Simple Autoencoder (linear).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0cbc51b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from six.moves import urllib\n",
    "opener = urllib.request.build_opener()\n",
    "opener.addheaders = [('User-agent', 'Mozilla/5.0')]\n",
    "urllib.request.install_opener(opener)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c5a0287a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import numpy as np\n",
    "from torchvision import datasets\n",
    "import torchvision.transforms as transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6a0ecbc5",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\reza\\anaconda3\\envs\\t7\\lib\\site-packages\\torchvision\\datasets\\mnist.py:498: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at  ..\\torch\\csrc\\utils\\tensor_numpy.cpp:180.)\n",
      "  return torch.from_numpy(parsed.astype(m[2], copy=False)).view(*s)\n"
     ]
    }
   ],
   "source": [
    "transform = transforms.ToTensor()\n",
    "train_data = datasets.MNIST(root='~/.pytorch/MNIST_data/', train=True,download=True, transform=transform)\n",
    "test_data = datasets.MNIST(root='~/.pytorch/MNIST_data/', train=False,download=True, transform=transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "77f5980c",
   "metadata": {},
   "outputs": [],
   "source": [
    "num_workers = 0\n",
    "batch_size = 20\n",
    "\n",
    "train_loader = torch.utils.data.DataLoader(train_data, batch_size=batch_size, num_workers=num_workers)\n",
    "test_loader = torch.utils.data.DataLoader(test_data, batch_size=batch_size, num_workers=num_workers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "e464e642",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1c972015788>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualization\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "    \n",
    "# obtain one batch of training images\n",
    "dataiter = iter(train_loader)\n",
    "images, labels = dataiter.next()\n",
    "images = images.numpy()\n",
    "\n",
    "# get one image from the batch\n",
    "img = np.squeeze(images[0])\n",
    "\n",
    "fig = plt.figure(figsize = (5,5)) \n",
    "ax = fig.add_subplot(111)\n",
    "ax.imshow(img, cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "215f1eb9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Autoencoder(\n",
      "  (fc1): Linear(in_features=784, out_features=32, bias=True)\n",
      "  (fc2): Linear(in_features=32, out_features=784, bias=True)\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class Autoencoder(nn.Module):\n",
    "    def __init__(self, encoding_dim):\n",
    "        super(Autoencoder, self).__init__()\n",
    "        ## encoder ##\n",
    "        # linear layer (784 -> encoding_dim)\n",
    "        self.fc1 = nn.Linear(28 * 28, encoding_dim)   \n",
    "        ## decoder ##\n",
    "        # linear layer (encoding_dim -> input size)\n",
    "        self.fc2 = nn.Linear(encoding_dim, 28*28)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.sigmoid(self.fc2(x))\n",
    "        return x\n",
    "\n",
    "    \n",
    "encoding_dim = 32\n",
    "model = Autoencoder(encoding_dim)\n",
    "print(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "0e8088e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8c1b402e",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\reza\\anaconda3\\envs\\t7\\lib\\site-packages\\torch\\nn\\functional.py:1805: UserWarning: nn.functional.sigmoid is deprecated. Use torch.sigmoid instead.\n",
      "  warnings.warn(\"nn.functional.sigmoid is deprecated. Use torch.sigmoid instead.\")\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch: 1 \tTraining Loss: 0.633195\n",
      "epoch: 2 \tTraining Loss: 0.323455\n",
      "epoch: 3 \tTraining Loss: 0.280142\n",
      "epoch: 4 \tTraining Loss: 0.271142\n",
      "epoch: 5 \tTraining Loss: 0.267526\n",
      "epoch: 6 \tTraining Loss: 0.265394\n",
      "epoch: 7 \tTraining Loss: 0.264026\n",
      "epoch: 8 \tTraining Loss: 0.263029\n",
      "epoch: 9 \tTraining Loss: 0.262292\n",
      "epoch: 10 \tTraining Loss: 0.261663\n",
      "epoch: 11 \tTraining Loss: 0.261129\n",
      "epoch: 12 \tTraining Loss: 0.260668\n",
      "epoch: 13 \tTraining Loss: 0.260258\n",
      "epoch: 14 \tTraining Loss: 0.259891\n",
      "epoch: 15 \tTraining Loss: 0.259546\n",
      "epoch: 16 \tTraining Loss: 0.259221\n",
      "epoch: 17 \tTraining Loss: 0.258912\n",
      "epoch: 18 \tTraining Loss: 0.258615\n",
      "epoch: 19 \tTraining Loss: 0.258334\n",
      "epoch: 20 \tTraining Loss: 0.258052\n",
      "epoch: 21 \tTraining Loss: 0.257768\n",
      "epoch: 22 \tTraining Loss: 0.257518\n",
      "epoch: 23 \tTraining Loss: 0.257284\n",
      "epoch: 24 \tTraining Loss: 0.257057\n",
      "epoch: 25 \tTraining Loss: 0.256831\n",
      "epoch: 26 \tTraining Loss: 0.256619\n",
      "epoch: 27 \tTraining Loss: 0.256414\n",
      "epoch: 28 \tTraining Loss: 0.256210\n",
      "epoch: 29 \tTraining Loss: 0.256005\n",
      "epoch: 30 \tTraining Loss: 0.255806\n"
     ]
    }
   ],
   "source": [
    "n_epochs = 30\n",
    "\n",
    "for epoch in range(1, n_epochs+1):\n",
    "    train_loss = 0.0\n",
    "    for data in train_loader:\n",
    "        images, _= data\n",
    "        images = images.view(images.size(0), -1)\n",
    "        optimizer.zero_grad()\n",
    "        outputs = model(images)\n",
    "        loss = criterion(outputs, images)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        train_loss += loss.item()*images.size(0)\n",
    "        \n",
    "    train_loss = train_loss/len(train_loader)\n",
    "    print('epoch: {} \\tTraining Loss: {:.6f}'.format(epoch, train_loss))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "19b7bff7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x288 with 20 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataiter = iter(test_loader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "images_flatten = images.view(images.size(0),-1)\n",
    "output = model(images_flatten)\n",
    "images = images.numpy()\n",
    "\n",
    "output = output.view(batch_size, 1, 28, 28)\n",
    "output = output.detach().numpy()\n",
    "\n",
    "fig, axes = plt.subplots(nrows=2, ncols=10, sharex=True, sharey=True, figsize=(25,4))\n",
    "\n",
    "for images, row in zip([images, output], axes):\n",
    "    for img,ax in zip(images, row):\n",
    "        ax.imshow(np.squeeze(img), cmap='gray')\n",
    "        ax.get_xaxis().set_visible(False)\n",
    "        ax.get_yaxis().set_visible(False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1a6394db",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}