ee046202_tutorial_07_deep_learn_pytorch_ae.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# <img src=\"https://img.icons8.com/dusk/64/000000/artificial-intelligence.png\" style=\"height:50px;display:inline\"> EE 046202 - Technion - Unsupervised Learning & Data Analysis\n",
    "---\n",
    "\n",
    "#### <a href=\"https://taldatech.github.io\">Tal Daniel</a>\n",
    "\n",
    "## Tutorial 07 - Deep Unsupervised Learning - PyTorch & Autoencoders\n",
    "---\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/bubbles/50/000000/checklist.png\" style=\"height:50px;display:inline\"> Agenda\n",
    "---\n",
    "* [Introduction & Motivation](#-Introduction-&-Motivation)\n",
    "* Up & Running with PyTorch\n",
    "    * [PyTorch Basics](#-PyTorch-Basics)\n",
    "    * [Datasets: MNIST, Fashion-MNIST](#-Datasets-in-PyTorch---MNIST-&-Fashion-MNIST)\n",
    "    * [Multi-Layer Perceptrons (MLPs)](#-Multi-Layer-Perceptron-(MLP)-in-PyTorch)\n",
    "    * [Extra: Convolutional Layers](#-Multi-Layer-Perceptron-(MLP)-in-PyTorch)\n",
    "* [Autoencoders](#-Autoencoders)\n",
    "    * [PyTorch Implementation]()\n",
    "* [Recommended Videos](#-Recommended-Videos)\n",
    "* [Credits](#-Credits)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "# imports for the tutorial\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import time\n",
    "\n",
    "# pytorch imports\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import DataLoader, Dataset, ConcatDataset\n",
    "import torchvision\n",
    "\n",
    "# scikit-learn imports\n",
    "from sklearn.manifold import LocallyLinearEmbedding, Isomap, TSNE\n",
    "from sklearn.decomposition import PCA, KernelPCA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## <img src=\"https://img.icons8.com/color/48/000000/fire-element.png\" style=\"height:50px;display:inline\" /> Introduction & Motivation\n",
    "---\n",
    "* Even though the idea of a neural network is a simple one, the implementation and calculations are not very sympathetic. \n",
    "* Neural networks are a strong tool that might be useful to many sorts of tasks and thus simple frameworks were established to help engineers focus on the task rather than the actual implemntation of the neural networks and making the backpropagation easy as one line of code.\n",
    "* There are several deep learning frameworks, where the two most popular ones are <a href=\"https://tensorflow.org\">Tensorflow</a> (maintained by Google) and <a href=\"https://pytorch.org\">PyTorch</a> (maintained by Facebook). \n",
    "* TF and PyTorch have different workflows, though in recent versions they share some of the workflows."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "<img src=\"./assets/tut_xiv_pytorch_logo.png\" style=\"height:100px\" />\n",
    "\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* PyTorch is \"an open source deep learning platform that provides a seamless path from research prototyping to production deployment\". \n",
    "* In simple words, PyTorch is meant to make deep learning programming more easier for Python programmers by adapting the Python workflow and \"easy-to-code\" language.  \n",
    "* Also, the PyTorch official website provides great tutorials to get you up and running in most deep learning fields (Computer Vision, Reinforcement Learning, Natural Language Processing)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "At its core, PyTorch provides two main features:\n",
    "\n",
    "* An n-dimensional Tensor, similar to numpy but can run on GPUs.\n",
    "* Automatic differentiation for building and training neural networks."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* The word **\"Tensor\"** might seem initmidating, but it is just an n-dimensional array or matrix. \n",
    "* A PyTorch Tensor is conceptually identical to a numpy array: a Tensor is an n-dimensional array, and PyTorch provides many functions for operating on these Tensors.\n",
    "* Behind the scenes, Tensors can **keep track of a computational graph and gradients**, but they’re also useful as a generic tool for scientific computing. \n",
    "* Moreover, Tensors can be placed on other devices than CPU, such as GPUs or TPUs, allowing to accelerate computations.\n",
    "* Most of the function and methods available in `numpy` are also available in PyTorch for tensors. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "#### Why not just use NumPy?\n",
    "\n",
    "Numpy is a great framework, but it cannot utilize GPUs to accelerate its numerical computations. For modern deep neural networks, GPUs often provide speedups of 50x or greater, so unfortunately numpy won’t be enough for modern deep learning."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/dusk/64/000000/diversity.png\" style=\"height:50px;display:inline\"> PyTorch Basics\n",
    "---\n",
    "We will now introduce some the basics. The main thing to keep in mind is that PyTorch was designed to have similar API as NumPy's. That means that almost every NumPy's function is in PyTorch, and most probably has the same name and same parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x: \n",
      " tensor([[-0.9316, -0.2761,  0.7423],\n",
      "        [-0.7188, -0.6464, -0.8161],\n",
      "        [ 0.2640,  0.6787, -0.3176],\n",
      "        [ 0.3718, -0.8876, -0.8888],\n",
      "        [-0.1430, -0.6374, -0.8229]])\n",
      "x size: torch.Size([5, 3])\n",
      "same as x shape: torch.Size([5, 3])\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# create a tensor initialized from a random uniform distribution\n",
    "x = torch.Tensor(5, 3).uniform_(-1, 1)\n",
    "print(\"x: \\n\", x)\n",
    "print(\"x size:\", x.size())\n",
    "print(\"same as x shape:\", x.shape)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "y: \n",
      " tensor([[0.5636, 0.0164, 0.2488],\n",
      "        [0.4739, 0.0954, 0.7210],\n",
      "        [0.1011, 0.9942, 0.8188],\n",
      "        [0.7817, 0.6225, 0.3460],\n",
      "        [0.9175, 0.1450, 0.6944]])\n",
      "x + y = \n",
      " tensor([[-0.3680, -0.2596,  0.9911],\n",
      "        [-0.2449, -0.5510, -0.0952],\n",
      "        [ 0.3651,  1.6729,  0.5012],\n",
      "        [ 1.1534, -0.2652, -0.5428],\n",
      "        [ 0.7745, -0.4924, -0.1285]])\n",
      "x - y = \n",
      " tensor([[-1.4953, -0.2925,  0.4935],\n",
      "        [-1.1927, -0.7417, -1.5371],\n",
      "        [ 0.1630, -0.3155, -1.1363],\n",
      "        [-0.4099, -1.5101, -1.2348],\n",
      "        [-1.0605, -0.7825, -1.5172]])\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# basic math oprations\n",
    "y = torch.rand(5, 3)\n",
    "print(\"y: \\n\", y)\n",
    "print(\"x + y = \\n\", x + y)\n",
    "print(\"x - y = \\n\", x - y)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x[3, :] =  tensor([ 0.3718, -0.8876, -0.8888])\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# indexing - same as NumPy\n",
    "print(\"x[3, :] = \", x[3, :])\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.ones([2, 2]): \n",
      " tensor([[1., 1.],\n",
      "        [1., 1.]])\n",
      "torch.zeros([1, 2]): \n",
      " tensor([[0., 0.]])\n",
      "y.reshape(-1, 1): \n",
      " tensor([[0.0746],\n",
      "        [0.9194],\n",
      "        [0.7642],\n",
      "        [0.7937],\n",
      "        [0.5881],\n",
      "        [0.2956],\n",
      "        [0.6239],\n",
      "        [0.1899],\n",
      "        [0.8838],\n",
      "        [0.4624],\n",
      "        [0.2099],\n",
      "        [0.3366],\n",
      "        [0.2956],\n",
      "        [0.7547],\n",
      "        [0.9789]])\n",
      "y.view(-1, 1): \n",
      " tensor([[0.0746],\n",
      "        [0.9194],\n",
      "        [0.7642],\n",
      "        [0.7937],\n",
      "        [0.5881],\n",
      "        [0.2956],\n",
      "        [0.6239],\n",
      "        [0.1899],\n",
      "        [0.8838],\n",
      "        [0.4624],\n",
      "        [0.2099],\n",
      "        [0.3366],\n",
      "        [0.2956],\n",
      "        [0.7547],\n",
      "        [0.9789]])\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# ones, zeros, reshaping - same as NumPy\n",
    "print(\"torch.ones([2, 2]): \\n\", torch.ones([2,2]))\n",
    "print(\"torch.zeros([1, 2]): \\n\", torch.zeros([1,2]))\n",
    "print(\"y.reshape(-1, 1): \\n\", y.reshape(-1, 1))\n",
    "# -1 just means 'don't care about the dimension this axis, make it happen'\n",
    "# more efficiently, use `view` as it doesn't copy the variable (saves memory)\n",
    "print(\"y.view(-1, 1): \\n\", y.view(-1, 1))\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "z: \n",
      " tensor([[1, 2],\n",
      "        [3, 4]], dtype=torch.int32)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# torch <-> numpy, preferred version\n",
    "z = torch.tensor(np.array([[1, 2], [3, 4]]))\n",
    "print(\"z: \\n\", z)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "z: \n",
      " tensor([[1, 2],\n",
      "        [3, 4]], dtype=torch.int32)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# torch <-> numpy, alternative verion\n",
    "a = np.array([[1, 2], [3, 4]])\n",
    "z = torch.from_numpy(a)\n",
    "print(\"z: \\n\", z)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true,
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "p = x.numpy(): \n",
      " [[-0.93162763 -0.27608442  0.7423003 ]\n",
      " [-0.71882963 -0.6463584  -0.8161427 ]\n",
      " [ 0.26402986  0.678723   -0.31756258]\n",
      " [ 0.3717742  -0.8876412  -0.88880277]\n",
      " [-0.14297056 -0.6374365  -0.8228916 ]]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "p = x.numpy()\n",
    "print(\"p = x.numpy(): \\n\", p)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "xy = x @ y.t(): \n",
      " tensor([[-0.3449,  0.0674,  0.2391, -0.6432, -0.3794],\n",
      "        [-0.6189, -0.9907, -1.3835, -1.2466, -1.3199],\n",
      "        [ 0.0810, -0.0391,  0.4415,  0.5190,  0.1202],\n",
      "        [-0.0262, -0.5493, -1.5727, -0.5695, -0.4048],\n",
      "        [-0.2958, -0.7218, -1.3220, -0.7933, -0.7950]])\n",
      "xy = torch.matmul(x, y.t()): \n",
      " tensor([[-0.3449,  0.0674,  0.2391, -0.6432, -0.3794],\n",
      "        [-0.6189, -0.9907, -1.3835, -1.2466, -1.3199],\n",
      "        [ 0.0810, -0.0391,  0.4415,  0.5190,  0.1202],\n",
      "        [-0.0262, -0.5493, -1.5727, -0.5695, -0.4048],\n",
      "        [-0.2958, -0.7218, -1.3220, -0.7933, -0.7950]])\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# tensor multiplication\n",
    "xy = x @ y.t()\n",
    "print(\"xy = x @ y.t(): \\n\", xy)\n",
    "# or\n",
    "xy = torch.matmul(x, y.t())\n",
    "print(\"xy = torch.matmul(x, y.t()): \\n\", xy)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false,
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "xy = x * y: \n",
      " tensor([[-0.5251, -0.0045,  0.1847],\n",
      "        [-0.3406, -0.0616, -0.5884],\n",
      "        [ 0.0267,  0.6748, -0.2600],\n",
      "        [ 0.2906, -0.5525, -0.3075],\n",
      "        [-0.1312, -0.0924, -0.5714]])\n",
      "xy = torch.mul(x, y): \n",
      " tensor([[-0.5251, -0.0045,  0.1847],\n",
      "        [-0.3406, -0.0616, -0.5884],\n",
      "        [ 0.0267,  0.6748, -0.2600],\n",
      "        [ 0.2906, -0.5525, -0.3075],\n",
      "        [-0.1312, -0.0924, -0.5714]])\n"
     ]
    }
   ],
   "source": [
    "# element-wise\n",
    "xy = x * y\n",
    "print(\"xy = x * y: \\n\", xy)\n",
    "# or\n",
    "xy = torch.mul(x, y)\n",
    "print(\"xy = torch.mul(x, y): \\n\", xy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/dusk/64/000000/processor.png\" style=\"height:50px;display:inline\"> Device - CPU or GPU?\n",
    "---\n",
    "* The greatest advantage of using the deep learning frameworks in the ability to utilize the GPUs. In PyTorch, we can code it to automatically choose which device to use.\n",
    "* A good practice is to always define the device at the beginning of your code, and then just send the models and tensors to that device using `.to(device)`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True\n",
      "current device:  0\n",
      "device:  cuda:0\n",
      "tensor([[-1.2017, -0.6636],\n",
      "        [ 1.2575, -0.0579]], device='cuda:0')\n"
     ]
    }
   ],
   "source": [
    "# check if there is a GPU available\n",
    "print(torch.cuda.is_available())\n",
    "# check what is the current available device\n",
    "if torch.cuda.is_available():\n",
    "    print(\"current device: \", torch.cuda.current_device())\n",
    "# automatically choose device\n",
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")  # use gpu 0 if it is available, o.w. use the cpu\n",
    "print(\"device: \", device)\n",
    "\n",
    "# create a random tensor and send it to the device\n",
    "a = torch.randn([2, 2]).to(device)\n",
    "print(a)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/cute-clipart/64/000000/documents-folder.png\" style=\"height:50px;display:inline\"> Datasets in PyTorch - MNIST & Fashion-MNIST\n",
    "---\n",
    "* The first thing we need to do before we build our networks is load the data, separate it to train and test sets (and sometimes also validation) and create batches of it to train using one of the Gradient Descent optimizers.\n",
    "* Fortunately, PyTorch provide a simple data structure to load the data (usually from files) called `Dataset` and another data structure, called `DataLoader` that creates batches out of the Dataset (and it even takes care of shuffling it if you wish).\n",
    "* We now introduce two very popular datasets:\n",
    "\n",
    "1. **MNIST**- The MNIST database (1998) of handwritten digits (0-9, a total of 10 digits) has a training set of 60,000 examples, and a test set of 10,000 examples. Images are of size 28 x 28 pixels with one color channel. It is a subset of a larger set available from NIST. The digits have been size-normalized and centered in a fixed-size image. It is a good database for people who want to try learning techniques and pattern recognition methods on real-world data while spending minimal efforts on preprocessing and formatting. MNIST is broadly used as a baseline to many ML papers, even today.\n",
    "2. **Fashion-MNIST** - Fashion-MNIST is a dataset of Zalando's article images—consisting of a training set of 60,000 examples and a test set of 10,000 examples. Each example is a 28x28 grayscale image, associated with a label from 10 classes. Zalando intends Fashion-MNIST to serve as a direct drop-in replacement for the original MNIST dataset for benchmarking machine learning algorithms. It shares the same image size and structure of training and testing splits.\n",
    "\n",
    "Let's load the dataset (PyTorch provides an auto-download feature)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true,
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape: \n",
      " torch.Size([128, 1, 28, 28])\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "batch_size = 128\n",
    "# MNIST\n",
    "mnist_train_dataset = torchvision.datasets.MNIST(root='./datasets/',\n",
    "                                           train=True, \n",
    "                                           transform=torchvision.transforms.ToTensor(),\n",
    "                                           download=True)\n",
    "\n",
    "mnist_test_dataset = torchvision.datasets.MNIST(root='./datasets',\n",
    "                                          train=False, \n",
    "                                          transform=torchvision.transforms.ToTensor())\n",
    "\n",
    "# Data loader\n",
    "mnist_train_loader = torch.utils.data.DataLoader(dataset=mnist_train_dataset,\n",
    "                                           batch_size=batch_size, \n",
    "                                           shuffle=True, drop_last=True)\n",
    "\n",
    "# We use drop_last=True to avoid the case where the data / batch_size != int\n",
    "\n",
    "mnist_test_loader = torch.utils.data.DataLoader(dataset=mnist_test_dataset,\n",
    "                                          batch_size=batch_size, \n",
    "                                          shuffle=False)\n",
    "\n",
    "# let's plot some of the samples from the test set\n",
    "examples = enumerate(mnist_test_loader)\n",
    "batch_idx, (example_data, example_targets) = next(examples)\n",
    "print(\"shape: \\n\", example_data.shape)\n",
    "fig = plt.figure()\n",
    "for i in range(6):\n",
    "    ax = fig.add_subplot(2,3,i+1)\n",
    "    ax.imshow(example_data[i][0], cmap='gray', interpolation='none')\n",
    "    ax.set_title(\"Ground Truth: {}\".format(example_targets[i]))\n",
    "    ax.set_axis_off()\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape: \n",
      " torch.Size([128, 1, 28, 28])\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Fashion-MNIST\n",
    "fmnist_train_dataset = torchvision.datasets.FashionMNIST(root='./datasets/',\n",
    "                                           train=True, \n",
    "                                           transform=torchvision.transforms.ToTensor(),\n",
    "                                           download=True)\n",
    "\n",
    "fmnist_test_dataset = torchvision.datasets.FashionMNIST(root='./datasets',\n",
    "                                          train=False, \n",
    "                                          transform=torchvision.transforms.ToTensor())\n",
    "\n",
    "# Data loader\n",
    "fmnist_train_loader = torch.utils.data.DataLoader(dataset=fmnist_train_dataset,\n",
    "                                           batch_size=batch_size, \n",
    "                                           shuffle=True, drop_last=True)\n",
    "\n",
    "fmnist_test_loader = torch.utils.data.DataLoader(dataset=fmnist_test_dataset,\n",
    "                                          batch_size=batch_size, \n",
    "                                          shuffle=False)\n",
    "\n",
    "# let's plot some of the samples from the test set\n",
    "examples = enumerate(fmnist_test_loader)\n",
    "batch_idx, (example_data, example_targets) = next(examples)\n",
    "print(\"shape: \\n\", example_data.shape)\n",
    "fig = plt.figure()\n",
    "for i in range(6):\n",
    "    ax = fig.add_subplot(2,3,i+1)\n",
    "    ax.imshow(example_data[i][0], cmap='gray', interpolation='none')\n",
    "    ax.set_title(\"Ground Truth: {}\".format(example_targets[i]))\n",
    "    ax.set_axis_off()\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/color/96/000000/broadcasting.png\" style=\"height:50px;display:inline\"> Multi-Layer Perceptron (MLP) in PyTorch\n",
    "---\n",
    "* An MLP (other names: Fully-Connected (FC) Network, Dense Network) is composed of one input layer, one or more hidden layers and a final output layer. \n",
    "* Every layer, except the output layer includes a bias neuron which is fully connected to the next layer.\n",
    "* When the number of hidden layers is larger than 2, the network is usually called a deep neural network (DNN).\n",
    "\n",
    "* The algorithm is composed of two main parts: **forward pass** and **backward pass**. \n",
    "* In the *forward pass*, for each training instance, the algorithm feeds it to the network and computes the output of every neuron in each consecutive layer (using the network for prediction is just doing a forward pass). Then, the output error (the difference between the desired output and the actual output from the network) is computed.\n",
    "\n",
    "* After the output error calculation, the network calculates how much each neuron in the last hidden layer contributed to the output error (using the chain rule). \n",
    "* It then proceeds to measure how much of these error contributions came from each neuron in the previous layers until reaching the input layer. \n",
    "* This is the *backward pass*: measuring the error gradient across all the connection weights in the network by propagating the error gradient backward in the network (this is the backpropagation process).\n",
    "\n",
    "\n",
    "* **In short**: for each training instance the backpropagation algorithm first makes a prediction (forward pass), measures the error, then goes in reverse to measure the error contribution from each connection (backward pass) and finally, using Gradient Descent, updates the weights in the direction that reduces the error."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "<img src=\"./assets/tut_xiv_mlp.jpg\" style=\"height:200px\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#### Implementing an MLP\n",
    "---\n",
    "We will now see how easy it is to implement a neural network using PyTorch. The first thing to know is that every neural network (NN) inherits from parent class `torch.nn.Module`, which establishes a shared workflow, that is, every (!) NN in PyTorch has the same properties in terms of functunality. This global design helps with building a common language. We will now meet PyTorch's basic blocks and see why it is so popular. \n",
    "\n",
    "We will implement the following network:\n",
    "* Input dimension: 28 * 28\n",
    "* Output dimension: 10 (as the number of classes)\n",
    "* Batch size: 128\n",
    "* Hidden Layers: 1\n",
    "* Hidden Units: 256\n",
    "* Optimizer: Adam (Learning Rate: 0.001)\n",
    "* Activation: ReLU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# define a two-layer MLP\n",
    "# method 1\n",
    "class TwoLayerNet(torch.nn.Module):\n",
    "    def __init__(self, D_in, H, D_out):\n",
    "        \"\"\"\n",
    "        In the constructor we instantiate two nn.Linear modules and assign them as\n",
    "        member variables.\n",
    "        Parameters:\n",
    "            D_in - dimensions of inputs\n",
    "            H - number of hidden units per layer\n",
    "            D_out - dimensions of outputs\n",
    "        \"\"\"\n",
    "        # initialzing the parent object (important!)\n",
    "        super(TwoLayerNet, self).__init__()\n",
    "        # define the first layer (hidden)\n",
    "        self.linear1 = torch.nn.Linear(D_in, H)\n",
    "        # define the second layer (output)\n",
    "        self.linear2 = torch.nn.Linear(H, D_out)\n",
    "        # define the activation function\n",
    "        self.relu = torch.nn.ReLU()\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        In the forward function we accept a Tensor of input data and we must return\n",
    "        a Tensor of output data. We can use Modules defined in the constructor as\n",
    "        well as arbitrary operators on Tensors.\n",
    "        Parameters:\n",
    "            x - tensor of inputs (shape: [BATCH_SIZE, D_in])\n",
    "        \"\"\"\n",
    "        h_relu = self.relu(self.linear1(x))\n",
    "        y_pred = self.linear2(h_relu)\n",
    "        return y_pred\n",
    "    \n",
    "# method 2\n",
    "class TwoLayerNetPiped(torch.nn.Module):\n",
    "    def __init__(self, D_in, H, D_out):\n",
    "        \"\"\"\n",
    "        In the constructor we instantiate two nn.Linear modules and assign them as\n",
    "        member variables.\n",
    "        Parameters:\n",
    "            D_in - dimensions of inputs\n",
    "            H - number of hidden units per layer\n",
    "            D_out - dimensions of outputs\n",
    "        \"\"\"\n",
    "        # initialzing the parent object (important!)\n",
    "        super(TwoLayerNetPiped, self).__init__()\n",
    "        # Create a pipeline - a sequence of layers\n",
    "        self.pipe = torch.nn.Sequential(\n",
    "            torch.nn.Linear(D_in, H), \n",
    "            torch.nn.ReLU(),\n",
    "            torch.nn.Linear(H, D_out))\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        In the forward function we accept a Tensor of input data and we must return\n",
    "        a Tensor of output data. We can use Modules defined in the constructor as\n",
    "        well as arbitrary operators on Tensors.\n",
    "        Parameters:\n",
    "            x - tensor of inputs (shape: [BATCH_SIZE, D_in])\n",
    "        \"\"\"\n",
    "        return self.pipe(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### When to use method 1 and when to use method 2?\n",
    "---\n",
    "If you want easy access to the weights of the layers (to do some manipulating on them or analyze them) then it is better to use method 1, otherwise, use method 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# hyper-parameters:\n",
    "num_epochs = 10\n",
    "learning_rate = 0.001\n",
    "\n",
    "# Device configuration, as before\n",
    "device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "# create model, send it to device\n",
    "model = TwoLayerNetPiped(D_in=28*28, H=256, D_out=10).to(device)\n",
    "\n",
    "# Loss and optimizer\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/10], Step [100/468], Loss: 0.4783, Time: 1.8672 secs\n",
      "Epoch [1/10], Step [200/468], Loss: 0.1721, Time: 3.2621 secs\n",
      "Epoch [1/10], Step [300/468], Loss: 0.1589, Time: 4.5964 secs\n",
      "Epoch [1/10], Step [400/468], Loss: 0.1732, Time: 5.9777 secs\n",
      "Epoch [2/10], Step [100/468], Loss: 0.2179, Time: 8.2756 secs\n",
      "Epoch [2/10], Step [200/468], Loss: 0.2117, Time: 9.6090 secs\n",
      "Epoch [2/10], Step [300/468], Loss: 0.1143, Time: 10.9415 secs\n",
      "Epoch [2/10], Step [400/468], Loss: 0.1430, Time: 12.2827 secs\n",
      "Epoch [3/10], Step [100/468], Loss: 0.1151, Time: 14.6643 secs\n",
      "Epoch [3/10], Step [200/468], Loss: 0.0915, Time: 16.0267 secs\n",
      "Epoch [3/10], Step [300/468], Loss: 0.2148, Time: 17.3831 secs\n",
      "Epoch [3/10], Step [400/468], Loss: 0.1106, Time: 18.6888 secs\n",
      "Epoch [4/10], Step [100/468], Loss: 0.0481, Time: 20.8634 secs\n",
      "Epoch [4/10], Step [200/468], Loss: 0.0289, Time: 22.1266 secs\n",
      "Epoch [4/10], Step [300/468], Loss: 0.0971, Time: 23.4104 secs\n",
      "Epoch [4/10], Step [400/468], Loss: 0.0954, Time: 24.7019 secs\n",
      "Epoch [5/10], Step [100/468], Loss: 0.0921, Time: 26.8355 secs\n",
      "Epoch [5/10], Step [200/468], Loss: 0.0399, Time: 28.0898 secs\n",
      "Epoch [5/10], Step [300/468], Loss: 0.0403, Time: 29.3446 secs\n",
      "Epoch [5/10], Step [400/468], Loss: 0.0204, Time: 30.6226 secs\n",
      "Epoch [6/10], Step [100/468], Loss: 0.0443, Time: 32.8190 secs\n",
      "Epoch [6/10], Step [200/468], Loss: 0.0593, Time: 34.2322 secs\n",
      "Epoch [6/10], Step [300/468], Loss: 0.1105, Time: 35.5956 secs\n",
      "Epoch [6/10], Step [400/468], Loss: 0.0168, Time: 36.9479 secs\n",
      "Epoch [7/10], Step [100/468], Loss: 0.0191, Time: 39.2662 secs\n",
      "Epoch [7/10], Step [200/468], Loss: 0.0214, Time: 40.6056 secs\n",
      "Epoch [7/10], Step [300/468], Loss: 0.0379, Time: 41.9390 secs\n",
      "Epoch [7/10], Step [400/468], Loss: 0.0780, Time: 43.2827 secs\n",
      "Epoch [8/10], Step [100/468], Loss: 0.0424, Time: 45.5493 secs\n",
      "Epoch [8/10], Step [200/468], Loss: 0.0149, Time: 46.8704 secs\n",
      "Epoch [8/10], Step [300/468], Loss: 0.0186, Time: 48.2039 secs\n",
      "Epoch [8/10], Step [400/468], Loss: 0.0212, Time: 49.5511 secs\n",
      "Epoch [9/10], Step [100/468], Loss: 0.0161, Time: 51.8537 secs\n",
      "Epoch [9/10], Step [200/468], Loss: 0.0053, Time: 53.0754 secs\n",
      "Epoch [9/10], Step [300/468], Loss: 0.0075, Time: 54.2974 secs\n",
      "Epoch [9/10], Step [400/468], Loss: 0.0357, Time: 55.5677 secs\n",
      "Epoch [10/10], Step [100/468], Loss: 0.0183, Time: 57.7369 secs\n",
      "Epoch [10/10], Step [200/468], Loss: 0.0526, Time: 58.9671 secs\n",
      "Epoch [10/10], Step [300/468], Loss: 0.0375, Time: 60.2029 secs\n",
      "Epoch [10/10], Step [400/468], Loss: 0.0186, Time: 61.4466 secs\n"
     ]
    }
   ],
   "source": [
    "# Train the model\n",
    "model.train()  # training mode\n",
    "total_step = len(mnist_train_loader)\n",
    "start_time = time.time()\n",
    "for epoch in range(num_epochs):\n",
    "    for i, (images, labels) in enumerate(mnist_train_loader):\n",
    "        # each i is a batch of 128 samples\n",
    "        images = images.to(device).view(batch_size, -1)  # represent images as column vectors\n",
    "        labels = labels.to(device)\n",
    "        \n",
    "        # Forward pass\n",
    "        outputs = model(images)\n",
    "        loss = criterion(outputs, labels)\n",
    "        \n",
    "        # Backward and optimize - ALWAYS IN THIS ORDER!\n",
    "        optimizer.zero_grad()  # zero (flush) the gradients from the previous iteration\n",
    "        loss.backward()  # calculate the gradients w.r.t the loss function (saved in tensor.grad field)\n",
    "        optimizer.step()  # apply the gradients\n",
    "        \n",
    "        if (i + 1) % 100 == 0:\n",
    "            print ('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}, Time: {:.4f} secs' \n",
    "                   .format(epoch + 1, num_epochs, i + 1, total_step, loss.item(), time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy of the model on the 10000 test images: 97.78 %\n"
     ]
    }
   ],
   "source": [
    "# Test the model\n",
    "model.eval()  # eval mode (batchnorm uses moving mean/variance instead of mini-batch mean/variance)\n",
    "with torch.no_grad(): # \"don't keep track of the gradients\" -> faster, can also use .detach()\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    for images, labels in mnist_test_loader:\n",
    "        images = images.to(device).view(images.size(0), -1)\n",
    "        labels = labels.to(device)\n",
    "        outputs = model(images)\n",
    "        _, predicted = torch.max(outputs.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "    print('Test Accuracy of the model on the 10000 test images: {} %'.format(100 * correct / total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/dusk/50/000000/xlarge-icons.png\" style=\"height:50px;display:inline\"> Extra - Convolutional Neural Networks\n",
    "---\n",
    "Let's first recap some of the basics:\n",
    "### 2D Convolution\n",
    "---\n",
    "Convolution is moving a window or filter across the image being studied. This moving window applies to a certain neighborhood of nodes as shown below – here, the filter applied is (0.5 × the node value):\n",
    "<img src=\"./assets/tut_xiv_filter.jpg\" style=\"height:200px\" />\n",
    "In the convolutional part of the CNN, we can imagine this 2 x 2 moving filter sliding across all the available nodes / pixels in the input image. This operation can also be illustrated using standard neural network node diagrams:\n",
    "<img src=\"./assets/tut_xiv_layer.jpg\" style=\"height:200px\" />\n",
    "The first position of the moving filter connections is illustrated by the blue connections, and the second is shown with the green lines. The weights of each of these connections, as stated previously, is 0.5.\n",
    "\n",
    "Calculating a convolutional layer output size:\n",
    "We define the following:\n",
    "* $W_{in}$ - the width of the input\n",
    "* $F$ - filter size\n",
    "* $P$ - padding\n",
    "* $S$ -stride\n",
    "The output width:\n",
    "$$W_{out} = \\frac{W_{in} - F + 2P}{S} + 1$$\n",
    "In our example, the output of the convolutional layers, just before the FC:\n",
    "$$W_{1, out} = \\frac{28 - 5 + 2*2}{1} + 1 = 28 \\rightarrow MaxPooling(2x2) \\rightarrow 28 / 2 = 14$$\n",
    "$$W_{2, out} = \\frac{14 - 5 + 2*2}{1} + 1 = 14 \\rightarrow MaxPooling(2x2) \\rightarrow 14 / 2 = 7$$\n",
    "So the input to the FC layer is $7x7=49$\n",
    "\n",
    "### Feature mapping and multiple channels\n",
    "---\n",
    "Since the weights of individual filters are held constant as they are applied over the input nodes, they can be trained to select certain features from the input data. In the case of images, it may learn to recognize common geometrical objects such as lines, edges and other shapes which make up objects. This is where the name feature mapping comes from. Because of this, any convolution layer needs multiple filters which are trained to detect different features.\n",
    "\n",
    "### Pooling\n",
    "---\n",
    "It is a sliding window type technique, but instead of applying weights, which can be trained, it applies a statistical function of some type over the contents of its window. The most common type of pooling is called max pooling, and it applies the max() function over the contents of the window.\n",
    "There are two main benefits to pooling in CNN's:\n",
    "1. It reduces the number of parameters in your model by a process called down-sampling\n",
    "2. It makes feature detection more robust to object orientation and scale changes\n",
    "\n",
    "Pooling generalizes over lower level, more complex information. Let’s imagine the case where we have convolutional filters that, during training, learn to detect the digit “9” in various orientations within the input images. In order for the Convolutional Neural Network to learn to classify the appearance of “9” in the image correctly, it needs to in some way “activate” whenever a “9” is found anywhere in the image, no matter what the size or orientation the digit is (except for when it looks like “6”, that is). Pooling can assist with this higher level, generalized feature selection. An example can be seen <a href=\"https://adventuresinmachinelearning.com/convolutional-neural-networks-tutorial-in-pytorch/\">here</a>.\n",
    "\n",
    "### Strides and down-sampling\n",
    "---\n",
    "In the pooling diagram below, you will notice that the pooling window shifts to the right each time by 2 places. This is called a stride of 2, which should be considered both in the x and y direction. In other words, the stride is actually specified as [2, 2]. One important thing to notice is that, if during pooling the stride is greater than 1, then the output size will be reduced. As can be observed below, the 5 x 5 input is reduced to a 3 x 3 output. This is a good thing – it is called down-sampling, and it reduces the number of trainable parameters in the model.\n",
    "\n",
    "### Padding\n",
    "---\n",
    "In the pooling diagram below is that there is an extra column and row added to the 5 x 5 input – this makes the effective size of the pooling space equal to 6 x 6. This is to ensure that the 2 x 2 pooling window can operate correctly with a stride of [2, 2] and is called padding. These nodes are basically dummy nodes – because the values of these dummy nodes is 0, they are basically invisible to the max pooling operation. Padding will need to be considered when constructing our Convolutional Neural Network in PyTorch.\n",
    "<img src=\"./assets/tut_xiv_max_pool.jpg\" style=\"height:200px\" />\n",
    "* <a href=\"adventuresinmachinelearning.com\">Image Source</a>\n",
    "\n",
    "### The FC Layer\n",
    "---\n",
    "The fully connected layer can be thought of as attaching a standard classifier onto the information-rich output of the network, to “interpret” the results and finally produce a classification result. In order to attach this fully connected layer to the network, the dimensions of the output of the Convolutional Neural Network need to be flattened.\n",
    "\n",
    "\n",
    "### CNN Vs. Fully Connected\n",
    "---\n",
    "Fully connected networks with a few layers can only do so much – to get close to state-of-the-art results in image classification it is necessary to go deeper. In other words, lots more layers are required in the network. However, by adding a lot of additional layers, we come across some problems. First, we can run into the vanishing gradient problem. However, this can be solved to an extent by using sensible activation functions, such as the ReLU family of activations. Another issue for deep fully connected networks is that the number of trainable parameters in the model (i.e. the weights) can grow rapidly. This means that the training slows down or becomes practically impossible, and also exposes the model to overfitting. CNNs try to solve this second problem by exploiting correlations between adjacent inputs in images. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# let's build the ConvNet\n",
    "# TRICK: if you don't want to manually calcualte the conv layers output size, you can use a dummy input\n",
    "# to test the output dimension.\n",
    "class ConvNet(nn.Module):\n",
    "    def __init__(self, input_shape, num_classes=10):\n",
    "        super(ConvNet, self).__init__()\n",
    "        self.layer1 = nn.Sequential(\n",
    "            nn.Conv2d(input_shape[0], 16, kernel_size=5, stride=1, padding=2),\n",
    "            nn.BatchNorm2d(16),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
    "        self.layer2 = nn.Sequential(\n",
    "            nn.Conv2d(16, 32, kernel_size=5, stride=1, padding=2),\n",
    "            nn.BatchNorm2d(32),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
    "#         self.fc = nn.Linear(7*7*32, num_classes)\n",
    "        # or, use our TRICK:\n",
    "        self.fc = nn.Linear(self._get_conv_out(input_shape), num_classes)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        out = self.layer1(x)\n",
    "        out = self.layer2(out)\n",
    "        out = out.view(out.size(0), -1)\n",
    "        out = self.fc(out)\n",
    "        return out\n",
    "    \n",
    "    def _get_conv_out(self, shape):\n",
    "        \"\"\"\n",
    "        Helper function to automatically calculate the conv layers output.\n",
    "        \"\"\"\n",
    "        o = self.layer2(self.layer1((torch.zeros(1, *shape))))\n",
    "        return int(np.prod(o.size()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# hyper-parameters:\n",
    "num_epochs = 5\n",
    "learning_rate = 0.001\n",
    "input_shape = [1, 28, 28] # num_channels, height, width\n",
    "\n",
    "# Device configuration, as before\n",
    "device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "# create model\n",
    "model = ConvNet(input_shape=input_shape, num_classes=10).to(device)\n",
    "\n",
    "# Loss and optimizer\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/5], Step [100/468], Loss: 0.1966, Time: 2.9221 secs\n",
      "Epoch [1/5], Step [200/468], Loss: 0.0972, Time: 4.5916 secs\n",
      "Epoch [1/5], Step [300/468], Loss: 0.0552, Time: 6.3140 secs\n",
      "Epoch [1/5], Step [400/468], Loss: 0.0602, Time: 8.0174 secs\n",
      "Epoch [2/5], Step [100/468], Loss: 0.0347, Time: 10.9237 secs\n",
      "Epoch [2/5], Step [200/468], Loss: 0.0090, Time: 12.6321 secs\n",
      "Epoch [2/5], Step [300/468], Loss: 0.0777, Time: 14.3456 secs\n",
      "Epoch [2/5], Step [400/468], Loss: 0.0507, Time: 15.9792 secs\n",
      "Epoch [3/5], Step [100/468], Loss: 0.0621, Time: 18.7488 secs\n",
      "Epoch [3/5], Step [200/468], Loss: 0.0225, Time: 20.4073 secs\n",
      "Epoch [3/5], Step [300/468], Loss: 0.0218, Time: 22.0461 secs\n",
      "Epoch [3/5], Step [400/468], Loss: 0.0728, Time: 23.6847 secs\n",
      "Epoch [4/5], Step [100/468], Loss: 0.0216, Time: 26.4405 secs\n",
      "Epoch [4/5], Step [200/468], Loss: 0.0211, Time: 28.0821 secs\n",
      "Epoch [4/5], Step [300/468], Loss: 0.0085, Time: 29.7826 secs\n",
      "Epoch [4/5], Step [400/468], Loss: 0.0094, Time: 31.5079 secs\n",
      "Epoch [5/5], Step [100/468], Loss: 0.0076, Time: 34.4580 secs\n",
      "Epoch [5/5], Step [200/468], Loss: 0.0206, Time: 36.2064 secs\n",
      "Epoch [5/5], Step [300/468], Loss: 0.0159, Time: 37.9407 secs\n",
      "Epoch [5/5], Step [400/468], Loss: 0.0098, Time: 39.6731 secs\n"
     ]
    }
   ],
   "source": [
    "# Train the model\n",
    "total_step = len(mnist_train_loader)\n",
    "start_time = time.time()\n",
    "for epoch in range(num_epochs):\n",
    "    for i, (images, labels) in enumerate(mnist_train_loader):\n",
    "        # each i is a batch of 128 samples\n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "        \n",
    "        # Forward pass\n",
    "        outputs = model(images)\n",
    "        loss = criterion(outputs, labels)\n",
    "        \n",
    "        # Backward and optimize\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        if (i + 1) % 100 == 0:\n",
    "            print ('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}, Time: {:.4f} secs' \n",
    "                   .format(epoch + 1, num_epochs, i + 1, total_step, loss.item(), time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy of the model on the 10000 test images: 98.83 %\n"
     ]
    }
   ],
   "source": [
    "# Test the model\n",
    "model.eval()  # eval mode (batchnorm uses moving mean/variance instead of mini-batch mean/variance)\n",
    "with torch.no_grad(): # \"don't keep track of the gradients\" ,can also use .detach()\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    for images, labels in mnist_test_loader:\n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "        outputs = model(images)\n",
    "        _, predicted = torch.max(outputs.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "    print('Test Accuracy of the model on the 10000 test images: {} %'.format(100 * correct / total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## <img src=\"https://img.icons8.com/color/96/000000/code.png\" style=\"height:50px;display:inline\"> Autoencoders\n",
    "---\n",
    "* Most of the natural data is high-dimensional, such as images. Consider the MNIST (hand-written digits) dataset, where each image has $28x28=784$ pixels, which means it can be represented by a vector of length 784. \n",
    "    * But do we really need 784 values to represent a digit? The answer is no. We believe that the data lies on a low-dimensional space which is enough to describe the observasions. In case of MNIST, we can choose to represent digits as one-hot vectors, which means we only need 10 dimensions. So we can **encode** high-dimensional observations in a low-dimensional space.\n",
    "    * But how can we learn meaningful low-dimensional representations? The general idea is to reconstruct or, **decode** the low-dimensional representation to the high-dimensional reperesentation, and use the reconstruction error to find the best representations (using the gradients of the error) . This is the core idea behind **autoencoders**.\n",
    "    * **Autoencoders** - models which take data as input and discover some latent state representation of that data.The input data is converted into an encoding vector where each dimension represents some learned attribute about the data. The most important detail to grasp here is that our encoder network is outputting a single value for each encoding dimension. The decoder network then subsequently takes these values and attempts to recreate the original input. \n",
    "        * Autoencoders have **three parts**: an encoder, a decoder, and a 'loss' function that maps one to the other. For the simplest autoencoders - the sort that compress and then reconstruct the original inputs from the compressed representation - we can think of the 'loss' as describing the amount of information lost in the process of reconstruction."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* Illustration: <img src=\"./assets/tut_xv_autoencoder.png\" style=\"height:250px\">\n",
    "        \n",
    "* The basic architecture of an autoencoder:\n",
    "<img src=\"./assets/tut_xiv_autoencoder.png\" style=\"height:250px\">\n",
    "* <a href=\"https://towardsdatascience.com/applied-deep-learning-part-3-autoencoders-1c083af4d798\">Image Source</a>\n",
    "    \n",
    "Let's implement it in PyTorch using what we have learnt so far!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "class AutoEncoder(nn.Module):\n",
    "    \n",
    "    def __init__(self, input_dim=28*28, hidden_dim=256, latent_dim=10):\n",
    "        super(AutoEncoder, self).__init__()\n",
    "        \n",
    "        self.input_dim = input_dim\n",
    "        self.hidden_dim = hidden_dim\n",
    "        self.latent_dim = latent_dim\n",
    "        \n",
    "        # define the encoder\n",
    "        self.encoder = nn.Sequential(nn.Linear(self.input_dim, self.hidden_dim),\n",
    "                                     nn.ReLU(), nn.Linear(self.hidden_dim, self.hidden_dim),\n",
    "                                     nn.ReLU(),\n",
    "                                     nn.Linear(self.hidden_dim, self.latent_dim)\n",
    "                                    )\n",
    "        \n",
    "        # define decoder\n",
    "        self.decoder = nn.Sequential(nn.Linear(self.latent_dim, self.hidden_dim),\n",
    "                                     nn.ReLU(),\n",
    "                                     nn.Linear(self.hidden_dim, self.hidden_dim),\n",
    "                                     nn.ReLU(),\n",
    "                                     nn.Linear(self.hidden_dim, self.input_dim),\n",
    "                                     nn.Sigmoid())\n",
    "        \n",
    "    def forward(self,x):\n",
    "        x = self.encoder(x)\n",
    "        x = self.decoder(x)\n",
    "        return x\n",
    "    \n",
    "    def get_latent_rep(self, x):\n",
    "        return self.encoder(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# hyper-parameters:\n",
    "num_epochs = 5\n",
    "learning_rate = 0.001\n",
    "\n",
    "# Device configuration, as before\n",
    "device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "# create model, send it to device\n",
    "model = AutoEncoder(input_dim=28 * 28, hidden_dim=128, latent_dim=10).to(device)\n",
    "\n",
    "# Loss and optimizer\n",
    "criterion = nn.BCELoss()  # binary cross entropy, as pixels are in [0,1]\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/5], Step [100/468], Loss: 0.3697, Time: 1.8620 secs\n",
      "Epoch [1/5], Step [200/468], Loss: 0.3328, Time: 3.6951 secs\n",
      "Epoch [1/5], Step [300/468], Loss: 0.3261, Time: 5.5362 secs\n",
      "Epoch [1/5], Step [400/468], Loss: 0.3315, Time: 7.3733 secs\n",
      "Epoch [2/5], Step [100/468], Loss: 0.3160, Time: 10.4680 secs\n",
      "Epoch [2/5], Step [200/468], Loss: 0.2970, Time: 12.3111 secs\n",
      "Epoch [2/5], Step [300/468], Loss: 0.3030, Time: 14.1701 secs\n",
      "Epoch [2/5], Step [400/468], Loss: 0.2982, Time: 16.0132 secs\n",
      "Epoch [3/5], Step [100/468], Loss: 0.3164, Time: 19.1069 secs\n",
      "Epoch [3/5], Step [200/468], Loss: 0.2908, Time: 20.9400 secs\n",
      "Epoch [3/5], Step [300/468], Loss: 0.2857, Time: 22.7841 secs\n",
      "Epoch [3/5], Step [400/468], Loss: 0.2917, Time: 24.5992 secs\n",
      "Epoch [4/5], Step [100/468], Loss: 0.3136, Time: 27.7169 secs\n",
      "Epoch [4/5], Step [200/468], Loss: 0.2937, Time: 29.5440 secs\n",
      "Epoch [4/5], Step [300/468], Loss: 0.3024, Time: 31.3981 secs\n",
      "Epoch [4/5], Step [400/468], Loss: 0.2916, Time: 33.2152 secs\n",
      "Epoch [5/5], Step [100/468], Loss: 0.3009, Time: 36.3059 secs\n",
      "Epoch [5/5], Step [200/468], Loss: 0.2981, Time: 38.1410 secs\n",
      "Epoch [5/5], Step [300/468], Loss: 0.3123, Time: 39.9771 secs\n",
      "Epoch [5/5], Step [400/468], Loss: 0.2789, Time: 41.8112 secs\n"
     ]
    }
   ],
   "source": [
    "# Train the model\n",
    "total_step = len(fmnist_train_loader)\n",
    "start_time = time.time()\n",
    "for epoch in range(num_epochs):\n",
    "    for i, (images, labels) in enumerate(fmnist_train_loader):\n",
    "        # each i is a batch of 128 samples\n",
    "        images = images.to(device).view(batch_size, -1)\n",
    "        \n",
    "        # Forward pass\n",
    "        outputs = model(images)\n",
    "        loss = criterion(outputs, images)\n",
    "        \n",
    "        # Backward and optimize - ALWAYS IN THIS ORDER!\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        if (i + 1) % 100 == 0:\n",
    "            print ('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}, Time: {:.4f} secs' \n",
    "                   .format(epoch + 1, num_epochs, i + 1, total_step, loss.item(), time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape: \n",
      " torch.Size([128, 1, 28, 28])\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# let's see some of the reconstructions\n",
    "model.eval()  # put in evaluation mode - no gradients\n",
    "examples = enumerate(fmnist_test_loader)\n",
    "batch_idx, (example_data, example_targets) = next(examples)\n",
    "print(\"shape: \\n\", example_data.shape)\n",
    "fig = plt.figure()\n",
    "for i in range(3):\n",
    "    ax = fig.add_subplot(2,3,i+1)\n",
    "    ax.imshow(example_data[i][0], cmap='gray', interpolation='none')\n",
    "    ax.set_title(\"Ground Truth: {}\".format(example_targets[i]))\n",
    "    ax.set_axis_off()\n",
    "    \n",
    "    ax = fig.add_subplot(2,3,i+4)\n",
    "    recon_img = model(example_data[i][0].view(1, -1).to(device)).data.cpu().numpy().reshape(28, 28)\n",
    "    ax.imshow(recon_img, cmap='gray')\n",
    "    ax.set_title(\"Reconstruction of: {}\".format(example_targets[i]))\n",
    "    ax.set_axis_off()\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# let's compare different dimensionality reduction methods\n",
    "n_neighbors = 10\n",
    "n_components = 2\n",
    "n_points = 500\n",
    "\n",
    "fmnist_test_loader = torch.utils.data.DataLoader(dataset=fmnist_test_dataset,\n",
    "                                          batch_size=n_points, \n",
    "                                          shuffle=False)\n",
    "X, labels = next(iter(fmnist_test_loader))\n",
    "latent_X = model.get_latent_rep(X.to(device).view(n_points, -1)).data.cpu().numpy()\n",
    "labels = labels.data.cpu().numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PCA time: 0.001 sec\n",
      "KPCA time: 0.009 sec\n",
      "LLE time: 0.083 sec\n",
      "Isomap time: 0.11 sec\n",
      "t-SNE time: 2.5 sec\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(15,8))\n",
    "\n",
    "# PCA\n",
    "t0 = time.time()\n",
    "x_pca = PCA(n_components).fit_transform(latent_X)\n",
    "t1 = time.time()\n",
    "print(\"PCA time: %.2g sec\" % (t1 - t0))\n",
    "ax = fig.add_subplot(2, 3, 1)\n",
    "ax.scatter(x_pca[:, 0], x_pca[:, 1], c=labels, cmap=plt.cm.Spectral)\n",
    "ax.set_title('PCA')\n",
    "\n",
    "# KPCA\n",
    "t0 = time.time()\n",
    "x_kpca = KernelPCA(n_components, kernel='rbf').fit_transform(latent_X)\n",
    "t1 = time.time()\n",
    "print(\"KPCA time: %.2g sec\" % (t1 - t0))\n",
    "ax = fig.add_subplot(2, 3, 2)\n",
    "ax.scatter(x_kpca[:, 0], x_kpca[:, 1], c=labels, cmap=plt.cm.Spectral)\n",
    "ax.set_title('KernelPCA')\n",
    "\n",
    "# LLE\n",
    "t0 = time.time()\n",
    "x_lle = LocallyLinearEmbedding(n_neighbors, n_components, eigen_solver='auto').fit_transform(latent_X)\n",
    "t1 = time.time()\n",
    "print(\"LLE time: %.2g sec\" % (t1 - t0))\n",
    "ax = fig.add_subplot(2, 3, 3)\n",
    "ax.scatter(x_lle[:, 0], x_lle[:, 1], c=labels, cmap=plt.cm.Spectral)\n",
    "ax.set_title('LLE')\n",
    "\n",
    "# Isomap\n",
    "t0 = time.time()\n",
    "x_isomap = Isomap(n_neighbors, n_components).fit_transform(latent_X)\n",
    "t1 = time.time()\n",
    "print(\"Isomap time: %.2g sec\" % (t1 - t0))\n",
    "ax = fig.add_subplot(2, 3, 4)\n",
    "ax.scatter(x_isomap[:, 0], x_isomap[:, 1], c=labels, cmap=plt.cm.Spectral)\n",
    "ax.set_title('Isomap')\n",
    "\n",
    "# t-SNE\n",
    "t0 = time.time()\n",
    "x_tsne = TSNE(n_components).fit_transform(latent_X)\n",
    "t1 = time.time()\n",
    "print(\"t-SNE time: %.2g sec\" % (t1 - t0))\n",
    "ax = fig.add_subplot(2, 3, 5)\n",
    "scatter = ax.scatter(x_tsne[:, 0], x_tsne[:, 1], c=labels, cmap=plt.cm.Spectral)\n",
    "ax.set_title('t-SNE')\n",
    "\n",
    "bounds = np.linspace(0, 10, 11)\n",
    "cb = plt.colorbar(scatter, spacing='proportional',ticks=bounds)\n",
    "cb.set_label('Classes Colors')\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### <img src=\"https://img.icons8.com/bubbles/50/000000/video-playlist.png\" style=\"height:50px;display:inline\"> Recommended Videos\n",
    "---\n",
    "#### <img src=\"https://img.icons8.com/cute-clipart/64/000000/warning-shield.png\" style=\"height:30px;display:inline\"> Warning!\n",
    "* These videos do not replace the lectures and tutorials.\n",
    "* Please use these to get a better understanding of the material, and not as an alternative to the written material.\n",
    "\n",
    "#### Video By Subject\n",
    "\n",
    "* PyTorch - <a href=\"https://www.youtube.com/watch?v=_H3aw6wkCv0&t=2s\">Stefan Otte: Deep Neural Networks with PyTorch | PyData Berlin 2018</a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## <img src=\"https://img.icons8.com/dusk/64/000000/prize.png\" style=\"height:50px;display:inline\"> Credits\n",
    "---\n",
    "* Icons from <a href=\"https://icons8.com/\">Icon8.com</a> - https://icons8.com\n",
    "* Datasets from <a href=\"https://www.kaggle.com/\">Kaggle</a> - https://www.kaggle.com/"
   ]
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}