transforms(img_size=img_size)\n",
    "\n",
    "dataloaders = create_dataloaders(\n",
    "    manifests_dir=PROJECT_ROOT / 'ml' / 'artifacts' / 'manifests',\n",
    "    data_root=PROJECT_ROOT / 'Cattle_Resized',\n",
    "    train_transform=train_transforms,\n",
    "    eval_transform=eval_transforms,\n",
    "    batch_size=batch_size,\n",
    "    num_workers=config['data'].get('num_workers', 4),\n",
    ")\n",
    "\n",
    "class_names = dataloaders['train'].dataset.classes\n",
    "print(f'Classes: {len(class_names)}')\n",
    "print(f'Train: {len(dataloaders[\"train\"].dataset)} | Val: {len(dataloaders[\"val\"].dataset)} | Test: {len(dataloaders[\"test\"].dataset)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62801587",
   "metadata": {},
   "source": [
    "## 2. Create Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "60fd0584",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = CattleResNet.from_config(config)\n",
    "\n",
    "# Count trainable vs frozen params\n",
    "trainable = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
    "total = sum(p.numel() for p in model.parameters())\n",
    "print(f'Total parameters:     {total:,}')\n",
    "print(f'Trainable (Phase 1):  {trainable:,} ({100*trainable/total:.1f}%)')\n",
    "print(f'Frozen:               {total - trainable:,}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a783655",
   "metadata": {},
   "source": [
    "## 3. Two-Phase Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "077006b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "trainer = Trainer(\n",
    "    model=model,\n",
    "    config=config,\n",
    "    dataloaders=dataloaders,\n",
    "    class_names=class_names,\n",
    "    device=device,\n",
    "    model_name='resnet',\n",
    ")\n",
    "\n",
    "# Two-phase training: frozen backbone → partial fine-tuning\n",
    "training_summary = trainer.train_with_phase_switch(\n",
    "    phase1_epochs=arch['unfreeze_after_epochs'],\n",
    "    phase2_lr=config['training']['fine_tune_lr'],\n",
    "    unfreeze_layers=arch['unfreeze_layers'],\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c62c99fb",
   "metadata": {},
   "source": [
    "## 4. Evaluate on Test Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b3c413c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_results = trainer.evaluate(split='test')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "98c3ee02",
   "metadata": {},
   "source": [
    "## 5. Save Artifacts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97f96e44",
   "metadata": {},
   "outputs": [],
   "source": [
    "trainer.save_artifacts()\n",
    "\n",
    "print('\\n=== ResNet50 Transfer Learning Summary ===')\n",
    "print(f'Best Val Loss:  {training_summary[\"best_val_loss\"]:.4f}')\n",
    "print(f'Best Val Acc:   {training_summary[\"best_val_accuracy\"]:.4f}')\n",
    "print(f'Test Accuracy:  {test_results[\"metrics\"][\"accuracy\"]:.4f}')\n",
    "print(f'Test Macro F1:  {test_results[\"metrics\"][\"macro_f1\"]:.4f}')\n",
    "print(f'Latency:        {test_results[\"latency\"][\"avg_ms\"]:.2f} ms')\n",
    "print(f'Model Size:     {test_results[\"model_size_mb\"]:.2f} MB')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5362b63",
   "metadata": {},
   "source": [
    "## 6. Classification Report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f3951b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(test_results['metrics']['classification_report'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "764286aa",
   "metadata": {},
   "source": [
    "---\n",
    "**✅ ResNet50 Transfer Learning complete.** Proceed to `04_vit_transfer_learning.ipynb`."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Cattle Classifier (Python 3.12)",
   "language": "python",
   "name": "cattle-classifier"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}