project1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "315e734e",
   "metadata": {},
   "source": [
    "1. Make sure you fill in all cells contain `YOUR CODE HERE` or `YOUR ANSWER HERE`.\n",
    "2. After you finished, `Restart the kernel & run` all cell in order."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7874ac4b",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "7328d3b709977c10d706c7fdcdcaaf45",
     "grade": false,
     "grade_id": "cell-61902e9b43fe1aa6",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "# Project I: Basics of Neural Network\n",
    "## Deadline: Oct 14, 11:59 pm\n",
    "\n",
    "You have learned about the basics of neural network training and testing during the class. Now let's gain hands-on experience on how to implement them! We adopt the image classification task as a proxy to check how good your model is. Concretely, this project is split into two parts:\n",
    "\n",
    "- **Part I: play with feed-forward neural network**\n",
    "    - Build your feed-forward network with different layers and activation functions\n",
    "    - Define the gradient descent function to update the parameters\n",
    "    - Adjust the learning rate to achieve better performance \n",
    "    - Run the evaluation function\n",
    "\n",
    "\n",
    "- **Part II: implement your own Convolutional Neural Network**\n",
    "    - Train the CNN and compare it with the feed-forward neural network\n",
    "\n",
    "\n",
    "Let's get started!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f5511f2",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "50c9d0dbe6381074fc313d01b19ce31a",
     "grade": false,
     "grade_id": "cell-6fde9518151c06f0",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "## 1. Package\n",
    "\n",
    "Let's first import all the packages that you will need.\n",
    "\n",
    "- **torch, torch.nn, torch.nn.functional** are the fundamental modules in pytorch library, supporting Python programs that facilitates building deep learning projects.\n",
    "- **torchvision** is a library for Computer Vision that goes hand in hand with PyTorch\n",
    "- **numpy** is the fundamental package for scientific computing with Python programs.\n",
    "- **matplotlib** is a library to plot graphs and images in Python.\n",
    "- **math, random** are the standard modules in Python."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "9aeb862f",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "bbc346bdb6a037e971378bba3ca48310",
     "grade": true,
     "grade_id": "cell-95d842eb9dc9d856",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Import packages successfully!\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import random\n",
    "import math\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from project1_utils import *\n",
    "\n",
    "print(\"Import packages successfully!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "670c0b55",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "cc81b07f23da1a57e93b64b312300325",
     "grade": false,
     "grade_id": "cell-b528308013dd165f",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "A helper function is provided:\n",
    "\n",
    "```python\n",
    "def set_seed(seed):\n",
    "    \"\"\"\n",
    "    Use random seed to ensure that results are reproducible.\n",
    "    \"\"\"\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ae4d8bb5",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "0233fdb270c744b0966fad9b95fd06fa",
     "grade": true,
     "grade_id": "cell-017397b54ab692ac",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "seed = 1\n",
    "set_seed(seed)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7a530f8",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "43ca57a2b2923b23634d6e3cb069efda",
     "grade": false,
     "grade_id": "cell-ff95df93105ccae5",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "## 2. Dataset\n",
    "\n",
    "You will use the \"cat vs. dog\" dataset for this assignment.\n",
    "\n",
    "Let's load the dataset first using pytorch dataset and loader modules."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c63334fa",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "13c8e8f00b0979f91282fe06f2e4241e",
     "grade": true,
     "grade_id": "cell-f798e30fc96e134e",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples: 10000\n",
      "Number of testing examples: 2000\n"
     ]
    }
   ],
   "source": [
    "# the number of images in a batch\n",
    "batch_size = 32\n",
    "\n",
    "# load dataset\n",
    "trainset = dataset(path='/u/cs298/Fall2022/project1/dataset/trainset.h5')\n",
    "trainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size,\n",
    "                                          shuffle=True, num_workers=2)\n",
    "testset = dataset(path='/u/cs298/Fall2022/project1/dataset/testset.h5')\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=batch_size,\n",
    "                                         shuffle=False, num_workers=2)\n",
    "\n",
    "# name of classes\n",
    "classes = ('cat', 'dog')\n",
    "\n",
    "print (\"Number of training examples: \" + str(trainset.length))\n",
    "print (\"Number of testing examples: \" + str(testset.length))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb2b51b3",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "e2321117efbe3ffaee6eea5560ba5352",
     "grade": false,
     "grade_id": "cell-7081b78dc3a8d0ca",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "Let's visualize some examples in the dataset, the tool to show images is provided as below:\n",
    "\n",
    "```python\n",
    "def imshow(images):\n",
    "    \"\"\"\n",
    "    Display the input images in a plot\n",
    "    \"\"\"\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "34752a1c",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "6e922c75b1c2c40309886a987e0b9d50",
     "grade": true,
     "grade_id": "cell-1ad90426f0c574d0",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "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": "stdout",
     "output_type": "stream",
     "text": [
      "  cat   cat   cat   dog\n"
     ]
    }
   ],
   "source": [
    "# get some random training images\n",
    "dataiter = iter(trainloader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "num_toshow = 4\n",
    "\n",
    "# show images\n",
    "imshow(torchvision.utils.make_grid(images[:num_toshow]))\n",
    "\n",
    "# print labels\n",
    "_, indexs = torch.max(labels, 1) \n",
    "print(' '.join('%5s' % classes[indexs[j]] for j in range(num_toshow)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "45c58c54",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "faa6af0be51add948686ace9ca2f041f",
     "grade": false,
     "grade_id": "cell-d65c5901e6efc6e1",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "# Part I\n",
    "\n",
    "---\n",
    "\n",
    "## 3. Build your feedforward neural network.\n",
    "\n",
    "In this cell, you will be required to build a **three-layer multilayer perceptron (MLP)** to classify images into different categories. \n",
    "\n",
    "<!-- As we know from the class, **each layer** of a MLP can be denoted as the following mathematical operation:\n",
    "\n",
    "$$z = W^T x + b$$ $$a = \\sigma(z)$$\n",
    "\n",
    "Here, $W, b$ denote the weights and biases, and $a, \\sigma$ denote activation output and activation function, respectively.\n",
    "**The function is parameterized by $W, b$ as well as the choice of $\\sigma(\\cdot)$**.\n",
    "\n",
    "Note that it is valid for $\\sigma(\\cdot)$ to be the identity function, or $z = \\sigma(z)$.\n",
    "\n",
    "----\n",
    "\n",
    "**Question 1 (6 points):** Now, let's implement functions at the layer level to do the following:\n",
    "\n",
    "Hint: To implement $W^Tx+b$ in PyTorch, one way is to write it as `x.mm(W) + b`. -->"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f616622",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "290e401fa444fe37762bec3d48da6909",
     "grade": false,
     "grade_id": "cell-5f86826cc8f017c2",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "The size of input images is a batch-like tensor $ X \\in \\mathbb{R}^{B \\times C \\times H \\times W}$, where $B$ denotes the batch size. Vectorize the image pixels equals to transforming into a vector $X_{vector} \\in \\mathbb{R}^{B \\times CHW}$.\n",
    "\n",
    "**Question 1 (4 points):** To process image data with feed-forward neural network, the image pixels should be vectorized. So first, implement a vectorized function (refer to lecture \"feedforward network\"). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "284f0306",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "d4ba7091c8beb94974af923d3a821caa",
     "grade": true,
     "grade_id": "cell-1d5278bf7c8bf138",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def image_vectorization(image_batch):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        image_batch: a batch of images with shape [b, c, h, w]\n",
    "    Output: \n",
    "        vectorized_image_batch: a batch of neurons\n",
    "    \"\"\"\n",
    "    \n",
    "    # vectorize the image pixels\n",
    "    vectorized_image_batch = []\n",
    "    for image in image_batch:\n",
    "        vectorized_image_batch.append(torch.flatten(image))\n",
    "    \n",
    "    return torch.stack(vectorized_image_batch)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1059bbbb",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "1db4cd393cb99536cdea1d033e0c9bf4",
     "grade": false,
     "grade_id": "cell-aee10b8c60cb444f",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "As we know from the class, **each layer** of a MLP can be denoted as the following mathematical operation:\n",
    "\n",
    "$$z = W^T x + b$$ \n",
    "\n",
    "Here, $W, b$ denote the weights and biases. The function is **parameterized by $W, b$**.\n",
    "\n",
    "**Question 2 (4 points):** Given the desired input, output dimensions, generate the parameters $W, b$. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "32728243",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "534aed699bf1cfd761fe20a33b87dbf9",
     "grade": true,
     "grade_id": "cell-24644c097079f609",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def get_layer_params(input_dim: int, output_dim: int):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        input_dim: number of neurons in the input\n",
    "        output_dim: number of neurons produced by the layer\n",
    "    Output: \n",
    "        a dictionary of generated parameters\n",
    "            - w: weights\n",
    "            - b: biases\n",
    "    \"\"\"\n",
    "    \n",
    "    # generate the parameters\n",
    "    w = torch.randn(input_dim, output_dim, requires_grad = True)\n",
    "    b = torch.randn(output_dim, requires_grad = True)\n",
    "    \n",
    "    return {'w': w,\n",
    "            'b': b}\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d7f1fa8",
   "metadata": {},
   "source": [
    "Following with the previous linear layer, an activation layer is required to add non-linearity to the network:\n",
    "\n",
    " $$a = \\sigma(z)$$\n",
    "\n",
    " $a, \\sigma$ denote activation output and activation function, respectively.\n",
    "The entire layer function is also **parameterized by choice of $\\sigma(\\cdot)$**.\n",
    "\n",
    "**Question 3 (4 points):** You need an activation wrapper function to support the following three activation functions (details are in the lecture slide).\n",
    "- Sigmoid\n",
    "- tanh\n",
    "- ReLU\n",
    "\n",
    "Note: don't directly call `torch.Sigmoid` and etc. Your implementation should follow their mathematical equation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "34390a64",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "b1e48e7e30fe99b03487f3e269501a06",
     "grade": true,
     "grade_id": "cell-4cd80a23d639a9cf",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def activation_wrapper(x, activation='relu'):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        x: the input neuron values\n",
    "        activation: name of activation, could be one in ['relu', 'sigmoid', 'tanh']\n",
    "    Output: \n",
    "        a: the corresponding activated output\n",
    "    \"\"\"\n",
    "\n",
    "    if activation == 'relu':\n",
    "        for neuron in x:\n",
    "            idxs = neuron < 0\n",
    "            neuron[idxs] = 0\n",
    "        a = x\n",
    "    elif activation == 'sigmoid':\n",
    "        a = 1 / (1 + torch.exp(-x))\n",
    "    elif activation == 'tanh':\n",
    "        a = (torch.exp(-x) - torch.exp(-x)) / (torch.exp(x) + torch.exp(-x))\n",
    " \n",
    "    return a\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8216ddce",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "8e8b4e28e72cb6de3fb0c20f631e88b4",
     "grade": false,
     "grade_id": "cell-37d560a619728dd8",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "**Question 4 (4 points):** Given the layer parameters $W, b$ and the choice of $\\sigma(\\cdot)$, compute the output for an MLP layer with input $x$. \n",
    "\n",
    "Hint:  `tensor.mm()` can be useful."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c1ed6494",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "4a7ffa1779e7005c993ee1d874b9222d",
     "grade": true,
     "grade_id": "cell-3b88b8f84dd5393b",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def layer_forward_computation(x, params, activation):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        x: the input to the layer\n",
    "        params: parameters of each layer\n",
    "        activation: activation type\n",
    "    Output: \n",
    "        a: the output after the activation\n",
    "    \"\"\"\n",
    "    \n",
    "    # compute the output for layer\n",
    "    layer_output = torch.mm(x, params['w']) + params['b']\n",
    "    a = activation_wrapper(layer_output, activation)\n",
    "    \n",
    "    return a"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "607eef82",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "cb331aebbde5d8fe52457510dc5095fd",
     "grade": false,
     "grade_id": "cell-aadf7e24fd04c923",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "---\n",
    "\n",
    "Back to building our three-layer MLP for classification. If you have implemented the functions above correctly,\n",
    "now the processing of putting everything together will be very easy.\n",
    "\n",
    "Just like other parts of your programming experience,\n",
    "knowing how to efficiently abstract and modularize components of your program will be critical in deep learning.\n",
    "\n",
    "**Architecture Requirement**:\n",
    "\n",
    "We now describe in details how our three-layer MLP should be built in PyTorch.\n",
    "\n",
    "1. In the dataset, the size of input image is a tensor $ X \\in \\mathbb{R}^{B \\times 3 \\times 32 \\times 32}$, where $B$ denotes the batch size.\n",
    "2. Vectorize the image pixels to a vector $X_{vector} \\in \\mathbb{R}^{B \\times 3072}$.\n",
    "3. We now begin describing the specific architecture of the model, although this is not the only design choice, and feel free to change the hidden dimensions of the parameters\n",
    "4. Layer1: set your parameters so the input is projected from $\\mathbb{R}^{B \\times 3072}$ to $\\mathbb{R}^{B \\times 256}$, use ReLU as your activation function\n",
    "5. Layer2: set your parameters so the input is projected from $\\mathbb{R}^{B \\times 256}$ to $\\mathbb{R}^{B \\times 128}$, use ReLU as your activation function\n",
    "6. Layer3: set your parameters so the input is projected from $\\mathbb{R}^{B \\times 128}$ to $\\mathbb{R}^{B \\times 2}$, use sigmoid function as your activation function\n",
    "\n",
    "---\n",
    "\n",
    "**Question 5 (4 points):** Let's build the three-layer MLP using the functions you have defined."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b25124b1",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "13cccd4e9498d87add4227c1fb297253",
     "grade": true,
     "grade_id": "cell-46335680e32c5b0f",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "layer1_params: dict = dict()\n",
    "layer2_params: dict = dict()\n",
    "layer3_params: dict = dict()\n",
    "\n",
    "def net(X, params, activations):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        X: the input images to the network\n",
    "        params: a dictionary of parameters(W and b) for the three different layers\n",
    "        activations: a dictionary of activation function names for the three different layers\n",
    "    Output: \n",
    "        output: the final output from the third layer\n",
    "    \"\"\"\n",
    "    \n",
    "    # build your network forward\n",
    "    output = image_vectorization(X)\n",
    "\n",
    "    for layer in params.keys():\n",
    "        output = layer_forward_computation(output, params[layer], activations[layer])\n",
    "\n",
    "    return output"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36679670",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "710dc3453a8f1ea60791abe66ba4e408",
     "grade": false,
     "grade_id": "cell-2de628ff93729043",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "**Question 6 (4 points):** All the required functions are defined right now. It's time to create your actual parameters and activation functions for the network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "abe70b2d",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "cd83a048b15490dc8969afc6fc292df5",
     "grade": true,
     "grade_id": "cell-8611eb15d3df224c",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "\"\"\" We prepare serval dictories to store the parameters and activations for different   \"\"\"\n",
    "layer1_params: dict = dict()\n",
    "layer2_params: dict = dict()\n",
    "layer3_params: dict = dict()\n",
    "params: dict = dict()\n",
    "activations: dict = dict()\n",
    "    \n",
    "# TODO: define your layer parameters and activations here. For example, layer1_params = ? \n",
    "# and pack your parameters with layer name, e.g., params['layer1'] = layer1_params\n",
    "\n",
    "layer1_params['input_dim'] = 3072\n",
    "layer1_params['output_dim'] = 256\n",
    "layer1_params.update(get_layer_params(layer1_params['input_dim'], layer1_params['output_dim']))\n",
    "\n",
    "layer2_params['input_dim'] = 256\n",
    "layer2_params['output_dim'] = 128\n",
    "layer2_params.update(get_layer_params(layer2_params['input_dim'], layer2_params['output_dim']))\n",
    "\n",
    "layer3_params['input_dim'] = 128\n",
    "layer3_params['output_dim'] = 2\n",
    "layer3_params.update(get_layer_params(layer3_params['input_dim'], layer3_params['output_dim']))\n",
    "\n",
    "params['layer1'] = layer1_params\n",
    "params['layer2'] = layer2_params\n",
    "params['layer3'] = layer3_params\n",
    "\n",
    "activations['layer1'] = 'relu'\n",
    "activations['layer2'] = 'relu'\n",
    "activations['layer3'] = 'sigmoid'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ef13176",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "d6df19e943fb3f604254e4c0b4ee3953",
     "grade": false,
     "grade_id": "cell-e014df9a9afed7e9",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "## 4. Backpropagation and optimization\n",
    "\n",
    "After finishing the forward pass, you now need to compute gradients for all Tensors with `requires_grad=True`, e.g., parameters of layer1. These gradients will be used to update parameters via gradient descent. \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "70a89c01",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "2577dc212e328a104775317a5d20fedd",
     "grade": false,
     "grade_id": "cell-25bcd84dc444c485",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "In this project, we don't require you to compute the gradients. Instead, you will use the gradients computed with Pytorch `autograd` mechanism. We want you to implement the Stochastic Gradient Descent (SGD). This will help you understand how the parameters are updated during training.\n",
    "\n",
    "---\n",
    "\n",
    "Gradient descent is a way to minimize the final objective function (loss) parameterized by a model's parameter $\\theta$ by updating the parameters in the opposite direction of the gradient $\\nabla_\\theta J(\\theta)$ w.r.t to the parameters. The learning rate $\\lambda$ determines the size of the steps you take to reach a (local) minimum.\n",
    "\n",
    "However, for the vanilla gradient descent, you need to run through all the samples in your training set and update once. This will be time-consuming with large-scale datasets. We choose Stochastic Gradient Descent, which only requires a subset of training samples to update the parameters. With the popular deep learning framework, the subset usually equals to the minibatch selected during training.\n",
    "\n",
    "Now, let's look at the equation to update parameters for each layer in your network.\n",
    "\n",
    "$$\\large \\theta = \\theta - \\lambda\\cdot\\nabla_\\theta J(\\theta)$$\n",
    "\n",
    "---\n",
    "\n",
    "**Question 7 (4 points):** Following the equation above, you will use the computed gradients to update the parameters for feedforward network. Function **update_params( )** should be completed.\n",
    "\n",
    "Hint: To get gradient and operate it in PyTorch, you can call `x.grad`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "758cceed",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "2b4f63b3e05a2bb8bb14eb8b7c7e37eb",
     "grade": true,
     "grade_id": "cell-3ace7576abd4e7f8",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def update_params(params, learning_rate):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        params: the dictornary to store all the layer parameters\n",
    "        learning_rate: the step length to update the parameters\n",
    "    Output: \n",
    "        params: the updated parameters\n",
    "    \"\"\"\n",
    "    with torch.no_grad():\n",
    "        for layer in params.keys():\n",
    "            params[layer]['w'] -= learning_rate * params[layer]['w'].grad\n",
    "            params[layer]['b'] -= learning_rate * params[layer]['b'].grad\n",
    "    \n",
    "    return params"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8026ef65",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "8cd4d30cf9052a7203af2ab335c38895",
     "grade": false,
     "grade_id": "cell-c487b54835c43813",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "Since you are updating the parameters for each batch of data iteratively, you will need to clear the gradients after each update. \n",
    "\n",
    "**Question 8 (4 points):** Complete the function **zero_grad( )** below to manually zero the gradients after updating weights. \n",
    "\n",
    "\n",
    "<!-- **Requirements:** You will need to complete the function **backprop( )** and **zero_grad( )** and implement them in the training process. \n",
    "\n",
    "Hint1: You can use `autograd` in PyTorch to compute gradients.\n",
    "\n",
    "Hint2: You should manually zero the gradients after updating weights. -->\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "50abdc66",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "6033ac84a5a191cb4e402eae4c7c4987",
     "grade": true,
     "grade_id": "cell-e276c24353feed3b",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def zero_grad(params):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        params: the dictornary to store all the layer parameters\n",
    "    Output: \n",
    "        params: the updated parameters with gradients clear\n",
    "    \"\"\"\n",
    "    #TODO: set the gradients with respect to parameters as zero\n",
    "    for layer in params.keys():\n",
    "        params[layer]['w'].grad = None\n",
    "        params[layer]['b'].grad = None\n",
    "    \n",
    "    return params\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d96138ea",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "627f2464dd1662519c7a95a8c81438c6",
     "grade": false,
     "grade_id": "cell-b3fa3df7f9879872",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "With the function **update_params( )** and **zero_grad( )** you have defined, you can move to the backpropagation process. The process includes computing gradients, updating parameters, reset gradients, think of how to combine them.\n",
    "\n",
    "**Question 9 (4 points):** Complete the function **backprop( )** below. \n",
    "\n",
    "Hint: You can use `backward()` in PyTorch to compute gradients.\n",
    "\n",
    "<!-- **Requirements:** You will need to complete the function **backprop( )** and **zero_grad( )** and implement them in the training process. \n",
    "\n",
    "Hint1: You can use `autograd` in PyTorch to compute gradients.\n",
    "\n",
    "Hint2: You should manually zero the gradients after updating weights. -->\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d41c6fff",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "9347587e1f37c138024725f5d7afda8e",
     "grade": true,
     "grade_id": "cell-fc06f560b737cec9",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def backprop(loss, params, learning_rate):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        loss: the loss tensor from the objective funtion that can be used to compute gradients\n",
    "        params: parameters of the three layers\n",
    "        learning_rate: the size of steps when updating parameters\n",
    "    Output:\n",
    "        params: parameters after one backpropogation\n",
    "    \"\"\"    \n",
    "    \n",
    "    loss.backward()\n",
    "    udpated_params = update_params(params, learning_rate)\n",
    "    params = zero_grad(udpated_params)\n",
    "\n",
    "    return params\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2de02f38",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "4909b7f8a4bb572ce7d55d65c9286618",
     "grade": false,
     "grade_id": "cell-54f264706094e49e",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "\n",
    "## 5. Training loop\n",
    "\n",
    "For this binary classification task, a standard objective function **Binary Cross-Entropy Loss** is used. Related detail is given as follows:\n",
    "\n",
    "$$\\large L = -\\frac{1}{N}\\sum_{i=1}^{N}( y_i \\cdot \\log(p(y_i))+(1-y_i)\\log(1-p(y_i)))$$\n",
    "\n",
    "where $y$ is the label (1 for dog and 0 for cat in our case) and $p(y)$ is the predicted probability, here $N$ equals to the batch_size.\n",
    "\n",
    "\n",
    "A initialization function is provided to help your network converge faster.\n",
    "\n",
    "```python\n",
    "def init_params(params):\n",
    "    \"\"\"\n",
    "    Initialize the parameters of each layer\n",
    "    \"\"\"\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ac75d74f",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "8515b8629f66b11ec06ce677c853db6f",
     "grade": false,
     "grade_id": "cell-968dae4544884930",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "Before moving into the training loop, it's usually a good practice to have a learning rate decay function. The reason is that when your model is training for a longer time, it's closer to the optimal convergence. Therefore, a lower learning rate will improve the learning of complex patterns.\n",
    "\n",
    "**Question 10 (4 points):** Complete the **adjust_lr** function, you can choose to decay by 0.1 every 15 epochs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "17a057c8",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "dec0de0a4200a50cb52289dd9c2fb8b4",
     "grade": true,
     "grade_id": "cell-66c9e14d5eb9189d",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "def adjust_lr(learning_rate, epoch):\n",
    "    \"\"\"\n",
    "    Input: \n",
    "        learning_rate: the input learning rate\n",
    "        epoch: which epoch you are in\n",
    "    Output:\n",
    "        learning_rate: the updated learning rate\n",
    "    \"\"\"    \n",
    "\n",
    "    if (epoch % 15 == 0):\n",
    "        learning_rate -= 0.001\n",
    "        \n",
    "    return learning_rate\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c9ed12c6",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "85d20fdd7bd1ef0b7cd90e875736c499",
     "grade": true,
     "grade_id": "cell-e6824e4b20c49f4a",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Epoch 1, Step   200] loss: 0.735\n",
      "[Epoch 2, Step   200] loss: 0.630\n",
      "[Epoch 3, Step   200] loss: 0.589\n",
      "[Epoch 4, Step   200] loss: 0.563\n",
      "[Epoch 5, Step   200] loss: 0.539\n",
      "[Epoch 6, Step   200] loss: 0.517\n",
      "[Epoch 7, Step   200] loss: 0.488\n",
      "[Epoch 8, Step   200] loss: 0.464\n",
      "[Epoch 9, Step   200] loss: 0.446\n",
      "[Epoch 10, Step   200] loss: 0.427\n",
      "[Epoch 11, Step   200] loss: 0.410\n",
      "[Epoch 12, Step   200] loss: 0.388\n",
      "[Epoch 13, Step   200] loss: 0.371\n",
      "[Epoch 14, Step   200] loss: 0.344\n",
      "[Epoch 15, Step   200] loss: 0.334\n",
      "[Epoch 16, Step   200] loss: 0.317\n",
      "[Epoch 17, Step   200] loss: 0.288\n",
      "[Epoch 18, Step   200] loss: 0.277\n",
      "[Epoch 19, Step   200] loss: 0.265\n",
      "[Epoch 20, Step   200] loss: 0.244\n",
      "[Epoch 21, Step   200] loss: 0.230\n",
      "[Epoch 22, Step   200] loss: 0.216\n",
      "[Epoch 23, Step   200] loss: 0.208\n",
      "[Epoch 24, Step   200] loss: 0.192\n",
      "[Epoch 25, Step   200] loss: 0.180\n",
      "[Epoch 26, Step   200] loss: 0.173\n",
      "[Epoch 27, Step   200] loss: 0.160\n",
      "[Epoch 28, Step   200] loss: 0.144\n",
      "[Epoch 29, Step   200] loss: 0.141\n",
      "[Epoch 30, Step   200] loss: 0.126\n",
      "Finished Training\n"
     ]
    }
   ],
   "source": [
    "# define the initial learning rate here\n",
    "learning_rate = 1e-2\n",
    "n_epochs = 30 # how many epochs to run\n",
    "\n",
    "# define loss function\n",
    "criterion = nn.BCELoss()\n",
    "\n",
    "# initialize network parameters\n",
    "init_params(params)\n",
    "    \n",
    "for epoch in range(n_epochs):  # loop over the dataset multiple times\n",
    "\n",
    "    running_loss = 0.0\n",
    "    for i, data in enumerate(trainloader, 0):\n",
    "\n",
    "        # get the inputs; data is a list of [inputs, labels]\n",
    "        inputs, labels = data\n",
    "        labels = labels.float()\n",
    "\n",
    "        # Forward \n",
    "        output = net(inputs, params, activations)\n",
    "        \n",
    "        # Compute the loss using the final output\n",
    "        loss = criterion(output, labels)\n",
    "\n",
    "        # Backpropagation\n",
    "        params = backprop(loss, params, learning_rate)\n",
    "        \n",
    "        # print statistics\n",
    "        running_loss += loss.item()\n",
    "        if i % 200 == 199:  # print every 200 mini-batches\n",
    "            print('[Epoch %d, Step %5d] loss: %.3f' %\n",
    "                  (epoch + 1, i + 1, running_loss / 200))\n",
    "            running_loss = 0.0\n",
    "            \n",
    "    # adjust learning rate\n",
    "    learning_rate = adjust_lr(learning_rate, epoch)\n",
    "print('Finished Training')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60078912",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "013d59c696855d6bc7354ec2cba558af",
     "grade": false,
     "grade_id": "cell-2a7f9c550151464e",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "## 6. Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "cb9dbfa1",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "2a8aac4443a5f12f8845d549a9c1204c",
     "grade": true,
     "grade_id": "cell-a5ef30cda9f342ab",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "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": "stdout",
     "output_type": "stream",
     "text": [
      "GroundTruth:    cat   cat   cat   cat\n",
      "Predicted:    cat   dog   dog   cat\n"
     ]
    }
   ],
   "source": [
    "dataiter = iter(testloader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "# print images\n",
    "_, labels = torch.max(labels, 1)\n",
    "imshow(torchvision.utils.make_grid(images))\n",
    "print('GroundTruth: ', ' '.join('%5s' % classes[labels[j]] for j in range(4)))\n",
    "\n",
    "output = net(images, params, activations)\n",
    "_, predicted = torch.max(output, 1)\n",
    "\n",
    "print('Predicted: ', ' '.join('%5s' % classes[predicted[j]]\n",
    "                              for j in range(4)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e01d5189",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "0272518b30b12353d9fbd4915d58359b",
     "grade": false,
     "grade_id": "cell-9d4eb6e5d234c194",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "**Evaluation**: Now testing with your trained model!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "5d45304e",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "7feb0df201f67601f3b831a78e786e33",
     "grade": true,
     "grade_id": "cell-ab0a206ec87856f5",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the 2000 test images: 60 %\n"
     ]
    }
   ],
   "source": [
    "correct = 0\n",
    "total = 0\n",
    "\n",
    "# since you're not training, you don't need to calculate the gradients for our outputs\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        images, labels = data\n",
    "        _, labels = torch.max(labels, 1)\n",
    "        \n",
    "        # calculate outputs by running images through the network\n",
    "        output = net(images, params, activations)\n",
    "\n",
    "        # the class with the highest energy is what we choose as prediction\n",
    "        _, predicted = torch.max(output.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "print('Accuracy of the network on the 2000 test images: %d %%' % (\n",
    "        100 * correct / total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "0dda3a70",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "ac1329a33c994bf1a50c89d85f94bd8f",
     "grade": true,
     "grade_id": "cell-77cad235a42d68f5",
     "locked": true,
     "points": 0,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy for class cat   is: 72.7 %\n",
      "Accuracy for class dog   is: 48.4 %\n"
     ]
    }
   ],
   "source": [
    "# prepare to count predictions for each class\n",
    "correct_pred = {classname: 0 for classname in classes}\n",
    "total_pred = {classname: 0 for classname in classes}\n",
    "\n",
    "# again no gradients needed\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        images, labels = data\n",
    "        _, labels = torch.max(labels, 1)\n",
    "        output = net(images, params, activations)\n",
    "        _, predictions = torch.max(output, 1)\n",
    "        # collect the correct predictions for each class\n",
    "        for label, prediction in zip(labels, predictions):\n",
    "            if label == prediction:\n",
    "                correct_pred[classes[label]] += 1\n",
    "            total_pred[classes[label]] += 1\n",
    "\n",
    "# print accuracy for each class\n",
    "for classname, correct_count in correct_pred.items():\n",
    "    accuracy = 100 * float(correct_count) / total_pred[classname]\n",
    "    print(\"Accuracy for class {:5s} is: {:.1f} %\".format(classname,\n",
    "                                                         accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23be2bff",
   "metadata": {
    "deletable": false,
    "editable": false,
    "nbgrader": {
     "cell_type": "markdown",
     "checksum": "a0daf457d2d6b94ffc8eaf4204abf259",
     "grade": false,
     "grade_id": "cell-a9fd8b383786bcf5",
     "locked": true,
     "schema_version": 3,
     "solution": false,
     "task": false
    }
   },
   "source": [
    "# Part II\n",
    "\n",
    "---\n",
    "\n",
    "## 7. Build your convolutional neural network.\n",
    "\n",
    "So far, you have tried feed-forward network on image classification, but the performance is not satisfying. Why? \n",
    "The reason is that vectorizing the images will lose some critical patterns of images, such as edges, corners, and local structures. Instead, convolutional neural network is very good at capturing these patterns. \n",
    "\n",
    "To explore this, let's build a CNN to see how good it is in image classification!\n",
    "\n",
    "---\n",
    "\n",
    "**Question 11 (8 points):** Build your CNN: You will build a two-layer CNN with a maxpooling layer in between.\n",
    "Note that one Fully-Connected (FC) layer will follow the CNN network to map image features into class features. Overall, this network is similar to the one you built above, with the first two feed-forward layers being replaced by the convolutional layers:\n",
    "\n",
    "            image -> [CNN layer 1] -> [CNN layer 2] -> vectorization -> [FC layer] -> prediction\n",
    "\n",
    "**Architecture Requirement**:\n",
    "\n",
    "1. CNN Layer1: suggests that **3 or 5** as your convolution kernel size; the number of output channels can be selected from **[16, 32, 64]**; use ReLU as your activation function\n",
    "2. CNN Layer2: suggests that **3 or 5** as your convolution kernel size; the number of output channels can be selected from **[128, 256]**; use ReLU as your activation function\n",
    "3. FC layer: set your parameters, so the input is projected from $\\mathbb{R}^{B \\times N}$ to $\\mathbb{R}^{B \\times 2}$, $N$ is defined by your CNN layers' parameters; use sigmoid function as your activation function\n",
    "\n",
    "Details and hints:\n",
    "- This time you don't need to write the whole network module. Instead, we provide `nn.Module` as a network wrapper (a base class for all neural network modules in `PyTorch` library). The parameters you defined in this module supports auto gradient computation. \n",
    "- For the convolutional layers, you can call `nn.Conv2d`. A convolutional layer usually contains a convolution, an activation and a pooling layer. You can refer to this [useful link](https://pytorch.org/docs/stable/nn.html) to find the layer you want to use.\n",
    "- Choose appropriate activation functions for all layers.\n",
    "- As you need to define the input/output dimension for an FC layer, this [tutorial](https://cs231n.github.io/convolutional-networks/) will help you understand how to calculate this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "fa6cce6e",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "8fb18111f63f223a1c834bc891886542",
     "grade": true,
     "grade_id": "cell-85f8f7ed1b8add0e",
     "locked": false,
     "points": 8,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [],
   "source": [
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        \n",
    "        # Define your layers here!! \n",
    "        self.layer1 = nn.Sequential(nn.Conv2d(3, 64, 3),\n",
    "                                    nn.ReLU(),\n",
    "                                    nn.MaxPool2d(2)\n",
    "                                    )\n",
    "        self.layer2 = nn.Sequential(nn.Conv2d(64, 256, 3),\n",
    "                                    torch.nn.ReLU(),\n",
    "                                    torch.nn.MaxPool2d(2)\n",
    "                                    )\n",
    "        self.layer3 = nn.Sequential(nn.Linear(256 * 6 * 6, 2),\n",
    "                                    nn.Sigmoid())\n",
    "        \n",
    "    def forward(self, x):\n",
    "        # Your forward pass with the defined layers\n",
    "        output = self.layer1(x)\n",
    "        output = self.layer2(output)\n",
    "        output = output.reshape(output.size(0), -1)\n",
    "        output = self.layer3(output)\n",
    "        return output"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4de994b4",
   "metadata": {},
   "source": [
    "**Question 12 (4 points):** Here you don't need to write in the details of updating network parameters, we provide `optim.SGD` to help you do this. However, you still need to finish the backpropagation procedure in the main training loop. \n",
    "\n",
    "Hint: the `optimizer` implement functions `step()` to update the parameters and `zero_grad()` to reset the gradients.\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "6b5ded0d",
   "metadata": {
    "deletable": false,
    "nbgrader": {
     "cell_type": "code",
     "checksum": "584db35ccf940f5b0c1c23385ccdb63b",
     "grade": true,
     "grade_id": "cell-258330ae36298fb1",
     "locked": false,
     "points": 4,
     "schema_version": 3,
     "solution": true,
     "task": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Epoch 1, Step   200] loss: 0.682\n",
      "[Epoch 2, Step   200] loss: 0.658\n",
      "[Epoch 3, Step   200] loss: 0.648\n",
      "[Epoch 4, Step   200] loss: 0.637\n",
      "[Epoch 5, Step   200] loss: 0.632\n",
      "[Epoch 6, Step   200] loss: 0.622\n",
      "[Epoch 7, Step   200] loss: 0.621\n",
      "[Epoch 8, Step   200] loss: 0.615\n",
      "[Epoch 9, Step   200] loss: 0.603\n",
      "[Epoch 10, Step   200] loss: 0.595\n",
      "[Epoch 11, Step   200] loss: 0.590\n",
      "[Epoch 12, Step   200] loss: 0.582\n",
      "[Epoch 13, Step   200] loss: 0.579\n",
      "[Epoch 14, Step   200] loss: 0.570\n",
      "[Epoch 15, Step   200] loss: 0.562\n",
      "[Epoch 16, Step   200] loss: 0.558\n",
      "[Epoch 17, Step   200] loss: 0.551\n",
      "[Epoch 18, Step   200] loss: 0.545\n",
      "[Epoch 19, Step   200] loss: 0.545\n",
      "[Epoch 20, Step   200] loss: 0.532\n",
      "[Epoch 21, Step   200] loss: 0.533\n",
      "[Epoch 22, Step   200] loss: 0.526\n",
      "[Epoch 23, Step   200] loss: 0.523\n",
      "[Epoch 24, Step   200] loss: 0.518\n",
      "[Epoch 25, Step   200] loss: 0.511\n",
      "[Epoch 26, Step   200] loss: 0.515\n",
      "[Epoch 27, Step   200] loss: 0.507\n",
      "[Epoch 28, Step   200] loss: 0.506\n",
      "[Epoch 29, Step   200] loss: 0.503\n",
      "[Epoch 30, Step   200] loss: 0.490\n",
      "Finished Training\n"
     ]
    }
   ],
   "source": [
    "# define the initial learning rate here\n",
    "learning_rate = 1e-2\n",
    "n_epochs = 30 # how many epochs to run\n",
    "# define loss function\n",
    "criterion = nn.BCELoss()\n",
    "cnn_net = Net()\n",
    "cnn_net = cnn_net.cuda()\n",
    "optimizer = torch.optim.SGD(cnn_net.parameters(), lr=learning_rate)\n",
    "\n",
    "for epoch in range(n_epochs):  # loop over the dataset multiple times\n",
    "\n",
    "    running_loss = 0.0\n",
    "    for i, data in enumerate(trainloader, 0):\n",
    "\n",
    "        # get the inputs; data is a list of [inputs, labels]\n",
    "        inputs, labels = data\n",
    "        labels = labels.float()\n",
    "        inputs = inputs.cuda()\n",
    "        labels = labels.cuda()\n",
    "\n",
    "        # Forward \n",
    "        output = cnn_net(inputs)\n",
    "        \n",
    "        # Compute the loss using the final output\n",
    "        loss = criterion(output, labels)\n",
    "\n",
    "        # Backpropagation\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        # print statistics\n",
    "        running_loss += loss.item()\n",
    "        if i % 200 == 199:  # print every 200 mini-batches\n",
    "            print('[Epoch %d, Step %5d] loss: %.3f' %\n",
    "                  (epoch + 1, i + 1, running_loss / 200))\n",
    "            running_loss = 0.0\n",
    "\n",
    "print('Finished Training')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "9bfc52fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the 2000 test images: 74 %\n"
     ]
    }
   ],
   "source": [
    "correct = 0\n",
    "total = 0\n",
    "\n",
    "# since you're not training, you don't need to calculate the gradients for our outputs\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        images, labels = data\n",
    "        images = images.cuda()\n",
    "        labels = labels.cuda() \n",
    "        _, labels = torch.max(labels, 1)\n",
    "        \n",
    "        # calculate outputs by running images through the network\n",
    "        output = cnn_net(images)\n",
    "\n",
    "        # the class with the highest energy is what we choose as prediction\n",
    "        _, predicted = torch.max(output.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "print('Accuracy of the network on the 2000 test images: %d %%' % (\n",
    "        100 * correct / total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "ea9d735d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy for class cat   is: 79.0 %\n",
      "Accuracy for class dog   is: 69.0 %\n"
     ]
    }
   ],
   "source": [
    "# prepare to count predictions for each class\n",
    "correct_pred = {classname: 0 for classname in classes}\n",
    "total_pred = {classname: 0 for classname in classes}\n",
    "\n",
    "# again no gradients needed\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        images, labels = data\n",
    "        images = images.cuda()\n",
    "        labels = labels.cuda() \n",
    "        _, labels = torch.max(labels, 1)\n",
    "        output = cnn_net(images)\n",
    "        _, predictions = torch.max(output, 1)\n",
    "        # collect the correct predictions for each class\n",
    "        for label, prediction in zip(labels, predictions):\n",
    "            if label == prediction:\n",
    "                correct_pred[classes[label]] += 1\n",
    "            total_pred[classes[label]] += 1\n",
    "\n",
    "# print accuracy for each class\n",
    "for classname, correct_count in correct_pred.items():\n",
    "    accuracy = 100 * float(correct_count) / total_pred[classname]\n",
    "    print(\"Accuracy for class {:5s} is: {:.1f} %\".format(classname,\n",
    "                                                         accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db386d42",
   "metadata": {},
   "source": [
    "# Submission\n",
    "1. At the menubar, click `Kernel` → `Restart & Run All`\n",
    "2. Pack the completed notebook and other project materials into a zip file and upload it via blackboard. The zip file should be named as \"[NetID]_proj1.zip\"\n",
    "\n",
    "1% deduction of late assignment total score per hour passing the deadline.\n",
    "\n",
    "We allow multiple submission attempts but only the latest submission will be considered."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6bb4b70c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "celltoolbar": "Edit 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": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}