2-Baseline_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import print_function, division\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torch.optim import lr_scheduler\n",
    "import numpy as np\n",
    "import torchvision\n",
    "from torchvision import datasets, models, transforms\n",
    "import matplotlib.pyplot as plt\n",
    "import time\n",
    "import os\n",
    "import copy\n",
    "\n",
    "plt.ion()   # interactive mode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data augmentation and normalization for training\n",
    "# Just normalization for validation\n",
    "data_transforms = {\n",
    "    'train': transforms.Compose([\n",
    "        transforms.RandomResizedCrop(224),\n",
    "        transforms.RandomHorizontalFlip(),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
    "    ]),\n",
    "    'val': transforms.Compose([\n",
    "        transforms.Resize(256),\n",
    "        transforms.CenterCrop(224),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
    "    ]),\n",
    "}\n",
    "\n",
    "data_dir = '../data/X_small/'\n",
    "image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),\n",
    "                                          data_transforms[x])\n",
    "                  for x in ['train', 'val']}\n",
    "dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,\n",
    "                                             shuffle=True, num_workers=4)\n",
    "              for x in ['train', 'val']}\n",
    "dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'val']}\n",
    "class_names = image_datasets['train'].classes\n",
    "\n",
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'train': 611, 'val': 263}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_sizes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Female', 'Male']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def imshow(inp, title=None):\n",
    "    \"\"\"Imshow for Tensor.\"\"\"\n",
    "    inp = inp.numpy().transpose((1, 2, 0))\n",
    "    mean = np.array([0.485, 0.456, 0.406])\n",
    "    std = np.array([0.229, 0.224, 0.225])\n",
    "    inp = std * inp + mean\n",
    "    inp = np.clip(inp, 0, 1)\n",
    "    plt.imshow(inp)\n",
    "    if title is not None:\n",
    "        plt.title(title)\n",
    "    plt.pause(0.001)  # pause a bit so that plots are updated\n",
    "\n",
    "\n",
    "# Get a batch of training data\n",
    "inputs, classes = next(iter(dataloaders['train']))\n",
    "\n",
    "# Make a grid from batch\n",
    "out = torchvision.utils.make_grid(inputs)\n",
    "\n",
    "imshow(out, title=[class_names[x] for x in classes])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(model, criterion, optimizer, scheduler, num_epochs=25):\n",
    "    since = time.time()\n",
    "\n",
    "    best_model_wts = copy.deepcopy(model.state_dict())\n",
    "    best_acc = 0.0\n",
    "\n",
    "    for epoch in range(num_epochs):\n",
    "        print('Epoch {}/{}'.format(epoch, num_epochs - 1))\n",
    "        print('-' * 10)\n",
    "\n",
    "        # Each epoch has a training and validation phase\n",
    "        for phase in ['train', 'val']:\n",
    "            if phase == 'train':\n",
    "                model.train()  # Set model to training mode\n",
    "            else:\n",
    "                model.eval()   # Set model to evaluate mode\n",
    "\n",
    "            running_loss = 0.0\n",
    "            running_corrects = 0\n",
    "\n",
    "            # Iterate over data.\n",
    "            for inputs, labels in dataloaders[phase]:\n",
    "                inputs = inputs.to(device)\n",
    "                labels = labels.to(device)\n",
    "\n",
    "                # zero the parameter gradients\n",
    "                optimizer.zero_grad()\n",
    "\n",
    "                # forward\n",
    "                # track history if only in train\n",
    "                with torch.set_grad_enabled(phase == 'train'):\n",
    "                    outputs = model(inputs)\n",
    "                    _, preds = torch.max(outputs, 1)\n",
    "                    loss = criterion(outputs, labels)\n",
    "\n",
    "                    # backward + optimize only if in training phase\n",
    "                    if phase == 'train':\n",
    "                        loss.backward()\n",
    "                        optimizer.step()\n",
    "\n",
    "                # statistics\n",
    "                running_loss += loss.item() * inputs.size(0)\n",
    "                running_corrects += torch.sum(preds == labels.data)\n",
    "            if phase == 'train':\n",
    "                scheduler.step()\n",
    "\n",
    "            epoch_loss = running_loss / dataset_sizes[phase]\n",
    "            epoch_acc = running_corrects.double() / dataset_sizes[phase]\n",
    "\n",
    "            print('{} Loss: {:.4f} Acc: {:.4f}'.format(\n",
    "                phase, epoch_loss, epoch_acc))\n",
    "\n",
    "            # deep copy the model\n",
    "            if phase == 'val' and epoch_acc > best_acc:\n",
    "                best_acc = epoch_acc\n",
    "                best_model_wts = copy.deepcopy(model.state_dict())\n",
    "\n",
    "        print()\n",
    "\n",
    "    time_elapsed = time.time() - since\n",
    "    print('Training complete in {:.0f}m {:.0f}s'.format(\n",
    "        time_elapsed // 60, time_elapsed % 60))\n",
    "    print('Best val Acc: {:4f}'.format(best_acc))\n",
    "\n",
    "    # load best model weights\n",
    "    model.load_state_dict(best_model_wts)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize_model(model, num_images=6):\n",
    "    was_training = model.training\n",
    "    model.eval()\n",
    "    images_so_far = 0\n",
    "    fig = plt.figure()\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for i, (inputs, labels) in enumerate(dataloaders['val']):\n",
    "            inputs = inputs.to(device)\n",
    "            labels = labels.to(device)\n",
    "\n",
    "            outputs = model(inputs)\n",
    "            _, preds = torch.max(outputs, 1)\n",
    "\n",
    "            for j in range(inputs.size()[0]):\n",
    "                images_so_far += 1\n",
    "                ax = plt.subplot(num_images//2, 2, images_so_far)\n",
    "                ax.axis('off')\n",
    "                ax.set_title('predicted: {}'.format(class_names[preds[j]]))\n",
    "                imshow(inputs.cpu().data[j])\n",
    "\n",
    "                if images_so_far == num_images:\n",
    "                    model.train(mode=was_training)\n",
    "                    return\n",
    "        model.train(mode=was_training)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Downloading: \"https://download.pytorch.org/models/resnet152-b121ed2d.pth\" to /home/zzl/.cache/torch/hub/checkpoints/resnet152-b121ed2d.pth\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "43a37e945d4f45e8afaef4ced31a45c2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0.00/230M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_ft = models.resnet152(pretrained=True)\n",
    "num_ftrs = model_ft.fc.in_features\n",
    "# Here the size of each output sample is set to 2.\n",
    "# Alternatively, it can be generalized to nn.Linear(num_ftrs, len(class_names)).\n",
    "model_ft.fc = nn.Linear(num_ftrs, 2)\n",
    "\n",
    "model_ft = model_ft.to(device)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "# Observe that all parameters are being optimized\n",
    "optimizer_ft = optim.SGD(model_ft.parameters(), lr=0.001, momentum=0.9)\n",
    "\n",
    "# Decay LR by a factor of 0.1 every 7 epochs\n",
    "exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0/24\n",
      "----------\n",
      "train Loss: 0.8742 Acc: 0.5368\n",
      "val Loss: 0.7510 Acc: 0.5665\n",
      "\n",
      "Epoch 1/24\n",
      "----------\n",
      "train Loss: 0.9534 Acc: 0.5434\n",
      "val Loss: 1.0690 Acc: 0.5437\n",
      "\n",
      "Epoch 2/24\n",
      "----------\n",
      "train Loss: 0.8138 Acc: 0.5859\n",
      "val Loss: 0.9888 Acc: 0.6008\n",
      "\n",
      "Epoch 3/24\n",
      "----------\n",
      "train Loss: 0.7643 Acc: 0.6023\n",
      "val Loss: 0.6551 Acc: 0.7110\n",
      "\n",
      "Epoch 4/24\n",
      "----------\n",
      "train Loss: 0.7113 Acc: 0.5974\n",
      "val Loss: 0.8668 Acc: 0.6502\n",
      "\n",
      "Epoch 5/24\n",
      "----------\n",
      "train Loss: 0.7193 Acc: 0.6268\n",
      "val Loss: 0.4905 Acc: 0.7871\n",
      "\n",
      "Epoch 6/24\n",
      "----------\n"
     ]
    }
   ],
   "source": [
    "model_ft = train_model(model_ft, criterion, optimizer_ft, exp_lr_scheduler,\n",
    "                       num_epochs=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = '../data/X_test/'\n",
    "test_list =['X_a', 'X_b', 'X_c']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test(model, data_dir, test_list):\n",
    "    model.eval()\n",
    "    image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),\n",
    "                                          data_transforms[x]) for x in test_list}\n",
    "    dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,\n",
    "                                                 shuffle=True, num_workers=4)\n",
    "    print(image_datasets)\n",
    "    for phase in test_list:\n",
    "        for inputs, labels in dataloaders[phase]:\n",
    "            outputs = model(inputs)\n",
    "            _, preds = torch.max(outputs, 1)\n",
    "            print(preds)\n",
    "            break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test(model_ft, data_dir, test_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import shutil"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_X_train = pd.read_csv(\"../data/X_train.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def file_splider(df, dest_loc, orgin_loc):\n",
    "    for index, row in df.iterrows():\n",
    "        if row[\"Sex (subj)\"]==1:\n",
    "            try:\n",
    "                origin_dest = orgin_loc + row[\"Image File\"]\n",
    "                dest_dest = dest_loc+\"Female/\"+ row[\"Image File\"]\n",
    "                shutil.copy(origin_dest, dest_dest)\n",
    "            except:\n",
    "                pass\n",
    "        else:\n",
    "            try:\n",
    "                origin_dest = orgin_loc + row[\"Image File\"]\n",
    "                dest_dest = dest_loc+\"Male/\"+ row[\"Image File\"]\n",
    "                shutil.copy(origin_dest, dest_dest)\n",
    "            except:\n",
    "                pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_splider(data_X_train, \"../data/train/\", \"../data/X_train/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}