lab-10_4_2_ImageFolder_2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 10-4 ImageFolder (2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "\n",
    "torch.manual_seed(777)\n",
    "if device =='cuda':\n",
    "    torch.cuda.manual_seed_all(777)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trans = transforms.Compose([\n",
    "    transforms.ToTensor()\n",
    "])\n",
    "\n",
    "train_data = torchvision.datasets.ImageFolder(root='./custom_data/train_data', transform=trans)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_loader = DataLoader(dataset = train_data, batch_size = 8, shuffle = True, num_workers=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(CNN, self).__init__()\n",
    "        self.layer1 = nn.Sequential(\n",
    "            nn.Conv2d(3,6,5),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(2),\n",
    "        )\n",
    "        self.layer2 = nn.Sequential(\n",
    "            nn.Conv2d(6,16,5),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(2),\n",
    "        )\n",
    "        self.layer3 = nn.Sequential(\n",
    "            nn.Linear(16*13*29, 120),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(120,2)\n",
    "        )\n",
    "        \n",
    "    def forward(self, x):\n",
    "        out = self.layer1(x)\n",
    "        out = self.layer2(out)\n",
    "        out = out.view(out.shape[0], -1)\n",
    "        out = self.layer3(out)\n",
    "        return out\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#testing \n",
    "net = CNN().to(device)\n",
    "test_input = (torch.Tensor(3,3,64,128)).to(device)\n",
    "test_out = net(test_input)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = optim.Adam(net.parameters(), lr=0.00005)\n",
    "loss_func = nn.CrossEntropyLoss().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_batch = len(data_loader)\n",
    "\n",
    "epochs = 7\n",
    "for epoch in range(epochs):\n",
    "    avg_cost = 0.0\n",
    "    for num, data in enumerate(data_loader):\n",
    "        imgs, labels = data\n",
    "        imgs = imgs.to(device)\n",
    "        labels = labels.to(device)\n",
    "        optimizer.zero_grad()\n",
    "        out = net(imgs)\n",
    "        loss = loss_func(out, labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        avg_cost += loss / total_batch\n",
    "        \n",
    "    print('[Epoch:{}] cost = {}'.format(epoch+1, avg_cost))\n",
    "print('Learning Finished!')   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(net.state_dict(), \"./model/model.pth\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_net = CNN().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_net.load_state_dict(torch.load('./model/model.pth'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(net.layer1[0])\n",
    "print(new_net.layer1[0])\n",
    "\n",
    "print(net.layer1[0].weight[0][0][0])\n",
    "print(new_net.layer1[0].weight[0][0][0])\n",
    "\n",
    "net.layer1[0].weight[0] == new_net.layer1[0].weight[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trans=torchvision.transforms.Compose([\n",
    "    transforms.Resize((64,128)),\n",
    "    transforms.ToTensor()\n",
    "])\n",
    "test_data = torchvision.datasets.ImageFolder(root='./custom_data/test_data', transform=trans)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_set = DataLoader(dataset = test_data, batch_size = len(test_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "with torch.no_grad():\n",
    "    for num, data in enumerate(test_set):\n",
    "        imgs, label = data\n",
    "        imgs = imgs.to(device)\n",
    "        label = label.to(device)\n",
    "        \n",
    "        prediction = net(imgs)\n",
    "        \n",
    "        correct_prediction = torch.argmax(prediction, 1) == label\n",
    "        \n",
    "        accuracy = correct_prediction.float().mean()\n",
    "        print('Accuracy:', accuracy.item())"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}