intro-to-pytorch/Part 4 - Fashion-MNIST (Exercises).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classifying Fashion-MNIST\n",
    "\n",
    "Now it's your turn to build and train a neural network. You'll be using the [Fashion-MNIST dataset](https://github.com/zalandoresearch/fashion-mnist), a drop-in replacement for the MNIST dataset. MNIST is actually quite trivial with neural networks where you can easily achieve better than 97% accuracy. Fashion-MNIST is a set of 28x28 greyscale images of clothes. It's more complex than MNIST, so it's a better representation of the actual performance of your network, and a better representation of datasets you'll use in the real world.\n",
    "\n",
    "<img src='assets/fashion-mnist-sprite.png' width=500px>\n",
    "\n",
    "In this notebook, you'll build your own neural network. For the most part, you could just copy and paste the code from Part 3, but you wouldn't be learning. It's important for you to write the code yourself and get it to work. Feel free to consult the previous notebooks though as you work through this.\n",
    "\n",
    "First off, let's load the dataset through torchvision."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz to C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\train-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████▉| 26320896/26421880 [00:59<00:00, 1239651.04it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\train-images-idx3-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz to C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\train-labels-idx1-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "0it [00:00, ?it/s]\n",
      "  0%|                                                | 0/29515 [00:00<?, ?it/s]\n",
      " 56%|██████████████████▎              | 16384/29515 [00:00<00:00, 48212.42it/s]\n",
      "32768it [00:00, 40554.33it/s]                                                  "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\train-labels-idx1-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz to C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\t10k-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "0it [00:00, ?it/s]\n",
      "  0%|                                              | 0/4422102 [00:00<?, ?it/s]\n",
      "  0%|                               | 16384/4422102 [00:00<00:47, 92994.60it/s]\n",
      "  1%|▎                             | 49152/4422102 [00:00<00:40, 108573.94it/s]\n",
      "  2%|▋                            | 106496/4422102 [00:00<00:31, 135339.24it/s]\n",
      "  3%|▉                            | 139264/4422102 [00:01<00:33, 128030.54it/s]\n",
      "  6%|█▊                           | 270336/4422102 [00:01<00:24, 169820.01it/s]\n",
      "  9%|██▋                          | 401408/4422102 [00:01<00:18, 220129.49it/s]\n",
      " 17%|████▉                        | 745472/4422102 [00:01<00:12, 306077.12it/s]\n",
      " 20%|█████▊                       | 884736/4422102 [00:02<00:10, 330255.08it/s]\n",
      " 30%|████████▌                   | 1343488/4422102 [00:02<00:06, 444370.86it/s]\n",
      " 40%|███████████▎                | 1785856/4422102 [00:02<00:04, 607614.49it/s]\n",
      " 46%|████████████▊               | 2015232/4422102 [00:02<00:03, 754883.42it/s]\n",
      " 53%|██████████████▋             | 2326528/4422102 [00:02<00:02, 973105.10it/s]\n",
      " 58%|███████████████▋           | 2572288/4422102 [00:02<00:01, 1148418.20it/s]\n",
      " 63%|█████████████████          | 2801664/4422102 [00:02<00:01, 1053328.00it/s]\n",
      " 69%|██████████████████▌        | 3047424/4422102 [00:03<00:01, 1263376.22it/s]\n",
      " 74%|███████████████████▊       | 3252224/4422102 [00:03<00:00, 1287392.72it/s]\n",
      " 78%|█████████████████████▏     | 3465216/4422102 [00:03<00:00, 1448932.98it/s]\n",
      " 83%|██████████████████████▎    | 3653632/4422102 [00:03<00:00, 1199923.33it/s]\n",
      " 88%|███████████████████████▊   | 3907584/4422102 [00:03<00:00, 1425507.11it/s]\n",
      " 93%|█████████████████████████  | 4112384/4422102 [00:03<00:00, 1512327.37it/s]\n",
      " 97%|██████████████████████████▎| 4300800/4422102 [00:03<00:00, 1220585.02it/s]\n",
      "4423680it [00:04, 1075798.85it/s]                                              "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\t10k-images-idx3-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz to C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\t10k-labels-idx1-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "0it [00:00, ?it/s]\n",
      "  0%|                                                 | 0/5148 [00:00<?, ?it/s]\n",
      "8192it [00:00, 16650.63it/s]                                                   "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting C:\\Users\\User/.pytorch/F_MNIST_data/FashionMNIST\\raw\\t10k-labels-idx1-ubyte.gz\n",
      "Processing...\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "from torchvision import datasets, transforms\n",
    "import helper\n",
    "\n",
    "# Define a transform to normalize the data\n",
    "transform = transforms.Compose([transforms.ToTensor(),\n",
    "                                transforms.Normalize((0.5,), (0.5,))])\n",
    "# Download and load the training data\n",
    "trainset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=True, train=True, transform=transform)\n",
    "trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True)\n",
    "\n",
    "# Download and load the test data\n",
    "testset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=True, train=False, transform=transform)\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we can see one of the images."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "26427392it [01:10, 1239651.04it/s]                                             "
     ]
    }
   ],
   "source": [
    "image, label = next(iter(trainloader))\n",
    "helper.imshow(image[0,:]);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Building the network\n",
    "\n",
    "Here you should define your network. As with MNIST, each image is 28x28 which is a total of 784 pixels, and there are 10 classes. You should include at least one hidden layer. We suggest you use ReLU activations for the layers and to return the logits or log-softmax from the forward pass. It's up to you how many layers you add and the size of those layers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.nn import functional as F\n",
    "from torch import nn as nn\n",
    "from torch import optim as optim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Define your network architecture here\n",
    "class Network(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.fc1 = nn.Linear(784, 256)\n",
    "        self.fc2 = nn.Linear(256, 128)\n",
    "        self.fc3 = nn.Linear(128, 64)\n",
    "        self.fc4 = nn.Linear(64, 10)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        x = x.view(x.shape[0],-1)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = F.relu(self.fc3(x))\n",
    "        x = F.log_softmax(self.fc4(x), dim=1)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the network\n",
    "\n",
    "Now you should create your network and train it. First you'll want to define [the criterion](http://pytorch.org/docs/master/nn.html#loss-functions) ( something like `nn.CrossEntropyLoss`) and [the optimizer](http://pytorch.org/docs/master/optim.html) (typically `optim.SGD` or `optim.Adam`).\n",
    "\n",
    "Then write the training code. Remember the training pass is a fairly straightforward process:\n",
    "\n",
    "* Make a forward pass through the network to get the logits \n",
    "* Use the logits to calculate the loss\n",
    "* Perform a backward pass through the network with `loss.backward()` to calculate the gradients\n",
    "* Take a step with the optimizer to update the weights\n",
    "\n",
    "By adjusting the hyperparameters (hidden units, learning rate, etc), you should be able to get the training loss below 0.4."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Create the network, define the criterion and optimizer\n",
    "model = Network()\n",
    "criterion = nn.NLLLoss()\n",
    "optimizer = optim.Adam(model.parameters(),lr=0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training loss: 0.5218964978925454\n",
      "Training loss: 0.3760990379398057\n"
     ]
    }
   ],
   "source": [
    "# TODO: Train the network here\n",
    "epochs = 2\n",
    "for e in range(epochs):\n",
    "    running_loss = 0\n",
    "    for images, labels in trainloader:\n",
    "        log_ps = model(images)\n",
    "        loss = criterion(log_ps, labels)\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        running_loss += loss.item()\n",
    "    else:\n",
    "        print(f\"Training loss: {running_loss/len(trainloader)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x648 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 206,
       "width": 423
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "import helper\n",
    "\n",
    "# Test out your network!\n",
    "\n",
    "dataiter = iter(testloader)\n",
    "images, labels = dataiter.next()\n",
    "img = images[0]\n",
    "# Convert 2D image to 1D vector\n",
    "img = img.resize_(1, 784)\n",
    "\n",
    "# TODO: Calculate the class probabilities (softmax) for img\n",
    "ps = torch.exp(model(img))\n",
    "\n",
    "# Plot the image and probabilities\n",
    "helper.view_classify(img.resize_(1, 28, 28), ps, version='Fashion')"
   ]
  }
 ],
 "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}