train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "import torchvision\n",
    "import torch.nn as nn\n",
    "import torchvision.models as models\n",
    "import os\n",
    "from PIL import Image\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import trange"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/Users/anthony/carNoCar'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "PATH = os.path.abspath(os.getcwd())\n",
    "PATH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomDset(Dataset):\n",
    "    def __init__(self, root_dir, transform):\n",
    "        self.root_dir = root_dir\n",
    "        self.images = os.listdir(root_dir)\n",
    "        self.transform = transform\n",
    "    \n",
    "    def __len__(self):\n",
    "        return len(os.listdir(self.root_dir))\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        img_loc = os.path.join(self.root_dir,self.images[index])\n",
    "        img = Image.open(img_loc).convert(\"RGB\")\n",
    "        label = 1 if os.path.basename(self.root_dir) == 'cars' else 0\n",
    "        label = torch.tensor(label).to(device)\n",
    "        \n",
    "        if self.transform is not None:\n",
    "            img = self.transform(img).to(device)\n",
    "#             img = img.reshape((3,150,150))\n",
    "    \n",
    "        return img,label\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "tfms = torchvision.transforms.Compose([\n",
    "    torchvision.transforms.Resize((150,150)),\n",
    "    torchvision.transforms.ToTensor()])\n",
    "carDset = CustomDset(PATH+'/data/cars', transform=tfms)\n",
    "openDset = CustomDset(PATH+'/data/open', transform=tfms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([3, 150, 150]), tensor(1))"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img,label = carDset.__getitem__(0)\n",
    "img.shape, label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8496, 2124)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "concat_dataset = torch.utils.data.ConcatDataset([carDset,openDset,openDset])\n",
    "lengths = [int(len(concat_dataset)*0.8), int(len(concat_dataset)*0.2)]\n",
    "train_set, val_set = torch.utils.data.random_split(concat_dataset, lengths)\n",
    "train_loader = DataLoader(train_set,batch_size=8,shuffle=True)\n",
    "len(train_set), len(val_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([8, 3, 150, 150]), torch.Size([8]))"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_iter = iter(train_loader)\n",
    "images,labels = train_iter.next()\n",
    "images.size(),labels.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'tensor([1, 1, 1, 1, 1, 0, 1, 0])')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "grid = torchvision.utils.make_grid(images.cpu())\n",
    "plt.imshow(grid.numpy().transpose((1,2,0)))\n",
    "plt.axis('off')\n",
    "plt.title(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "#model\n",
    "class ApdNet(torch.nn.Module):\n",
    "    def __init__(self):\n",
    "        super(ApdNet, self).__init__()\n",
    "        self.layer1 = nn.Sequential(\n",
    "            nn.Conv2d(3, 32, kernel_size=5, stride=1, padding=2),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
    "        self.layer2 = nn.Sequential(\n",
    "            nn.Conv2d(32, 64, kernel_size=5, stride=1, padding=2),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
    "        self.drop_out = nn.Dropout()\n",
    "        self.fc1 = nn.Linear(87616, 1000)\n",
    "        self.fc2 = nn.Linear(1000, 2)\n",
    "    def forward(self,x):\n",
    "        #Max pooling over (2,2) window\n",
    "        out = self.layer1(x)\n",
    "        out = self.layer2(out)\n",
    "        out = out.reshape(out.size(0), -1)\n",
    "        out = self.drop_out(out)\n",
    "        out = self.fc1(out)\n",
    "        out = self.fc2(out)\n",
    "        return out\n",
    "    \n",
    "    def num_flat_features(self,x):\n",
    "        size = x.size()[1:]\n",
    "        num_features = 1\n",
    "        for s in size:\n",
    "            num_features *= s\n",
    "        return num_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<function torchvision.models.resnet.resnet18(pretrained=False, progress=True, **kwargs)>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models.resnet18"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|          | 0/2 [00:00<?, ?it/s]"
     ]
    }
   ],
   "source": [
    "#training\n",
    "model = models.resnet18(pretrained=True)\n",
    "num_ftrs = model.fc.in_features\n",
    "model.fc = nn.Linear(num_ftrs, 2)\n",
    "model.to(device)\n",
    "\n",
    "loss_list, acc_list = [],[]\n",
    "loss_func = nn.CrossEntropyLoss().to(device)\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)\n",
    "num_epochs = 2\n",
    "total_step = len(train_loader)\n",
    "for epoch in (t := trange(num_epochs)):\n",
    "    for i, (images, labels) in enumerate(train_loader):\n",
    "        outputs = model(images)\n",
    "#         label = torch.max(labels,1)[1]\n",
    "\n",
    "        loss = loss_func(outputs,labels)\n",
    "        loss_list.append(loss.item())\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        total = labels.size(0)\n",
    "        _, predicted = torch.max(outputs.data,1)\n",
    "        correct = (predicted == labels).sum().item()\n",
    "        acc_list.append(correct/total)\n",
    "        \n",
    "        t.set_description('Epoch [{}/{}], Step [{},{}], Loss: {:.4f}, Accuracy: {:.2f}%'\n",
    "                .format(epoch+1, num_epochs, i+1, total_step, loss.item(),\n",
    "                       (correct/total)*100))\n",
    "    \n",
    "plt.ylim(-0.1,1.1)\n",
    "plt.plot(loss_list)\n",
    "plt.plot(acc_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# testing validation set\n",
    "model.eval()\n",
    "accuracy_list = []\n",
    "for img, label in val_set:\n",
    "    img = img.unsqueeze(0)\n",
    "    with torch.no_grad():\n",
    "        output = model(img)\n",
    "    predicted = torch.argmax(output)\n",
    "    correct = 1 if predicted == label else 0\n",
    "    accuracy_list.append(correct)\n",
    "\n",
    "acc = sum(accuracy_list)/len(val_set)\n",
    "print(acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.eval()\n",
    "img = Image.open(PATH+'/test_images/no_car.png').convert(\"RGB\")\n",
    "img = torchvision.transforms.functional.resize(img, (150,150))\n",
    "img = torchvision.transforms.functional.to_tensor(img)\n",
    "img = img.unsqueeze(0).to('cuda')\n",
    "print(img.shape)\n",
    "with torch.no_grad():\n",
    "    output = model(img)\n",
    "output\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sm = torch.jit.script(model)\n",
    "sm.save(\"carClassify.pt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model.state_dict(), 'models/bad_model.pt')"
   ]
  },
  {
   "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}