FILIOT_ALEXANDRE.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "FILIOT_ALEXANDRE.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/afiliot/DLiP-TP3/blob/master/FILIOT_ALEXANDRE.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "PAqKzYI_1wpo",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Small data and deep learning\n",
        "This mini-project proposes to study several techniques for improving challenging context, in which few data and resources are available."
      ]
    },
    {
      "metadata": {
        "id": "7zcDTpSB1wpw",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Training set creation\n",
        "__Question 1:__ Propose a dataloader or modify the file located at https://github.com/pytorch/vision/blob/master/torchvision/datasets/cifar.py in order to obtain a training loader that will only use the first 100 samples of the CIFAR-10 training set. "
      ]
    },
    {
      "metadata": {
        "id": "unSUwk3NNN8V",
        "colab_type": "code",
        "outputId": "71bf13c4-00b4-4329-f395-43f832d50bd4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 32
        }
      },
      "cell_type": "code",
      "source": [
        "#from google.colab import drive\n",
        "#drive.mount('/content/drive')"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ig8wOjoWCVGN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> **Set your path: it should end by /FILIOT_ALEXANDRE/**. The folder is the zip file which contains the folders simulations and weights."
      ]
    },
    {
      "metadata": {
        "id": "hgqju02ECUJR",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "path = input()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "YexKar2iuA1F",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> In the wording, we assume that we \"have a limited computational budget (e.g., no GPUs)\". On collab, one can have access to GPUs, which make the trainings much faster and convenient for doing cross-validations, gridsearch/finetuning, avering the network performance on several runs, etc... For more convenience, I set ``device = 'cuda' if torch.cuda.is_available() else 'cpu'``which will be equal to `'cuda'` on Collab. The code is totally adaptable when `device='cpu', i.e. nothing has to be changed."
      ]
    },
    {
      "metadata": {
        "id": "-BPf_ioPbSE3",
        "colab_type": "code",
        "outputId": "b56c148b-1cde-4db0-ae3c-a3b67b270360",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 32
        }
      },
      "cell_type": "code",
      "source": [
        "# Basic packages\n",
        "import os\n",
        "import matplotlib\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import time\n",
        "import pickle\n",
        "from copy import deepcopy\n",
        "from tqdm import tqdm\n",
        "\n",
        "# Packages for DL\n",
        "from torchvision.datasets.cifar import CIFAR10 as _CIFAR10\n",
        "import torchvision.transforms as transforms\n",
        "import torchvision.models as models\n",
        "import torchvision.utils as vutils\n",
        "import torch\n",
        "import torchvision\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "import torch.backends.cudnn as cudnn\n",
        "\n",
        "# Data Augmentation\n",
        "#!pip install Augmentor\n",
        "import Augmentor\n",
        "\n",
        "# Wavelet transform\n",
        "#!pip install kymatio\n",
        "import kymatio\n",
        "\n",
        "# Skorch\n",
        "#!pip install skorch\n",
        "import skorch\n",
        "from skorch import NeuralNetClassifier\n",
        "from skorch.callbacks import EarlyStopping, ProgressBar, LRScheduler, WarmRestartLR\n",
        "\n",
        "\n",
        "# Sklearn\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.model_selection import StratifiedKFold\n",
        "\n",
        "# Remove warnings introduced by the small number of members per class in Kfold\n",
        "import warnings\n",
        "warnings.filterwarnings('ignore')\n",
        "\n",
        "# Set device (on collab, device='cuda', on personal laptop, device='cpu')\n",
        "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
        "print('Working with 1 GPU: '+torch.cuda.get_device_name(0))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Working with 1 GPU: Tesla K80\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "vTVXQZ1nftIH",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Data loader inspired from  https://github.com/pytorch/vision/blob/master/torchvision/datasets/cifar.py. "
      ]
    },
    {
      "metadata": {
        "id": "BumGCvFZaKxp",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "class CIFAR10(_CIFAR10):\n",
        "    def __init__(self, root, start=None, end=None, is_labeled=True, train=True, transform=None, target_transform=None, download=False):\n",
        "        \"\"\"\n",
        "        Args:\n",
        "            start: first index of the set\n",
        "            end: last index of the set\n",
        "            is_labeled: for the X set, which is unlabeled\n",
        "        \"\"\"\n",
        "        super(CIFAR10, self).__init__(root, train, transform, target_transform, download)                           \n",
        "        self.is_labeled = is_labeled\n",
        "        self.data = self.data[start:end]\n",
        "        self.targets = self.targets[start:end]\n",
        "        \n",
        "    def __getitem__(self, index):\n",
        "        img, target = super(CIFAR10, self).__getitem__(index)\n",
        "\n",
        "        if self.is_labeled:\n",
        "            return img, target\n",
        "        return img"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "63D6qabFgHJf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Data vizualization of first 100 samples."
      ]
    },
    {
      "metadata": {
        "id": "h1HxH41kgA_9",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def plot_cifar(X, size):\n",
        "    \"\"\"\n",
        "    Plot the whole training data set on a grid.\n",
        "    \"\"\"\n",
        "    fig, axes1 = plt.subplots(size, size, figsize=(size-3, size-3))\n",
        "    for j in range(size):\n",
        "        for k in range(size):\n",
        "            i = size * j + k\n",
        "            axes1[j][k].set_axis_off()\n",
        "            axes1[j][k].imshow(X[i])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "PgGVKm_a1wp_",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "This is our dataset $\\mathcal{X}_{\\text{train}}$, it will be used until the end of this project. The remaining samples correspond to $\\mathcal{X}$. The testing set $\\mathcal{X}_{\\text{test}}$ corresponds to the whole testing set of CIFAR-10."
      ]
    },
    {
      "metadata": {
        "id": "DTu_f4U-1wqC",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Testing procedure\n",
        "__Question 2:__ Explain why the evaluation of the training procedure is difficult. Propose several solutions."
      ]
    },
    {
      "metadata": {
        "id": "xSm4ih-W1wqE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Many issues arise from small data.\n",
        "- First,  over-fitting becomes much harder to avoid. Worse, over-fit is more likely to occur not only on the training data, but also on the validation set as well. Indeed, the number of observations in training and validation is too few to give meaningful performance estimate. That's why finetuning parameters should be done carefully. In fact, one should not over-use cross-validation data for hyper-parameter optimization, or not at all. Here, 100 training samples of observations results in about 25 samples for validation, which is way too few to guide parameter search optimally. \n",
        "- Then, models are more sensitive to outliers. Here, even few outliers (1 to 10) can significantly alter the model and dramatically impact the performance. Moreover, noise in general becomes a real issue, be it in the target variable or in some images.\n",
        "- At last, one can speak about all form of uncertainty that, in presence of small data, are more significant: measurement errors, missing values, sampling bias). \n",
        ">\n",
        ">\n",
        "> Now, how to tackle those issues (generally speaking).\n",
        ">\n",
        ">- First, use simpler models with lesser degrees of freedom, which contribute to more robust estimates. \n",
        "- Control the data in terms of pre-processing (outliers, missing values, etc...).\n",
        "- Reduce overfitting by ensemble learning (stacking, combining classifiers), regularizing, and augmenting the data set. \n",
        "- Always report marging of error as models on small data have larger confidence intervals in terms of predictions. \n",
        "- Use domain expertise as in weak-supervised learning. \n",
        ">\n",
        ">In the TP, one will take a look at several techniques to tackle the issues of small data. Let's specialize to our data, CIFAR-10. \n",
        ">- *What I do [1/2]*: one may effectively look for using simple deep learning models, i.e. \"simple\" neural network. In fact, VGG or ResNet architecture are not quite simple in terms of numbers of parameters... But we will be very careful on regularization through, for instance, learning rate schedule, dropout or weight decay. To anticipate the next parts,  what will be the systematic approach to best predict the classes is the following:\n",
        "  - Stick to the original architectures. Namely, don't use dropout neither in ResNet nor in VGG. \n",
        "  - Apply 2-fold cross-validation to determine the optimal learning rate, maximal number of epochs and batch size (hoping that the model won't overfit on the validation set). Weight decay is fixed set to $10^{-4}$ or $5\\cdot 10^{-4}$.\n",
        "  - Set the following Learning Rate Schedule: divise the learning rate by $10$ each $10$ epochs (I do this for simplicity as I struggled to use build-in function on Pytorch). \n",
        "  - Use the fine tuned hyperparameters and train the network with it multiple times in order to build confidence intervals. **Sometimes I will choose different parameters from the optimized ones, in particular for ResNet18**\n",
        "  - **The accuracies of reference on the full data set are available in the conclusion.**\n",
        "- *What I do [2/2]*: a bit of data augmentation is done as a basis for training. Namely, `randomCrop(32, padding=4)` and `RandomHorizontalFlip()`. ** Question 8, I compare more complex data augmentation strategies and no data augmentation at all, the latter case is thus not done for question 3 and 4 (misunderstanding of the wording)**.\n",
        "- *What I don't do*: I won't proceed in data pre-processing as CIFAR-10 data set is pretty clean already. I won't proceed neither in stacking models (for simplicity) and in order to stick to the assumption of a limited computational budget.\n",
        ">\n",
        "> Finally, one may talk about transfer, semi-supervised and weakly-supervised learning that are very useful to handle small data. All techniques seen in the lecture will be applied in the next sections and are considered as bringing state-of-the-art performance on CIFAR-10 for very few training samples. \n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "gcSrxGFW1wqH",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Raw approach: the baseline"
      ]
    },
    {
      "metadata": {
        "id": "D9nbW7Fu1wqJ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "In this section, the goal is to train a CNN on $\\mathcal{X}_{\\text{train}}$ and compare its performances with reported number from the litterature. You will have to re-use and/or design a standard classification pipeline. You should optimize your pipeline to obtain the best performances (image size, data augmentation by flip, ...).\n",
        "\n",
        "The key ingredients for training a CNN are the batch size, as well as the learning rate schedule, i.e. how to decrease the learning rate as a function of the number of epochs. A possible schedule is to start the learning rate at 0.1 and decreasing it every 30 epochs by 10. In case of divergence, reduce the laerning rate. A potential batch size could be 10, yet this can be cross-validated.\n",
        "\n",
        "You can get some baselines accuracies in this paper: http://openaccess.thecvf.com/content_cvpr_2018/papers/Keshari_Learning_Structure_and_CVPR_2018_paper.pdf. Obviously, it is a different context, as those researchers had access to GPUs."
      ]
    },
    {
      "metadata": {
        "id": "-k-6fF4Q1wqK",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## ResNet architectures"
      ]
    },
    {
      "metadata": {
        "id": "Hb3oavQd1wqM",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "__Question 3:__ Write a classification pipeline for $\\mathcal{X}_{\\text{train}}$, train from scratch and evaluate a *ResNet-18* architecture specific to CIFAR10 (details about the ImageNet model can be found here: https://arxiv.org/abs/1409.1556 ). If possible, please report the accuracy obtained on the whole dataset, as well as the reference paper/GitHub link you might have used.\n",
        "\n",
        "*Hint:* You can re-use the following code: https://github.com/kuangliu/pytorch-cifar. During a training of 10 epochs, a batch size of 10 and a learning rate of 0.01, one obtains 40% accuracy on $\\mathcal{X}_{\\text{train}}$ (~2 minutes) and 20% accuracy on $\\mathcal{X}_{\\text{test}}$ (~5 minutes)."
      ]
    },
    {
      "metadata": {
        "id": "icuW7uK126ZU",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Here, one stick to the ResNet18 classical architecture from He et al. \"Deep Residual Learning for Image Recognition\" (2015). I re-used the suggested piece of code available on GitHub. In order to apply hyperparameters tuning, I use ``skorch`` package, which needs to be installed beforehand. It is really helpful for doing gridsearching in a compact way. \n",
        ">\n",
        "> Kuang Liu proposes a bunch of models applied to CIFAR-10 data set. "
      ]
    },
    {
      "metadata": {
        "id": "jOz98dwVrQ06",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "class BasicBlock(nn.Module):\n",
        "    \"\"\"\n",
        "    Define a block from ResNet architecture.\n",
        "    Source: https://github.com/kuangliu/pytorch-cifar\n",
        "    \"\"\"\n",
        "    expansion = 1\n",
        "    def __init__(self, in_planes, planes, stride=1):\n",
        "        super(BasicBlock, self).__init__()\n",
        "        self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)\n",
        "        self.bn1 = nn.BatchNorm2d(planes)\n",
        "        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1, padding=1, bias=False)\n",
        "        self.bn2 = nn.BatchNorm2d(planes)\n",
        "        self.shortcut = nn.Sequential()\n",
        "        if stride != 1 or in_planes != self.expansion*planes:\n",
        "            self.shortcut = nn.Sequential(\n",
        "                nn.Conv2d(in_planes, self.expansion*planes, kernel_size=1, stride=stride, bias=False),\n",
        "                nn.BatchNorm2d(self.expansion*planes))\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = F.relu(self.bn1(self.conv1(x)))\n",
        "        out = self.bn2(self.conv2(out))\n",
        "        out += self.shortcut(x)\n",
        "        out = F.relu(out)\n",
        "        return out\n",
        "      \n",
        "class ResNet(nn.Module):\n",
        "    \"\"\"\n",
        "    ResNet network from  He et al. \"Deep Residual Learning for Image Recognition\" (2015).\n",
        "    Source: https://github.com/kuangliu/pytorch-cifar\n",
        "    \"\"\"\n",
        "    def __init__(self, block=BasicBlock, num_blocks=[2, 2, 2, 2], num_classes=10):\n",
        "        super(ResNet, self).__init__()\n",
        "        self.in_planes = 64\n",
        "        self.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1, bias=False)\n",
        "        self.bn1 = nn.BatchNorm2d(64)\n",
        "        self.layer1 = self._make_layer(block, 64, num_blocks[0], stride=1)\n",
        "        self.layer2 = self._make_layer(block, 128, num_blocks[1], stride=2)\n",
        "        self.layer3 = self._make_layer(block, 256, num_blocks[2], stride=2)\n",
        "        self.layer4 = self._make_layer(block, 512, num_blocks[3], stride=2)\n",
        "        self.linear = nn.Linear(512*block.expansion, num_classes)\n",
        "\n",
        "    def _make_layer(self, block, planes, num_blocks, stride):\n",
        "        strides = [stride] + [1]*(num_blocks-1)\n",
        "        layers = []\n",
        "        for stride in strides:\n",
        "            layers.append(block(self.in_planes, planes, stride))\n",
        "            self.in_planes = planes * block.expansion\n",
        "        return nn.Sequential(*layers)\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = F.relu(self.bn1(self.conv1(x)))\n",
        "        out = self.layer1(out)\n",
        "        out = self.layer2(out)\n",
        "        out = self.layer3(out)\n",
        "        out = self.layer4(out)\n",
        "        out = F.avg_pool2d(out, 4)\n",
        "        out = out.view(out.size(0), -1)\n",
        "        out = self.linear(out)\n",
        "        return out"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "rBlLo0rugi0m",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Now I apply transformation on the CIFAR-10 data set. Those transformations are directly inspired from Kuang Liu. **In fact, I struggled pretty hard to retrieve the 20% accuracy from the wording.** This is due to the fact that I didn't apply those transformations (except normalization). My cross-validations and training were giving at most 18% accuracy on test set, which was not expected. I don't know if those transformations should have been applied. That's why I choose to test the models with no data augmentation at all in question 8."
      ]
    },
    {
      "metadata": {
        "id": "Ths9eG2qgbeF",
        "colab_type": "code",
        "outputId": "fb1c47af-017d-48f7-f539-473377c95a9f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49
        }
      },
      "cell_type": "code",
      "source": [
        "# Transform on train\n",
        "transform_train =  transforms.Compose([transforms.RandomCrop(32, padding=4), \n",
        "                                       transforms.RandomHorizontalFlip(), \n",
        "                                       transforms.ToTensor(),\n",
        "                                       transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "# Transform on test\n",
        "transform_test =  transforms.Compose([transforms.ToTensor(),\n",
        "                                      transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "\n",
        "# Load data set \n",
        "dataset_train = CIFAR10(path, end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(path, train=False, transform=transform_test, download=True)\n",
        "\n",
        "# Extract X, y\n",
        "X_train, y_train, X_test, y_test = dataset_train.data, dataset_train.targets, dataset_test.data, dataset_test.targets"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\r0it [00:00, ?it/s]"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to /content/cifar-10-python.tar.gz\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "100%|█████████▉| 170418176/170498071 [00:43<00:00, 1864659.91it/s]"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "mJ_XM4JFse77",
        "colab_type": "code",
        "outputId": "c242f4f2-b6c1-4487-db6c-b72571910ade",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        }
      },
      "cell_type": "code",
      "source": [
        "# Plot training samples\n",
        "plot_cifar(X_train, 10)"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAakAAAGfCAYAAADlDy3rAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 504x504 with 100 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "HTDzmGaA6Q_i",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Let's try cross-validation on some parameters. One may only look at the maximal number of epochs, learning rate and batch-size. Indeed, He et al. (2015) do not use any dropout. They use SGD solver. "
      ]
    },
    {
      "metadata": {
        "id": "beW-tpuwLfZc",
        "colab_type": "code",
        "outputId": "b81ff010-20dd-4ac6-b129-4312383b2a3c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 103
        }
      },
      "cell_type": "code",
      "source": [
        "# Define the network. Verbose=0 in order to avoid too many callbacks.\n",
        "# Learning rate are automatically scheduled thanks to LRScheduler.\n",
        "ResNet18 = NeuralNetClassifier(\n",
        "    ResNet().double(),\n",
        "    iterator_train__shuffle=True,\n",
        "    criterion=nn.CrossEntropyLoss,\n",
        "    optimizer=torch.optim.SGD,\n",
        "    callbacks=[EarlyStopping(monitor='valid_loss', patience=5, threshold=1e-4)],\n",
        "    device=device,\n",
        "    verbose=0)\n",
        "\n",
        "# Allowing many combinations is doable as, in Collab, GPU is available.\n",
        "params = {'lr': [0.005, 0.01, 0.02, 0.05],\n",
        "          'batch_size': [5, 10, 15, 20],\n",
        "          'max_epochs': [5, 10, 20, 30]}\n",
        "\n",
        "# I do not use X_train and y_train here in order to avoid errors. \n",
        "### First with the Basic Block\n",
        "# CV is done over 2 stratified folds.\n",
        "gs_bb = GridSearchCV(ResNet18, params, refit=True, cv=StratifiedKFold(n_splits=2), scoring='accuracy', verbose=0)\n",
        "print('...Gridsearch with ResNet18...')\n",
        "t = time.time()\n",
        "gs_bb.fit(np.array([X.reshape(3, 32, 32).astype('double') for X in X_train]), np.array(y_train).astype('int64'))\n",
        "print('Cross-validation finished. Time elapsed: {:5.2f} sec.'.format(time.time()-t))\n",
        "print('Best accuracy on val set: {:0.4f}, best parameters: '.format(gs_bb.best_score_), gs_bb.best_params_)\n",
        "pickle.dump(gs_bb, open(os.path.join(path, 'cv_resnet18.pk'), 'wb'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Gridsearch with ResNet18...\n",
            "Cross-validation finished. Time elapsed: 783.64 sec.\n",
            "Best accuracy on val set: 0.2200, best parameters:  {'batch_size': 5, 'lr': 0.005, 'max_epochs': 10}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "QZ9EbwdqMhGf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Now that hyperparameters has been tuned, let's see how our network behaves on the test set."
      ]
    },
    {
      "metadata": {
        "id": "5cAOFv3iNuT2",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "lr, momentum, weight_decay = 0.005, 0.9, 5e-4\n",
        "max_epochs = 20 # twice the optimal one /!\\\n",
        "batch_size = 5"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "eXNqclRtaDj7",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def init_resnet():\n",
        "  ResNet18 = ResNet().to(device)\n",
        "  ResNet18 = ResNet18.to(device)\n",
        "  if device == 'cuda':\n",
        "      ResNet18 = torch.nn.DataParallel(ResNet18)\n",
        "      cudnn.benchmark = True\n",
        "  return ResNet18\n",
        "\n",
        "ResNet18 = init_resnet()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "4JawFSezL1Er",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Define loss and optimizer\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = optim.SGD(ResNet18.parameters(), lr=lr, momentum=0.9, weight_decay=5e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Uiz4J4_4P9RU",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Here I code a function to train and/or test several times a given network. Those functions are very useful for the whole homework."
      ]
    },
    {
      "metadata": {
        "id": "WUTTbbqWqJ-m",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Training\n",
        "def train(net, epoch, optimizer, criterion, train_loader, verbose, scattering):\n",
        "    net.train()\n",
        "    train_loss = 0; correct = 0; total = 0\n",
        "    for batch_idx, (inputs, targets) in enumerate(train_loader):\n",
        "        inputs, targets = inputs.to(device), targets.to(device)\n",
        "        optimizer.zero_grad()\n",
        "        if scattering is None:\n",
        "          outputs = net(inputs)\n",
        "        else:\n",
        "          outputs = net(scattering(inputs))\n",
        "        loss = criterion(outputs, targets)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        train_loss += loss.item()\n",
        "        _, predicted = outputs.max(1)\n",
        "        total += targets.size(0)\n",
        "        correct += predicted.eq(targets).sum().item()\n",
        "    loss, acc =  train_loss / (batch_idx+1), 100 * correct / total\n",
        "    if verbose==1:\n",
        "      print('\\n...Epoch %d...\\n...Train_loss: %.3f | Train_acc: %.3f%% |... ' % (epoch, loss, acc))  \n",
        "    return loss, acc\n",
        "\n",
        "# Testing\n",
        "def test(net, epoch, criterion, test_loader, verbose, scattering):\n",
        "    net.eval()\n",
        "    test_loss = 0; correct = 0; total = 0\n",
        "    with torch.no_grad():\n",
        "        for batch_idx, (inputs, targets) in enumerate(test_loader):\n",
        "            inputs, targets = inputs.to(device), targets.to(device)\n",
        "            if scattering is None:\n",
        "              outputs = net(inputs)\n",
        "            else:\n",
        "              outputs = net(scattering(inputs))\n",
        "            loss = criterion(outputs, targets)\n",
        "            test_loss += loss.item()\n",
        "            _, predicted = outputs.max(1)\n",
        "            total += targets.size(0)\n",
        "            correct += predicted.eq(targets).sum().item()\n",
        "    loss, acc =  test_loss / (batch_idx+1), 100 * correct / total\n",
        "    if verbose==1:\n",
        "      print('...Test_loss: %.3f | Test_acc: %.3f%% |... ' % (loss, acc))  \n",
        "    return loss, acc\n",
        "\n",
        "# Both\n",
        "def train_test(init, name, max_epochs, criterion, train_loader, test_loader, N=1, verbose=0, LRreduce=0, scattering=None):\n",
        "    train_loss, test_loss, train_acc, test_acc = np.zeros((N, max_epochs)), np.zeros((N, max_epochs)), np.zeros((N, max_epochs)), np.zeros((N, max_epochs))\n",
        "    for n in range(N):\n",
        "      t = time.time()\n",
        "      print('...Training n°%d...'% (n+1))\n",
        "      lr_ = deepcopy(lr)\n",
        "      net = init()\n",
        "      optimizer = optim.SGD(net.parameters(), lr=lr_, momentum=0.9, weight_decay=5e-4)\n",
        "      for epoch in range(1, max_epochs+1):\n",
        "          if epoch%10==0 and LRreduce == 1:\n",
        "              lr_*=0.1\n",
        "              optimizer = torch.optim.SGD(net.parameters(), lr=lr_, momentum=0.9,\n",
        "                                          weight_decay=5e-4)\n",
        "          l_tr, a_tr = train(net, epoch, optimizer, criterion, train_loader, verbose, scattering)\n",
        "          l_te, a_te = test(net, epoch, criterion, test_loader, verbose, scattering)\n",
        "          train_loss[n, epoch-1]=l_tr\n",
        "          test_loss[n, epoch-1]=l_te\n",
        "          train_acc[n, epoch-1]=a_tr\n",
        "          test_acc[n, epoch-1]=a_te\n",
        "          pickle.dump([train_loss, test_loss, train_acc, test_acc], open(os.path.join(path, '/Simulations/N'+str(N)+'LRr'+str(LRreduce)+'_train_test_'+name+'.pk'), 'wb')) \n",
        "      i0 = np.argmax(test_acc[n])\n",
        "      print('Time elapsed: %4.2f sec. | Best test_acc: %.3f%% | Best train_acc: %.3f%%' % (time.time()-t, test_acc[n, i0], train_acc[n, i0]))\n",
        "    return train_loss, test_loss, train_acc, test_acc"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "XHkYZCMZqMb1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 345
        },
        "outputId": "bd6a4704-0b49-479e-95e7-733263aaa5ea"
      },
      "cell_type": "code",
      "source": [
        "# Let's see what it has\n",
        "train_loss_r18, test_loss_r18, train_acc_r18, test_acc_r18 = train_test(init_resnet, name='resnet18', max_epochs=max_epochs,\n",
        "                                                                        criterion=criterion,\n",
        "                                                                        train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                        N=10, verbose=0)"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Training n°1...\n",
            "Time elapsed: 403.30 sec. | Best test_acc: 24.490% | Best train_acc: 40.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 403.00 sec. | Best test_acc: 23.950% | Best train_acc: 37.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 402.90 sec. | Best test_acc: 23.440% | Best train_acc: 30.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 402.29 sec. | Best test_acc: 23.830% | Best train_acc: 30.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 403.46 sec. | Best test_acc: 25.370% | Best train_acc: 31.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 400.96 sec. | Best test_acc: 24.120% | Best train_acc: 37.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 403.07 sec. | Best test_acc: 23.350% | Best train_acc: 44.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 402.29 sec. | Best test_acc: 23.520% | Best train_acc: 32.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 402.36 sec. | Best test_acc: 23.310% | Best train_acc: 31.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 404.01 sec. | Best test_acc: 25.300% | Best train_acc: 35.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "kvh25pOT1wqT",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## VGG-like architectures"
      ]
    },
    {
      "metadata": {
        "id": "2dIkWDqe1wqV",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "__Question 4:__ Same question as before, but with a *VGG*. Which model do you recommend?"
      ]
    },
    {
      "metadata": {
        "id": "1eLIzAjV1wqW",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "cfg = {\n",
        "    'VGG11': [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],\n",
        "    'VGG13': [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],\n",
        "    'VGG16': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M'],\n",
        "    'VGG19': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512, 'M'],\n",
        "}\n",
        "\n",
        "\n",
        "class VGG(nn.Module):\n",
        "    \"\"\"\n",
        "    VGG architecture.\n",
        "    Source: https://github.com/kuangliu/pytorch-cifar\n",
        "    \"\"\"\n",
        "    def __init__(self, vgg_name):\n",
        "        super(VGG, self).__init__()\n",
        "        self.features = self._make_layers(cfg[vgg_name])\n",
        "        self.classifier = nn.Linear(512, 10)\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = self.features(x)\n",
        "        out = out.view(out.size(0), -1)\n",
        "        out = self.classifier(out)\n",
        "        return out\n",
        "\n",
        "    def _make_layers(self, cfg):\n",
        "        layers = []\n",
        "        in_channels = 3\n",
        "        for x in cfg:\n",
        "            if x == 'M':\n",
        "                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]\n",
        "            else:\n",
        "                layers += [nn.Conv2d(in_channels, x, kernel_size=3, padding=1),\n",
        "                           nn.BatchNorm2d(x),\n",
        "                           nn.ReLU(inplace=True)]\n",
        "                in_channels = x\n",
        "        layers += [nn.AvgPool2d(kernel_size=1, stride=1)]\n",
        "        return nn.Sequential(*layers)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "NLc7HbTsZWEu",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "VGG16 = NeuralNetClassifier(\n",
        "    VGG('VGG16').double(),\n",
        "    iterator_train__shuffle=True,\n",
        "    criterion=nn.CrossEntropyLoss,\n",
        "    optimizer=torch.optim.SGD,\n",
        "    optimizer__momentum=0.95,\n",
        "    optimizer__weight_decay=1e-4,\n",
        "    device=device,\n",
        "    callbacks=[EarlyStopping(monitor='valid_loss', patience=5, threshold=1e-4)],\n",
        "    verbose=0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Ck-HyjgPcKYM",
        "colab_type": "code",
        "outputId": "ac553483-b9b4-4dcf-c883-149431dca53d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 65
        }
      },
      "cell_type": "code",
      "source": [
        "params = {'lr': [0.005, 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1],\n",
        "          'batch_size': [5, 10, 15, 20, 25, 30],\n",
        "          'max_epochs': [10, 20, 30, 50, 80, 100]}\n",
        "\n",
        "# I do not use X_train and y_train here in order to avoid errors. \n",
        "### First with the Basic Block\n",
        "# CV is done over 2 stratified folds.\n",
        "gs_vgg16 = GridSearchCV(VGG16, params, refit=False, cv=StratifiedKFold(n_splits=2), scoring='accuracy', verbose=0)\n",
        "print('...Gridsearch for VGG16...')\n",
        "t = time.time()\n",
        "gs_vgg16.fit(np.array([X.reshape(3, 32, 32).astype('double') for X in X_train]), np.array(y_train).astype('int64'))\n",
        "print('Cross-validation finished. Time elapsed: {:5.2f} sec.'.format(time.time()-t))\n",
        "print('Best accuracy on val set: {:0.4f}, best parameters: '.format(gs_vgg16.best_score_), gs_vgg16.best_params_)\n",
        "pickle.dump(gs_vgg16, open(os.path.join(path, 'cv_vgg16.pk'), 'wb'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Gridsearch for VGG16...\n",
            "Cross-validation finished. Time elapsed: 1899.75 sec.\n",
            "Best accuracy on val set: 0.2000, best parameters:  {'batch_size': 5, 'lr': 0.05, 'max_epochs': 20}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "M8TYTT3xcuhg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "lr, momentum, weight_decay = 0.05, 0.9, 1e-4\n",
        "max_epochs = 20\n",
        "batch_size = 5"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "H6T1HROIZSHh",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def init_vgg():\n",
        "  VGG16 = VGG('VGG16').to(device)\n",
        "  if device == 'cuda':\n",
        "      net = torch.nn.DataParallel(VGG16)\n",
        "      cudnn.benchmark = True\n",
        "  return VGG16\n",
        "\n",
        "VGG16 = init_vgg()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "_KrTRgAlcyns",
        "colab_type": "code",
        "outputId": "47487bb1-b8c2-42d6-edbc-70645ede2e3b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 345
        }
      },
      "cell_type": "code",
      "source": [
        "# Define the loss function\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "\n",
        "# And the optimizer\n",
        "optimizer = optim.SGD(VGG16.parameters(), lr=lr, momentum=0.9, weight_decay=1e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "# Let's see what it has\n",
        "train_loss_vgg16, test_loss_vgg16, train_acc_vgg16, test_acc_vgg16 = train_test(init_resnet, name='vgg16', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Training n°1...\n",
            "Time elapsed: 433.49 sec. | Best test_acc: 19.450% | Best train_acc: 24.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 432.06 sec. | Best test_acc: 18.790% | Best train_acc: 21.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 439.13 sec. | Best test_acc: 18.990% | Best train_acc: 23.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 439.03 sec. | Best test_acc: 14.260% | Best train_acc: 16.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 438.80 sec. | Best test_acc: 20.980% | Best train_acc: 25.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 437.91 sec. | Best test_acc: 17.120% | Best train_acc: 20.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 439.30 sec. | Best test_acc: 17.300% | Best train_acc: 22.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 447.42 sec. | Best test_acc: 20.120% | Best train_acc: 19.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 443.33 sec. | Best test_acc: 20.570% | Best train_acc: 27.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 432.79 sec. | Best test_acc: 17.070% | Best train_acc: 16.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "xFmC9JnOd0RS",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Comparison : the cell below compare the accuracies on train and test set averaged on 10 independent simulations with same hyperparameters. From those numbers, one can strongly recommand using ResNet18. It has 25% less trainable parameters, is faster to train and produces higher train and test accuracies with comparable or better standard deviations. \n",
        "\n",
        "| Model | Number of  epochs  | Train accuracy | Test accuracy |\n",
        "|------|------|------|------|\n",
        "|   ResNet18  | 20 | 35.500 ± 3.828% | 24.672 ± 0.777% |\n",
        "|   VGG16  | 20 | 21.300 ± 3.466% | 18.465 ± 1.940% |\n"
      ]
    },
    {
      "metadata": {
        "id": "bDCdnz-wARRA",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> To run the following cell, you should set a path that like `./simulations/` as all files of type \"train_test\" are available in this folder."
      ]
    },
    {
      "metadata": {
        "id": "TGR-VeK5RTWF",
        "colab_type": "code",
        "outputId": "8a7ee41f-76e6-4e04-b993-37b74e1ea16e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49
        }
      },
      "cell_type": "code",
      "source": [
        "# Load results from multiple trainings \n",
        "train_loss_r18, test_loss_r18, train_acc_r18, test_acc_r18 = pickle.load(open(os.path.join(path, '/Simulations/N10LRr0_train_test_resnet18.pk'), 'rb'))\n",
        "train_loss_vgg16, test_loss_vgg16, train_acc_vgg16, test_acc_vgg16 = pickle.load(open(os.path.join(path, '/Simulations/N10LRr0_train_test_vgg16.pk'), 'rb'))\n",
        "\n",
        "# Define a small function to return mean and std\n",
        "def stat(train_acc, test_acc):\n",
        "  imax = np.argmax(test_acc, axis=1)\n",
        "  train_acc_, test_acc_ = train_acc[np.arange(10), imax], test_acc[np.arange(10), imax]\n",
        "  return np.mean(train_acc_), np.std(train_acc_), np.mean(test_acc_), np.std(test_acc_)\n",
        "\n",
        "# Compute corresponding stats for ResNet18 and VGG16\n",
        "m_tr_r18, std_tr_r18, m_te_r18, std_te_r18 = stat(train_acc_r18, test_acc_r18)\n",
        "m_tr_v16, std_tr_v16, m_te_v16, std_te_v16 = stat(train_acc_vgg16, test_acc_vgg16)\n",
        "\n",
        "# Print comparison\n",
        "tpvgg16 = sum(p.numel() for p in VGG16.parameters() if p.requires_grad)\n",
        "tpresnet18 = sum(p.numel() for p in ResNet18.parameters() if p.requires_grad)\n",
        "print('ResNet18: %d training parameters | Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (tpresnet18, m_tr_r18, std_tr_r18, m_te_r18, std_te_r18))\n",
        "print('VGG16:    %d training parameters | Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (tpvgg16, m_tr_v16, std_tr_v16, m_te_v16, std_te_v16))"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ResNet18: 11173962 training parameters | Train accuracy: 34.700 ± 4.518% | Test accuracy: 24.068 ± 0.725%\n",
            "VGG16:    14728266 training parameters | Train accuracy: 21.300 ± 3.466% | Test accuracy: 18.465 ± 1.940%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "QWrhIaod1wqa",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Transfer learning"
      ]
    },
    {
      "metadata": {
        "id": "yANzK3581wqc",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "We propose to use pre-trained models on a classification and generative task, in order to improve the results of our setting."
      ]
    },
    {
      "metadata": {
        "id": "RhHYKXV71wqd",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## ImageNet features"
      ]
    },
    {
      "metadata": {
        "id": "YLOyIoT21wqf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Now, we will use some pre-trained models on ImageNet and see how well they compare on CIFAR. A list is available on: https://pytorch.org/docs/stable/torchvision/models.html.\n",
        "\n",
        "__Question 5:__ Pick a model from the list above, adapt it to CIFAR and retrain its final layer (or a block of layers, depending on the resources to which you have access to). Report its accuracy."
      ]
    },
    {
      "metadata": {
        "id": "e_OIXWp9iTIp",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> I choose VGG16."
      ]
    },
    {
      "metadata": {
        "id": "jOd4S-9i1wqg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def init_vgg_pretrained():\n",
        "  VGG16_pretrained = models.vgg16(pretrained=True)\n",
        "\n",
        "  # Freeze model weights\n",
        "  for param in VGG16_pretrained.parameters():\n",
        "      param.requires_grad = False\n",
        "\n",
        "  # Retrain last layer only\n",
        "  VGG16_pretrained.classifier[6] = nn.Sequential(\n",
        "                        nn.Linear(4096, 256), \n",
        "                        nn.ReLU(), \n",
        "                        nn.Dropout(0.4),\n",
        "                        nn.Linear(256, 10),                   \n",
        "                        nn.LogSoftmax(dim=1))\n",
        "  # Move to gpu\n",
        "  VGG16_pretrained = VGG16_pretrained.to('cuda')\n",
        "  # Distribute across 2 gpus\n",
        "  VGG16_pretrained = nn.DataParallel(VGG16_pretrained)\n",
        "  return VGG16_pretrained\n",
        "\n",
        "VGG16_pretrained = init_vgg_pretrained()\n",
        "\n",
        "# Find total parameters and trainable parameters\n",
        "total_params = sum(p.numel() for p in VGG16_pretrained.parameters())\n",
        "print(f'{total_params:,} total parameters.')\n",
        "total_trainable_params = sum(p.numel() for p in VGG16_pretrained.parameters() if p.requires_grad)\n",
        "print(f'{total_trainable_params:,} training parameters.')\n",
        "\n",
        "\n",
        "# Use skorch\n",
        "\n",
        "VGG16_pretrained_ = NeuralNetClassifier(\n",
        "    VGG16_pretrained.double(),\n",
        "    iterator_train__shuffle=True,\n",
        "    criterion=nn.CrossEntropyLoss,\n",
        "    optimizer=torch.optim.SGD,\n",
        "    optimizer__momentum=0.95,\n",
        "    optimizer__weight_decay=1e-4,\n",
        "    device=device,\n",
        "    callbacks=[EarlyStopping(monitor='valid_loss', patience=5, threshold=1e-4)],\n",
        "    verbose=0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "KcEQ5jBYE5nz",
        "colab_type": "code",
        "outputId": "5b69a1ae-24eb-46f4-bf7e-ffff4d797154",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 65
        }
      },
      "cell_type": "code",
      "source": [
        "# Allowing many combinations is doable as, in Collab, GPU is available.\n",
        "params = {'lr': [0.0001, 0.0005, 0.001, 0.005, 0.01, 0.02, 0.05],\n",
        "          'batch_size': [15, 20, 25, 30, 35, 40, 50],\n",
        "          'max_epochs': [10, 20, 30, 40, 50, 70, 80, 100]}\n",
        "\n",
        "\n",
        "# I do not use X_train and y_train here in order to avoid errors. \n",
        "### First with the Basic Block\n",
        "# CV is done over 2 stratified folds.\n",
        "gs_vgg16_ptrained = GridSearchCV(VGG16_pretrained_, params, refit=False, cv=StratifiedKFold(n_splits=2), scoring='accuracy', verbose=0)\n",
        "print('...Gridsearch for pretrained VGG16 (only last layer learned)...')\n",
        "t = time.time()\n",
        "gs_vgg16_ptrained.fit(np.array([X.reshape(3, 32, 32).astype('double') for X in X_train]), np.array(y_train).astype('int64'))\n",
        "print('Cross-validation finished. Time elapsed: {:5.2f} sec.'.format(time.time()-t))\n",
        "print('Best accuracy on val set: {:0.4f}, best parameters: '.format(gs_vgg16_ptrained.best_score_), gs_vgg16_ptrained.best_params_)\n",
        "pickle.dump(gs_vgg16_ptrained, open(os.path.join(path,'cv_vgg16_pretrained.pk'), 'wb'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Gridsearch for pretrained VGG16 (only last layer learned)...\n",
            "Cross-validation finished. Time elapsed: 2007.26 sec.\n",
            "Best accuracy on val set: 0.2500, best parameters:  {'batch_size': 20, 'lr': 0.0001, 'max_epochs': 40}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "rqAZgBMBW6-e",
        "colab_type": "code",
        "outputId": "56662e8f-9ee5-4fc8-e4cf-96e2bf395e4c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 345
        }
      },
      "cell_type": "code",
      "source": [
        "lr, momentum, weight_decay = 1e-3, 0.9, 1e-4\n",
        "max_epochs = 40\n",
        "batch_size = 20\n",
        "\n",
        "# Define the loss function\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "\n",
        "# And the optimizer\n",
        "optimizer = optim.SGD(VGG16_pretrained.parameters(), lr=lr, momentum=0.9, weight_decay=1e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "# Let's see what it has\n",
        "train_loss_vgg16p, test_loss_vgg16p, train_acc_vgg16p, test_acc_vgg16p = train_test(init_vgg_pretrained, name='vgg16pretrained', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Training n°1...\n",
            "Time elapsed: 437.82 sec. | Best test_acc: 37.960% | Best train_acc: 65.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 433.83 sec. | Best test_acc: 37.560% | Best train_acc: 71.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 435.52 sec. | Best test_acc: 37.790% | Best train_acc: 62.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 435.49 sec. | Best test_acc: 38.830% | Best train_acc: 75.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 434.41 sec. | Best test_acc: 36.780% | Best train_acc: 73.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 433.69 sec. | Best test_acc: 37.780% | Best train_acc: 69.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 433.63 sec. | Best test_acc: 39.030% | Best train_acc: 65.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 433.92 sec. | Best test_acc: 37.950% | Best train_acc: 69.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 432.96 sec. | Best test_acc: 37.780% | Best train_acc: 80.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 433.08 sec. | Best test_acc: 37.630% | Best train_acc: 65.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "HRjrjDwRYSO3",
        "colab_type": "code",
        "outputId": "e8353e4b-47d8-441d-b814-be1ad93e3fad",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 32
        }
      },
      "cell_type": "code",
      "source": [
        "# Load results from multiple trainings \n",
        "train_loss_vgg16p, test_loss_vgg16p, train_acc_vgg16p, test_acc_vgg16p = pickle.load(open(os.path.join(path, '/Simulations/N10LRr0_train_test_vgg16pretrained.pk'), 'rb'))\n",
        "\n",
        "# Compute corresponding stats for ResNet18 and VGG16\n",
        "m_tr_v16p, std_tr_v16p, m_te_v16p, std_te_v16p = stat(train_acc_vgg16p, test_acc_vgg16p)\n",
        "\n",
        "# Print comparison\n",
        "tpvgg16p = sum(p.numel() for p in VGG16_pretrained.parameters() if p.requires_grad)\n",
        "print('Pretrained VGG16: %d training parameters | Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (tpvgg16p, m_tr_v16p, std_tr_v16p, m_te_v16p, std_te_v16p))"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Pretrained VGG16: 1051402 training parameters | Train accuracy: 69.400 ± 5.219% | Test accuracy: 37.909 ± 0.603%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "9Oqg9upHYI7j",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Here are the accuracies: \n",
        "\n",
        "| Model | Number of  epochs  | Train accuracy | Test accuracy |\n",
        "|------|------|------|------|\n",
        "|   ResNet18  | 20 | 35.500 ± 3.828% | 24.672 ± 0.777% |\n",
        "|   VGG16  | 20 | 21.300 ± 3.466% | 18.465 ± 1.940% |\n",
        "|   Pretrained VGG16  | 40 | 69.400 ± 5.219% | 37.909 ± 0.603% |\n",
        "\n",
        "Impressive !\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "ts6V0BgJ1wqi",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## DCGan features"
      ]
    },
    {
      "metadata": {
        "id": "C2AZiQTv1wqj",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "GANs correspond to an unsupervised technique for generating images. In https://arxiv.org/pdf/1511.06434.pdf, Sec. 5.1 shows that the representation obtained from the Discriminator has some nice generalization properties on CIFAR10."
      ]
    },
    {
      "metadata": {
        "id": "CcbrmJ-n1wqk",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "__Question 6:__  Using for instance a pretrained model from https://github.com/soumith/dcgan.torch combined with https://github.com/pytorch/examples/tree/master/dcgan, propose a model to train on $\\mathcal{X}_{\\text{train}}$. Train it and report its accuracy.\n",
        "\n",
        "*Hint:* You can use the library: https://github.com/bshillingford/python-torchfile to load the weights of a model from torch(Lua) to pytorch(python)."
      ]
    },
    {
      "metadata": {
        "id": "qQyeAhWDibkh",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> For this part, I used the following weights: https://github.com/csinva/pytorch_gan_pretrained/tree/master/cifar10_dcgan/weights. There are obtained by training DCGAN on CIFAR-10 for 200 epochs. The architectures of the discriminator and generator are exactly the same as in https://github.com/pytorch/examples/tree/master/dcgan. I did that as I couldn't find a way to proper use the weights you suggested. "
      ]
    },
    {
      "metadata": {
        "id": "FxCpNK1PdmmM",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "class Generator(nn.Module):\n",
        "    def __init__(self, ngpu, nc=3, nz=100, ngf=64):\n",
        "        super(Generator, self).__init__()\n",
        "        self.ngpu = ngpu\n",
        "        self.main = nn.Sequential(\n",
        "            # input is Z, going into a convolution\n",
        "            nn.ConvTranspose2d(nz, ngf * 8, 4, 1, 0, bias=False),\n",
        "            nn.BatchNorm2d(ngf * 8),\n",
        "            nn.ReLU(True),\n",
        "            # state size. (ngf*8) x 4 x 4\n",
        "            nn.ConvTranspose2d(ngf * 8, ngf * 4, 4, 2, 1, bias=False),\n",
        "            nn.BatchNorm2d(ngf * 4),\n",
        "            nn.ReLU(True),\n",
        "            # state size. (ngf*4) x 8 x 8\n",
        "            nn.ConvTranspose2d(ngf * 4, ngf * 2, 4, 2, 1, bias=False),\n",
        "            nn.BatchNorm2d(ngf * 2),\n",
        "            nn.ReLU(True),\n",
        "            # state size. (ngf*2) x 16 x 16\n",
        "            nn.ConvTranspose2d(ngf * 2, ngf, 4, 2, 1, bias=False),\n",
        "            nn.BatchNorm2d(ngf),\n",
        "            nn.ReLU(True),\n",
        "            nn.ConvTranspose2d(ngf, nc, kernel_size=1, stride=1, padding=0, bias=False),\n",
        "            nn.Tanh())\n",
        "\n",
        "    def forward(self, input):\n",
        "        if input.is_cuda and self.ngpu > 1:\n",
        "            output = nn.parallel.data_parallel(self.main, input, range(self.ngpu))\n",
        "        else:\n",
        "            output = self.main(input)\n",
        "        return output\n",
        "\n",
        "\n",
        "class Discriminator(nn.Module):\n",
        "    def __init__(self, ngpu, nc=3, ndf=64):\n",
        "        super(Discriminator, self).__init__()\n",
        "        self.ngpu = ngpu\n",
        "        self.main = nn.Sequential(\n",
        "            # input is (nc) x 64 x 64\n",
        "            nn.Conv2d(nc, ndf, 4, 2, 1, bias=False),\n",
        "            nn.LeakyReLU(0.2, inplace=True),\n",
        "            # state size. (ndf) x 32 x 32\n",
        "            nn.Conv2d(ndf, ndf * 2, 4, 2, 1, bias=False),\n",
        "            nn.BatchNorm2d(ndf * 2),\n",
        "            nn.LeakyReLU(0.2, inplace=True),\n",
        "            # state size. (ndf*2) x 16 x 16\n",
        "            nn.Conv2d(ndf * 2, ndf * 4, 4, 2, 1, bias=False),\n",
        "            nn.BatchNorm2d(ndf * 4),\n",
        "            nn.LeakyReLU(0.2, inplace=True),\n",
        "            # state size. (ndf*4) x 8 x 8\n",
        "            nn.Conv2d(ndf * 4, ndf * 8, 4, 2, 1, bias=False),\n",
        "            nn.BatchNorm2d(ndf * 8),\n",
        "            nn.LeakyReLU(0.2, inplace=True),\n",
        "            # state size. (ndf*8) x 4 x 4\n",
        "            nn.Conv2d(ndf * 8, 1, 2, 2, 0, bias=False),\n",
        "            nn.Sigmoid())\n",
        "\n",
        "    def forward(self, input):\n",
        "        output = self.main(input)\n",
        "        return output.view(-1, 1).squeeze(1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "2C1OhvapdmYj",
        "colab_type": "code",
        "outputId": "2ad1ac43-349b-456e-a75c-2974f8ccff39",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49
        }
      },
      "cell_type": "code",
      "source": [
        "# Transform on train\n",
        "transform_train =  transforms.Compose([transforms.ToTensor(),\n",
        "                                       transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "# Transform on test\n",
        "transform_test =  transforms.Compose([transforms.ToTensor(),\n",
        "                                      transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "\n",
        "# Load data set \n",
        "dataset_train = CIFAR10(path, end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(path, train=False, transform=transform_test, download=True)\n",
        "\n",
        "# Extract X, y\n",
        "X_train, y_train, X_test, y_test = dataset_train.data, dataset_train.targets, dataset_test.data, dataset_test.targets\n",
        "\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=10, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=10, shuffle=False, num_workers=1)"
      ],
      "execution_count": 132,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "EJYYMglldmSR",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "netD = Discriminator(ngpu=1).to(device)\n",
        "\n",
        "# load weights\n",
        "weights_D = torch.load(os.path.join(path, '/weights/netD_epoch_199.pth'))\n",
        "# weights_G = torch.load(os.path.join(path, '/weights/netG_epoch_199.pth'))\n",
        "netD.load_state_dict(weights_D)\n",
        "# netG.load_state_dict(weights_G)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "fxxlYv_R8RzP",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> I train a linear SVC on top of the discriminator."
      ]
    },
    {
      "metadata": {
        "id": "k9dstv43taip",
        "colab_type": "code",
        "outputId": "5bbae751-356f-4a58-d02c-4994836ebb49",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 214
        }
      },
      "cell_type": "code",
      "source": [
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=1, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=1, shuffle=False, num_workers=1)\n",
        "\n",
        "def get_features(dataloader):\n",
        "  X, y = [], []\n",
        "  model = nn.Sequential(*[netD.main[i].to(device) for i in range(11)])\n",
        "  model.maxpool2 = nn.MaxPool2d(kernel_size=2)\n",
        "  for _, (inputs, targets) in enumerate(dataloader):\n",
        "    inputs, targets = inputs.to(device), targets.to(device)\n",
        "    x = model(inputs)\n",
        "    X.append(x.cpu().data.numpy().flatten())\n",
        "    y.append(np.array(targets.cpu()))\n",
        "  return X, y\n",
        "\n",
        "X_train_, y_train_ = get_features(train_loader)\n",
        "X_test_, y_test_ = get_features(test_loader)\n",
        "\n",
        "from sklearn.model_selection import cross_validate as CV\n",
        "from sklearn.svm import LinearSVC as SVC\n",
        "for C in [1e-4, 5e-4, 1e-3, 5e-3, 1e-2, 5e-2, 1e-1, 5e-1, 1, 5, 10, 50]:\n",
        "  clf = SVC(C=C)\n",
        "  print('C =',C,' | CV scores: ', CV(estimator=clf, X=X_train_, y=y_train_, cv=StratifiedKFold(n_splits=2), scoring='accuracy')['test_score'])"
      ],
      "execution_count": 136,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "C = 0.0001  | CV scores:  [0.1509434  0.17021277]\n",
            "C = 0.0005  | CV scores:  [0.1509434  0.17021277]\n",
            "C = 0.001  | CV scores:  [0.1509434  0.17021277]\n",
            "C = 0.005  | CV scores:  [0.18867925 0.21276596]\n",
            "C = 0.01  | CV scores:  [0.18867925 0.23404255]\n",
            "C = 0.05  | CV scores:  [0.20754717 0.21276596]\n",
            "C = 0.1  | CV scores:  [0.20754717 0.21276596]\n",
            "C = 0.5  | CV scores:  [0.20754717 0.21276596]\n",
            "C = 1  | CV scores:  [0.20754717 0.21276596]\n",
            "C = 5  | CV scores:  [0.20754717 0.21276596]\n",
            "C = 10  | CV scores:  [0.20754717 0.21276596]\n",
            "C = 50  | CV scores:  [0.20754717 0.21276596]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "VvjXE6PbZixJ",
        "colab_type": "code",
        "outputId": "abf78a18-7373-4322-f46d-873d64f5228b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 181
        }
      },
      "cell_type": "code",
      "source": [
        "for n in range(10):\n",
        "  clf = SVC(C=0.5)\n",
        "  clf.fit(X_train_, y_train_)\n",
        "  print('Simulation {:d}: test_acc {}, train_acc {:.3f}'.format(n+1, clf.score(X_test_, y_test_), clf.score(X_train_, y_train_)))"
      ],
      "execution_count": 138,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Simulation 1: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 2: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 3: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 4: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 5: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 6: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 7: test_acc 0.1979, train_acc 1.000\n",
            "Simulation 8: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 9: test_acc 0.1978, train_acc 1.000\n",
            "Simulation 10: test_acc 0.1978, train_acc 1.000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "vTCF6ITLjtGw",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Here are the accuracies obtained so far (with DCGAN): \n",
        "\n",
        "| Model | Number of  epochs  | Train accuracy | Test accuracy |\n",
        "|------|------|------|------|\n",
        "|   ResNet18  | 20 | 35.500 ± 3.828% | 24.672 ± 0.777% |\n",
        "|   VGG16  | 20 | 21.300 ± 3.466% | 18.465 ± 1.940% |\n",
        "|   Pretrained VGG16  | 40 | 69.400 ± 5.219% | 37.909 ± 0.603% |\n",
        "| DCGAN | 200 |100.000 ± 0.000% | 19.780 ± 0.000% |\n"
      ]
    },
    {
      "metadata": {
        "id": "FacYjTPs1wqo",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Incorporating *a priori*\n",
        "Geometrical *a priori* are appealing for image classification tasks. For now, we only consider linear transformations $\\mathcal{T}$ of the inputs $x:\\mathbb{S}^2\\rightarrow\\mathbb{R}$ where $\\mathbb{S}$ is the support of an image, meaning that:\n",
        "\n",
        "$$\\forall u\\in\\mathbb{S}^2,\\mathcal{T}(\\lambda x+\\mu y)(u)=\\lambda \\mathcal{T}(x)(u)+\\mu \\mathcal{T}(y)(u)\\,.$$\n",
        "\n",
        "For instance if an image had an infinite support, a translation $\\mathcal{T}_a$ by $a$ would lead to:\n",
        "\n",
        "$$\\forall u, \\mathcal{T}_a(x)(u)=x(u-a)\\,.$$\n",
        "\n",
        "Otherwise, one has to handle several boundary effects.\n",
        "\n",
        "__Question 7:__ Explain the issues when dealing with translations, rotations, scaling effects, color changes on $32\\times32$ images. Propose several ideas to tackle them."
      ]
    },
    {
      "metadata": {
        "id": "Hdz9muNm1wqp",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> The use of augmentation in deep learning is ubiquitous as it generally allows a network to better generalize. It serves as a type of regularization, reducing the chance of overfitting by extracting more general information from the database and passing it to the network. What we are interested in is the core difference between supervised and unsupervised data augmentation.\n",
        ">\n",
        "> - **Unsupervised augmentation**:  the data expansion task is done regardless of the label of the sample: adding noise, rotating, flipping, shearing the images. Pytorch provides a simple way to implement some random transformations in the data set thanks to `Augmentor` package. The issue when dealing with really poor resolution images is first that random transformations are more likely to result in non-sensical images.  For example, blurring or adding some noise to the training images is not worth it. Moreover, augmenting some images could result in new images with different labels or at least confusing association between the latter and their labels. This is particularly true in medical images, less here. Likewise, a system that uses deep learning to classify or interpret road signs may become incapable of discerning left and right arrows if the training set was augmented with by indiscriminate flipping of images. Just before, we said that \"if an image [do not] have an infinite support, a translation $\\mathcal{T}_a$ by $a$ would lead to several boundary effects\". Those boundary effects are indeed non-neligeable on a $32\\times 32$ images as they occur on $124$ pixels, i.e. $12,1\\%$ of the images. Anyway, as data augmentation is one of the most powerful tool available for solving the issue of small data, one would want to better benefit from it for poor resolution images. \n",
        ">\n",
        "> - **Supervised augmentation**: the data expansion task is done by mixing different samples with the same label in feature space in order to generate a new sample with the same label. The generated sample has to be recognizable as a valid data sample, and also as a sample representative of that specific class, since the label of the data is used to generate the new sample. More generally, I will refer as supervised augmentation any technique that do not randomly draw transformations.\n",
        ">>\n",
        ">>  -  *GAN*: One can thus cite adversarial examples. Goodfellow et al. (\"Generative adversarial nets\", 2012) note that, although augmentation is usually done with the goal of creating images that are as similar as possible to the natural images one expects in the testing set, this does not need to be the case. They further demonstrate that training with adversarial examples can increase the generalization capacity of a network. GAN learn until the synthetic data is nearly indistinguishable from what one would expect real data to look like. That's why they can be used to generate images that augment datasets, as in the strategy employed by Shrivastav et al. (\"Learning from simulated and unsupervised images through adversarial training\", 2016). \n",
        ">>  - *Smart Augmentation*:  proposed by Lemley et al. (\"Smart Augmentation: Learning an Optimal Data Augmentation Strategy\", 2017). SA is the process of learning suitable augmentations when training deep neural networks. The goal of SA is to learn the best augmentation strategy for a given class of input data. It does this by learning to merge two or more samples in one class. This merged sample is then used to train a target network. The loss of the target network is used to inform the augmenter at the same time. This has the result of generating more data for use by the target network. This process often includes letting the network come up with unusual or unexpected but highly performant augmentation strategies.\n",
        ">>  - *Spatial Transformer Networks*: proposed by Jaderberg et al. (\"Spatial Transformer Networks\", 2015). STN combines a localisation network, a grid generator, and a sampler form a spatial transformer. This is a self-contained module which can be dropped into a CNN architecture at any point, and in any number, giving rise to spatial transformer networks. Placing spatial transformers within a CNN allows the network to **learn how to actively transform the feature maps to help minimise the overall cost function of the network during training. The knowledge of how to transform each training sample is compressed and cached in the weights of the localisation network (and also the weights of the layers previous to a spatial transformer) during training** through a parameter $\\theta$ that encodes the transformation.\n",
        ">\n",
        "> Thus, a main solution for boosting the impact of data augmentation is to use some of the non-exhaustive techniques of supervised DA. To make a link with the wording, the goal here is to develop strategies that optimize the transformations (linear or not) $\\mathcal{T}$. In particular, taking the example of translation $\\mathcal{T}_a$, learning the transformations would not necessarily lead in a proeminent weight of $\\mathcal{T}_a$ because of the pitfalls of boundary effects, but rather give more weight to rotation or rescaling.\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "9coNhJ9z1wqq",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Data augmentations"
      ]
    },
    {
      "metadata": {
        "id": "_cC2inrR1wqr",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "__Question 8:__ Propose a set of geometric transformation beyond translation, and incorporate them in your training pipeline. Train the model of the __Question 3__ and __Question 4__ with them and report the accuracies."
      ]
    },
    {
      "metadata": {
        "id": "EjdlJ7j3enLi",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### RESNET18"
      ]
    },
    {
      "metadata": {
        "id": "cxIwcmY1bGGE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Models of the question 3 and 4 have been trained using small transformations (horizontal flip and crop). The latter provide interesting accuracies on both train and test sets. The following section aims at trying some others  \"fancier\" transformations and no DA (data augmentation) at all. "
      ]
    },
    {
      "metadata": {
        "id": "TXGZsnMS1wqs",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def transformations_with_DA():\n",
        "    p = Augmentor.Pipeline()\n",
        "    p.rotate(probability=0.7, max_left_rotation=10, max_right_rotation=10)\n",
        "    p.shear(probability=0.7, max_shear_left=10, max_shear_right=10)\n",
        "    p.histogram_equalisation(probability=0.7)\n",
        "    p.skew_left_right(probability=0.2)\n",
        "    p.flip_random(probability=0.7)\n",
        "    p.random_distortion(probability=0.5, grid_height=16, grid_width=16, magnitude=10)\n",
        "    p.crop_random(probability=0.5, percentage_area=0.4)\n",
        "    return transforms.Compose([p.torch_transform(),\n",
        "                               transforms.Resize((32, 32)),\n",
        "                               transforms.ToTensor(),\n",
        "                               transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "\n",
        "basic_transforms = transforms.Compose([transforms.ToTensor(),\n",
        "                                      transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "qb-1A_N-ctVv",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> I do not use cross-validation here. Indeed, I tried this approach but it often gave me parameters that were clearly non optimal. The set of hyperparameters are the same as for question 3 (resnet18) and 4 (vgg16). First, let's see the impact of the transformations."
      ]
    },
    {
      "metadata": {
        "id": "TWQJ2es3ygDn",
        "colab_type": "code",
        "outputId": "b0cbd1aa-b129-4149-dc39-3c68e83d2991",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        }
      },
      "cell_type": "code",
      "source": [
        "lr, momentum, weight_decay = 0.005, 0.9, 5e-4\n",
        "max_epochs = 20\n",
        "batch_size = 10\n",
        "\n",
        "transform_train =  transformations_with_DA()  \n",
        "transform_test =  basic_transforms\n",
        "dataset_train = CIFAR10(path, end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(path, train=False, transform=transform_test, download=True)\n",
        "  \n",
        "ResNet18_DA = init_resnet()\n",
        "\n",
        "# Define loss and optimizer\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = optim.SGD(ResNet18_DA.parameters(), lr=lr, momentum=0.9, weight_decay=5e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "# Let's see what it has \n",
        "train_loss_r18da, test_loss_r18da, train_acc_r18da, test_acc_r18da = train_test(init_resnet, name='resnet18da', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0)lr, momentum, weight_decay = 0.05, 0.9, 1e-4"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n",
            "...Training n°1...\n",
            "Time elapsed: 263.25 sec. | Best test_acc: 17.680% | Best train_acc: 19.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 263.18 sec. | Best test_acc: 17.200% | Best train_acc: 20.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 263.24 sec. | Best test_acc: 17.650% | Best train_acc: 10.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 263.31 sec. | Best test_acc: 19.150% | Best train_acc: 14.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 262.56 sec. | Best test_acc: 19.330% | Best train_acc: 19.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 262.22 sec. | Best test_acc: 20.830% | Best train_acc: 21.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 262.83 sec. | Best test_acc: 18.920% | Best train_acc: 17.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 262.68 sec. | Best test_acc: 20.420% | Best train_acc: 21.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 262.57 sec. | Best test_acc: 17.670% | Best train_acc: 19.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 262.98 sec. | Best test_acc: 19.520% | Best train_acc: 21.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "MykTUb2pdXgx",
        "colab_type": "code",
        "outputId": "247cbe3f-9780-48b6-e3ed-aa2ee66246b2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        }
      },
      "cell_type": "code",
      "source": [
        "lr, momentum, weight_decay = 0.005, 0.9, 5e-4\n",
        "max_epochs = 20\n",
        "batch_size = 10\n",
        "\n",
        "transform_train =  basic_transforms\n",
        "transform_test =  basic_transforms\n",
        "dataset_train = CIFAR10(path, end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(path, train=False, transform=transform_test, download=True)\n",
        "  \n",
        "ResNet18_NODA = init_resnet()\n",
        "\n",
        "# Define loss and optimizer\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = optim.SGD(ResNet18_NODA.parameters(), lr=lr, momentum=0.9, weight_decay=5e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "# Let's see what it has \n",
        "train_loss_r18noda, test_loss_r18noda, train_acc_r18noda, test_acc_r18noda = train_test(init_resnet, name='resnet18noda', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n",
            "...Training n°1...\n",
            "Time elapsed: 261.28 sec. | Best test_acc: 23.220% | Best train_acc: 98.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 263.35 sec. | Best test_acc: 22.380% | Best train_acc: 96.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 263.88 sec. | Best test_acc: 22.800% | Best train_acc: 94.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 263.31 sec. | Best test_acc: 23.710% | Best train_acc: 100.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 261.47 sec. | Best test_acc: 23.620% | Best train_acc: 97.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 262.34 sec. | Best test_acc: 22.600% | Best train_acc: 98.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 262.10 sec. | Best test_acc: 23.610% | Best train_acc: 98.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 263.39 sec. | Best test_acc: 23.840% | Best train_acc: 100.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 263.34 sec. | Best test_acc: 22.480% | Best train_acc: 99.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 264.07 sec. | Best test_acc: 23.390% | Best train_acc: 93.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "YcHKQM86ef7P",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> And now with no transformations at all (maybe the goal of questions 3 and 4 ?)."
      ]
    },
    {
      "metadata": {
        "id": "vg5OSznheqqz",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### VGG16"
      ]
    },
    {
      "metadata": {
        "id": "FwFbxdvzfYeE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> First with the transformations."
      ]
    },
    {
      "metadata": {
        "id": "y3Nh8Mh1fR31",
        "colab_type": "code",
        "outputId": "2ac0c873-f58c-4da9-e2cf-4f5cb86520ff",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        }
      },
      "cell_type": "code",
      "source": [
        "max_epochs = 20\n",
        "batch_size = 10\n",
        "\n",
        "transform_train =  transformations_with_DA()\n",
        "transform_test =  basic_transforms\n",
        "dataset_train = CIFAR10(path, end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(path, train=False, transform=transform_test, download=True)\n",
        "\n",
        "VGG16_DA = init_vgg()\n",
        "\n",
        "# Define the loss function\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "\n",
        "# And the optimizer\n",
        "optimizer = optim.SGD(VGG16_DA.parameters(), lr=lr, momentum=0.9, weight_decay=1e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "# Let's see what it has\n",
        "train_loss_vgg16da, test_loss_vgg16da, train_acc_vgg16da, test_acc_vgg16da = train_test(init_vgg, name='vgg16da', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n",
            "...Training n°1...\n",
            "Time elapsed: 215.76 sec. | Best test_acc: 13.300% | Best train_acc: 7.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 215.11 sec. | Best test_acc: 11.730% | Best train_acc: 7.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 214.31 sec. | Best test_acc: 11.160% | Best train_acc: 18.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 213.83 sec. | Best test_acc: 11.410% | Best train_acc: 14.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 213.27 sec. | Best test_acc: 13.890% | Best train_acc: 21.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 213.43 sec. | Best test_acc: 13.000% | Best train_acc: 8.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 213.19 sec. | Best test_acc: 10.500% | Best train_acc: 17.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 213.26 sec. | Best test_acc: 10.740% | Best train_acc: 13.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 212.91 sec. | Best test_acc: 13.180% | Best train_acc: 8.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 213.43 sec. | Best test_acc: 13.480% | Best train_acc: 16.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "99484_lWfbhC",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> And without."
      ]
    },
    {
      "metadata": {
        "id": "BYTEinG1x5AJ",
        "colab_type": "code",
        "outputId": "9f5bc270-f8f2-49ec-808f-d3f1e292c030",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        }
      },
      "cell_type": "code",
      "source": [
        "lr, momentum, weight_decay = 0.05, 0.9, 1e-4\n",
        "max_epochs = 20\n",
        "batch_size = 10\n",
        "\n",
        "transform_train =  basic_transforms\n",
        "transform_test =  basic_transforms\n",
        "dataset_train = CIFAR10(path, end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(path, train=False, transform=transform_test, download=True)\n",
        "VGG16_NODA = init_vgg()\n",
        "\n",
        "# Define the loss function\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "\n",
        "# And the optimizer\n",
        "optimizer = optim.SGD(VGG16_NODA.parameters(), lr=lr, momentum=0.9, weight_decay=1e-4)\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "# Let's see what it has\n",
        "train_loss_vgg16noda, test_loss_vgg16noda, train_acc_vgg16noda, test_acc_vgg16noda = train_test(init_vgg, name='vgg16noda', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n",
            "...Training n°1...\n",
            "Time elapsed: 212.70 sec. | Best test_acc: 13.090% | Best train_acc: 9.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 212.86 sec. | Best test_acc: 12.140% | Best train_acc: 19.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 211.91 sec. | Best test_acc: 12.140% | Best train_acc: 12.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 212.07 sec. | Best test_acc: 11.370% | Best train_acc: 15.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 212.10 sec. | Best test_acc: 13.500% | Best train_acc: 10.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 212.04 sec. | Best test_acc: 11.440% | Best train_acc: 13.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 212.22 sec. | Best test_acc: 14.260% | Best train_acc: 23.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 212.28 sec. | Best test_acc: 12.030% | Best train_acc: 13.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 211.90 sec. | Best test_acc: 11.550% | Best train_acc: 20.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 212.20 sec. | Best test_acc: 13.760% | Best train_acc: 15.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "6x_97CXn0xjt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 82
        },
        "outputId": "eb1c764a-84e7-4f12-affc-094d7e468976"
      },
      "cell_type": "code",
      "source": [
        "# Load results from multiple trainings \n",
        "train_loss_r18da, test_loss_r18da, train_acc_r18da, test_acc_r18da = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18da.pk'), 'rb'))\n",
        "train_loss_r18noda, test_loss_r18noda, train_acc_r18noda, test_acc_r18noda = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18noda.pk'), 'rb'))\n",
        "train_loss_vgg16da, test_loss_vgg16da, train_acc_vgg16da, test_acc_vgg16da = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_vgg16da.pk'), 'rb'))\n",
        "train_loss_vgg16noda, test_loss_vgg16noda, train_acc_vgg16noda, test_acc_vgg16noda = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_vgg16noda.pk'), 'rb'))\n",
        "\n",
        "\n",
        "# Compute corresponding stats for ResNet18 and VGG16\n",
        "m_tr_v16da, std_tr_v16da, m_te_v16da, std_te_v16da = stat(train_acc_vgg16da, test_acc_vgg16da)\n",
        "m_tr_v16noda, std_tr_v16noda, m_te_v16noda, std_te_v16noda = stat(train_acc_vgg16noda, test_acc_vgg16noda)\n",
        "m_tr_r18da, std_tr_r18da, m_te_r18da, std_te_r18da = stat(train_acc_r18da, test_acc_r18da)\n",
        "m_tr_r18noda, std_tr_r18noda, m_te_r18noda, std_te_r18noda = stat(train_acc_r18noda, test_acc_r18noda)\n",
        "\n",
        "# Print comparison\n",
        "print('ResNet18 with DA2   : Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (m_tr_r18da, std_tr_r18da, m_te_r18da, std_te_r18da))\n",
        "print('ResNet18 with no DA : Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (m_tr_r18noda, std_tr_r18noda, m_te_r18noda, std_te_r18noda))\n",
        "print('VGG16 with DA2      : Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (m_tr_v16da, std_tr_v16da, m_te_v16da, std_te_v16da))\n",
        "print('VGG16 with no DA    : Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (m_tr_v16noda, std_tr_v16noda, m_te_v16noda, std_te_v16noda))"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ResNet18 with DA2   : Train accuracy: 18.100 ± 3.390% | Test accuracy: 18.837 ± 1.187%\n",
            "ResNet18 with no DA : Train accuracy: 97.300 ± 2.238% | Test accuracy: 23.165 ± 0.524%\n",
            "VGG16 with DA2      : Train accuracy: 12.900 ± 4.867% | Test accuracy: 12.239 ± 1.193%\n",
            "VGG16 with no DA    : Train accuracy: 14.900 ± 4.277% | Test accuracy: 12.528 ± 0.990%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "RUMrCoTyfeGm",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Comparison of accuracies: \n",
        "\n",
        "| Model | Number of  epochs  | Train accuracy | Test accuracy |\n",
        "|------|------|------|------|\n",
        "|   ResNet18 no DA  | 20 | 97.300 ± 2.238%  |23.165 ± 0.524% |\n",
        "|   VGG16 no DA | 20 | 14.900 ± 4.277% |  12.528 ± 0.990%|\n",
        "|   ResNet18  DA1 | 20 | 35.500 ± 3.828% | 24.672 ± 0.777% |\n",
        "|   VGG16  DA1 | 20 | 21.300 ± 3.466% | 18.465 ± 1.940% |\n",
        "|   ResNet18  DA2 | 20 | 18.100 ± 3.390% |  18.837 ± 1.187%|\n",
        "|   VGG16  DA2 | 20 | 12.900 ± 4.867% |12.239 ± 1.193%|\n",
        "| DCGAN | 200 (no training) |100.000 ± 0.000% | 19.780 ± 0.000% |\n",
        "|   **Pretrained VGG16**  | 40 | 69.400 ± 5.219% | **37.909 ± 0.603%** |\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "VqjvB_gC1wqu",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Wavelets\n",
        "\n",
        "__Question 9:__ Use a Scattering Transform as an input to a ResNet-like architecture. You can find a baseline here: https://arxiv.org/pdf/1703.08961.pdf.\n",
        "\n",
        "*Hint:* You can use the following package: https://www.kymat.io/"
      ]
    },
    {
      "metadata": {
        "id": "1Fjk70OvBtQr",
        "colab_type": "code",
        "outputId": "8802e62c-dc80-4003-e646-721001df5f74",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49
        }
      },
      "cell_type": "code",
      "source": [
        "from kymatio import Scattering2D\n",
        "import kymatio.datasets as scattering_datasets\n",
        "\n",
        "# Transform on train\n",
        "transform_train =  transforms.Compose([transforms.RandomCrop(32, padding=4), \n",
        "                                       transforms.RandomHorizontalFlip(), \n",
        "                                       transforms.ToTensor(),\n",
        "                                       transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "# Transform on test\n",
        "transform_test =  transforms.Compose([transforms.ToTensor(),\n",
        "                                      transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))])\n",
        "\n",
        "# Load data set \n",
        "dataset_train = CIFAR10(scattering_datasets.get_dataset_dir('CIFAR'), end=100, transform=transform_train, download=True)\n",
        "dataset_test = CIFAR10(scattering_datasets.get_dataset_dir('CIFAR'), train=False, transform=transform_test, download=True)\n",
        "\n",
        "# Extract X, y\n",
        "X_train, y_train, X_test, y_test = dataset_train.data, dataset_train.targets, dataset_test.data, dataset_test.targets"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "l3e8u6MtA2kI",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def conv3x3(in_planes, out_planes, stride=1):\n",
        "    \"3x3 convolution with padding\"\n",
        "    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,\n",
        "                     padding=1, bias=False)\n",
        "\n",
        "\n",
        "class BasicBlock_(nn.Module):\n",
        "    def __init__(self, inplanes, planes, stride=1, downsample=None):\n",
        "        super(BasicBlock_, self).__init__()\n",
        "        self.conv1 = conv3x3(inplanes, planes, stride)\n",
        "        self.bn1 = nn.BatchNorm2d(planes)\n",
        "        self.relu = nn.ReLU(inplace=True)\n",
        "        self.conv2 = conv3x3(planes, planes)\n",
        "        self.bn2 = nn.BatchNorm2d(planes)\n",
        "        self.downsample = downsample\n",
        "        self.stride = stride\n",
        "\n",
        "    def forward(self, x):\n",
        "        residual = x\n",
        "\n",
        "        out = self.conv1(x)\n",
        "        out = self.bn1(out)\n",
        "        out = self.relu(out)\n",
        "\n",
        "        out = self.conv2(out)\n",
        "        out = self.bn2(out)\n",
        "\n",
        "        if self.downsample is not None:\n",
        "            residual = self.downsample(x)\n",
        "\n",
        "        out += residual\n",
        "        out = self.relu(out)\n",
        "\n",
        "        return out\n",
        "\n",
        "\n",
        "class Scattering2dResNet(nn.Module):\n",
        "    def __init__(self, in_channels,  k=2, n=4, num_classes=10):\n",
        "        super(Scattering2dResNet, self).__init__()\n",
        "        self.inplanes = 16 * k\n",
        "        self.ichannels = 16 * k\n",
        "        self.K = in_channels\n",
        "        self.init_conv = nn.Sequential(\n",
        "            nn.BatchNorm2d(in_channels, eps=1e-5, affine=False),\n",
        "            nn.Conv2d(in_channels, self.ichannels,\n",
        "                  kernel_size=3, stride=1, padding=1, bias=False),\n",
        "            nn.BatchNorm2d(self.ichannels),\n",
        "            nn.ReLU(True)\n",
        "        )\n",
        "\n",
        "        self.layer2 = self._make_layer(BasicBlock_, 32 * k, n)\n",
        "        self.layer3 = self._make_layer(BasicBlock_, 64 * k, n)\n",
        "        self.avgpool = nn.AdaptiveAvgPool2d(2)\n",
        "        self.fc = nn.Linear(64 * k * 4, num_classes)\n",
        "\n",
        "    def _make_layer(self, block, planes, blocks, stride=1):\n",
        "        downsample = None\n",
        "        if stride != 1 or self.inplanes != planes:\n",
        "            downsample = nn.Sequential(\n",
        "                nn.Conv2d(self.inplanes, planes, kernel_size=1, stride=stride, bias=False),\n",
        "                nn.BatchNorm2d(planes))\n",
        "        layers = []\n",
        "        layers.append(block(self.inplanes, planes, stride, downsample))\n",
        "        self.inplanes = planes\n",
        "        for i in range(1, blocks):\n",
        "            layers.append(block(self.inplanes, planes))\n",
        "        return nn.Sequential(*layers)\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = x.view(x.size(0), self.K, 8, 8)\n",
        "        x = self.init_conv(x)\n",
        "        x = self.layer2(x)\n",
        "        x = self.layer3(x)\n",
        "        x = self.avgpool(x)\n",
        "        x = x.view(x.size(0), -1)\n",
        "        x = self.fc(x)\n",
        "        return x"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "RnaAte_4CxOp",
        "colab_type": "code",
        "outputId": "2fba85ce-bbe0-45e7-ae28-c42066080317",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 65
        }
      },
      "cell_type": "code",
      "source": [
        "scattering = Scattering2D(J=2, shape=(32, 32), max_order=1).cuda()\n",
        "K = 17*3\n",
        "\n",
        "ResNet18_sc = NeuralNetClassifier(\n",
        "    Scattering2dResNet(in_channels=K),\n",
        "    iterator_train__shuffle=True,\n",
        "    criterion=nn.CrossEntropyLoss,\n",
        "    optimizer=torch.optim.SGD,\n",
        "    optimizer__momentum=0.95,\n",
        "    optimizer__weight_decay=1e-4,\n",
        "    device=device,\n",
        "    callbacks=[EarlyStopping(monitor='valid_loss', patience=5, threshold=1e-4)],\n",
        "    verbose=0)\n",
        "\n",
        "params = {'lr': [0.005, 0.01, 0.02, 0.05, 0.1, 0.5],\n",
        "          'batch_size': [5, 10, 15, 20, 25, 30],\n",
        "          'max_epochs': [10, 20, 30, 50]}\n",
        "\n",
        "\n",
        "# I do not use X_train and y_train here in order to avoid errors. \n",
        "### First with the Basic Block\n",
        "# CV is done over 2 stratified folds.\n",
        "gs_resnet18_sc = GridSearchCV(ResNet18_sc, params, refit=False, cv=StratifiedKFold(n_splits=2), scoring='accuracy', verbose=0)\n",
        "print('...Gridsearch for ResNet18 with scaterring...')\n",
        "t = time.time()\n",
        "X_sc = np.array([scattering(torch.cuda.FloatTensor(X.reshape(3, 32, 32).astype('double'))) for X in X_train])\n",
        "gs_resnet18_sc.fit(X_sc, np.array(y_train).astype('int64'))\n",
        "print('Cross-validation finished. Time elapsed: {:5.2f} sec.'.format(time.time()-t))\n",
        "print('Best accuracy on val set: {:0.4f}, best parameters: '.format(gs_resnet18_sc.best_score_), gs_resnet18_sc.best_params_)\n",
        "pickle.dump(gs_resnet18_sc, open(os.path.join(path,'cv_resnet18_sc.pk'), 'wb'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Gridsearch for ResNet18 with scaterring...\n",
            "Cross-validation finished. Time elapsed: 205.09 sec.\n",
            "Best accuracy on val set: 0.2200, best parameters:  {'batch_size': 10, 'lr': 0.005, 'max_epochs': 30}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "pfIYCR9n_VyQ",
        "colab_type": "code",
        "outputId": "c293a7c8-b106-410f-e381-ba46e8c8d0ee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 345
        }
      },
      "cell_type": "code",
      "source": [
        "# Let's see what it has\n",
        "lr, momentum, weight_decay = 0.005, 0.9, 5e-4\n",
        "max_epochs = 30\n",
        "batch_size = 10\n",
        "\n",
        "# Format data with data.DataLoader\n",
        "train_loader = torch.utils.data.DataLoader(dataset_train, batch_size=batch_size, shuffle=True, num_workers=2)\n",
        "test_loader = torch.utils.data.DataLoader(dataset_test, batch_size=batch_size, shuffle=False, num_workers=1)\n",
        "\n",
        "def init_resnet_sc():\n",
        "  ResNet18_sc = Scattering2dResNet(K).to(device)\n",
        "  ResNet18_sc = ResNet18_sc.to(device)\n",
        "  if device == 'cuda':\n",
        "      ResNet18_sc = torch.nn.DataParallel(ResNet18_sc)\n",
        "      cudnn.benchmark = True\n",
        "  return ResNet18_sc\n",
        "\n",
        "ResNet18_sc = init_resnet_sc()\n",
        "\n",
        "# Define loss and optimizer\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = optim.SGD(ResNet18_sc.parameters(), lr=lr, momentum=0.9, weight_decay=5e-4)\n",
        "\n",
        "\n",
        "# Let's see what it has\n",
        "train_loss_r18sc, test_loss_r18sc, train_acc_r18sc, test_acc_r18sc = train_test(init_resnet_sc, name='resnet18_scattered', max_epochs=max_epochs,\n",
        "                                                                                criterion=criterion,\n",
        "                                                                                train_loader=train_loader, test_loader=test_loader,\n",
        "                                                                                N=10, verbose=0, scattering=scattering)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "...Training n°1...\n",
            "Time elapsed: 998.90 sec. | Best test_acc: 28.680% | Best train_acc: 74.000%\n",
            "...Training n°2...\n",
            "Time elapsed: 992.01 sec. | Best test_acc: 31.580% | Best train_acc: 65.000%\n",
            "...Training n°3...\n",
            "Time elapsed: 985.33 sec. | Best test_acc: 29.890% | Best train_acc: 65.000%\n",
            "...Training n°4...\n",
            "Time elapsed: 962.95 sec. | Best test_acc: 30.600% | Best train_acc: 65.000%\n",
            "...Training n°5...\n",
            "Time elapsed: 949.15 sec. | Best test_acc: 28.930% | Best train_acc: 60.000%\n",
            "...Training n°6...\n",
            "Time elapsed: 943.85 sec. | Best test_acc: 29.860% | Best train_acc: 75.000%\n",
            "...Training n°7...\n",
            "Time elapsed: 933.05 sec. | Best test_acc: 31.540% | Best train_acc: 61.000%\n",
            "...Training n°8...\n",
            "Time elapsed: 970.42 sec. | Best test_acc: 29.620% | Best train_acc: 72.000%\n",
            "...Training n°9...\n",
            "Time elapsed: 984.51 sec. | Best test_acc: 28.200% | Best train_acc: 66.000%\n",
            "...Training n°10...\n",
            "Time elapsed: 976.11 sec. | Best test_acc: 30.860% | Best train_acc: 79.000%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "KQAuQY1C2qJq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 32
        },
        "outputId": "f387b904-f966-490d-e0b7-4ce96134265d"
      },
      "cell_type": "code",
      "source": [
        "# Load results from multiple trainings \n",
        "train_loss_r18sc, test_loss_r18sc, train_acc_r18sc, test_acc_r18sc = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18_scattered.pk'), 'rb'))\n",
        "\n",
        "# Compute corresponding stats for ResNet18 with scattering\n",
        "m_tr_r18sc, std_tr_r18sc, m_te_r18sc, std_te_r18sc = stat(train_acc_r18sc, test_acc_r18sc)\n",
        "\n",
        "# Print comparison\n",
        "print('ResNet18 with scattering   : Train accuracy: %0.3f ± %2.3f%% | Test accuracy: %0.3f ± %2.3f%%' % (m_tr_r18sc, std_tr_r18sc, m_te_r18sc, std_te_r18sc))"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ResNet18 with scattering   : Train accuracy: 68.200 ± 6.046% | Test accuracy: 29.976 ± 1.107%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "bXfqUxs_2q0F",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> Comparison of accuracies: \n",
        "\n",
        "| Model | Number of  epochs  | Train accuracy | Test accuracy |\n",
        "|------|------|------|------|\n",
        "|   ResNet18 no DA  | 20 | 97.30 ± 2.238%  |23.165 ± 0.524% |\n",
        "|   VGG16 no DA | 20 | 14.900 ± 4.277% |  12.528 ± 0.990%|\n",
        "|   ResNet18  DA1 | 20 | 35.500 ± 3.828% | 24.672 ± 0.777% |\n",
        "|   VGG16  DA1 | 20 | 21.300 ± 3.466% | 18.465 ± 1.940% |\n",
        "|   ResNet18  DA2 | 20 | 18.100 ± 3.390% |  18.837 ± 1.187%|\n",
        "|   VGG16  DA2 | 20 | 12.900 ± 4.867% |12.239 ± 1.193%|\n",
        "| DCGAN | 200 (no training) |100.000 ± 0.000% | 19.780 ± 0.000% |\n",
        "|   **Pretrained VGG16**  | 40 | 69.400 ± 5.219% | **37.909 ± 0.603%** |\n",
        "| Scattering ResNet18 | 30 | 68.200 ± 6.046% | 29.976 ± 1.107%|\n",
        "\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "0AZBZX9q1wqv",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Weak supervision"
      ]
    },
    {
      "metadata": {
        "id": "3mPLTCoo1wqw",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Weakly supervised techniques permit to tackle the issue of labeled data. An introduction to those techniques can be found here: https://hazyresearch.github.io/snorkel/blog/ws_blog_post.html.\n",
        "\n",
        "__(Open) Question 10:__ Pick a weakly supervised method that will potentially use $\\mathcal{X}\\cup\\mathcal{X}_{\\text{train}}$ to train a representation (a subset of $\\mathcal{X}$ is also fine). Evaluate it and report the accuracies. You should be careful in the choice of your method, in order to avoid heavy computational effort."
      ]
    },
    {
      "metadata": {
        "id": "GkfbXMbNkdCt",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> I didn't do this one. "
      ]
    },
    {
      "metadata": {
        "id": "IEWfGIMR1wqx",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Conclusions"
      ]
    },
    {
      "metadata": {
        "id": "HZZwObBI1wqz",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "__Question 11:__ Write a short report explaining the pros and the cons of each methods that you implemented. 25% of the grade of this project will correspond to this question, thus, it should be done carefully. In particular, please add a plot that will summarize all your numerical results."
      ]
    },
    {
      "metadata": {
        "id": "DCfBSzFekhqH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 716
        },
        "outputId": "2ec7f1f9-7db2-4402-b3c3-4803882c2fc3"
      },
      "cell_type": "code",
      "source": [
        "# Load results from multiple trainings \n",
        "train_loss_r18sc, test_loss_r18sc, train_acc_r18sc, test_acc_r18sc = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18_scattered.pk'), 'rb'))\n",
        "train_loss_r18da, test_loss_r18da, train_acc_r18da, test_acc_r18da = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18da.pk'), 'rb'))\n",
        "train_loss_r18noda, test_loss_r18noda, train_acc_r18noda, test_acc_r18noda = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18noda.pk'), 'rb'))\n",
        "train_loss_vgg16da, test_loss_vgg16da, train_acc_vgg16da, test_acc_vgg16da = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_vgg16da.pk'), 'rb'))\n",
        "train_loss_vgg16noda, test_loss_vgg16noda, train_acc_vgg16noda, test_acc_vgg16noda = pickle.load(open(os.path.join(path,'/Simulations/DLIP-TP3/N10LRr0_train_test_vgg16noda.pk'), 'rb'))\n",
        "train_loss_vgg16p, test_loss_vgg16p, train_acc_vgg16p, test_acc_vgg16p = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_vgg16pretrained.pk'), 'rb'))\n",
        "train_loss_r18, test_loss_r18, train_acc_r18, test_acc_r18 = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_resnet18.pk'), 'rb'))\n",
        "train_loss_vgg16, test_loss_vgg16, train_acc_vgg16, test_acc_vgg16 = pickle.load(open(os.path.join(path,'/Simulations/N10LRr0_train_test_vgg16.pk'), 'rb'))\n",
        "\n",
        "# Define a small function to return mean and std\n",
        "def stat(train_acc, test_acc):\n",
        "  imax = np.argmax(test_acc, axis=1)\n",
        "  train_acc_, test_acc_ = train_acc[np.arange(10), imax], test_acc[np.arange(10), imax]\n",
        "  return np.mean(train_acc_), np.std(train_acc_), np.mean(test_acc_), np.std(test_acc_)\n",
        "\n",
        "# Compute means, std\n",
        "m_tr_v16da, std_tr_v16da, m_te_v16da, std_te_v16da = stat(train_acc_vgg16da, test_acc_vgg16da)\n",
        "m_tr_v16noda, std_tr_v16noda, m_te_v16noda, std_te_v16noda = stat(train_acc_vgg16noda, test_acc_vgg16noda)\n",
        "m_tr_r18da, std_tr_r18da, m_te_r18da, std_te_r18da = stat(train_acc_r18da, test_acc_r18da)\n",
        "m_tr_r18noda, std_tr_r18noda, m_te_r18noda, std_te_r18noda = stat(train_acc_r18noda, test_acc_r18noda)\n",
        "m_tr_r18sc, std_tr_r18sc, m_te_r18sc, std_te_r18sc = stat(train_acc_r18sc, test_acc_r18sc)\n",
        "m_tr_v16p, std_tr_v16p, m_te_v16p, std_te_v16p = stat(train_acc_vgg16p, test_acc_vgg16p)\n",
        "m_tr_r18, std_tr_r18, m_te_r18, std_te_r18 = stat(train_acc_r18, test_acc_r18)\n",
        "m_tr_v16, std_tr_v16, m_te_v16, std_te_v16 = stat(train_acc_vgg16, test_acc_vgg16)\n",
        "m_tr_dcgan, std_tr_dcgan, m_te_dcgan, std_te_dcgan = 100, 0, 19.78, 0\n",
        "\n",
        "\n",
        "font = {'family' : 'normal',\n",
        "        'weight' : 'bold',\n",
        "        'size'   : 12}\n",
        "matplotlib.rc('font', **font)\n",
        "plt.figure(figsize=(18, 12))\n",
        "plt.subplot(2,1,1)\n",
        "plt.boxplot([[100], train_acc_r18noda.flatten(),  train_acc_vgg16p.flatten(),\n",
        "             train_acc_r18sc.flatten(), train_acc_r18.flatten(),\n",
        "             train_acc_vgg16.flatten(), train_acc_r18da.flatten(),\n",
        "             train_acc_vgg16noda.flatten(), train_acc_vgg16da.flatten()], vert=False)\n",
        "plt.yticks([1, 2, 3, 4, 5, 6, 7, 8, 9], ['DCGAN', 'ResNet18 no DA', 'Pretrained VGG16',\n",
        "                                      'ResNet18+scattering', 'ResNet18+DA1',\n",
        "                                      'VGG16+DA1', 'ResNet18+DA2',\n",
        "                                      'VGG16 no DA', 'VGG16+DA2'])\n",
        "plt.title('Training accuracies for the 8 models', fontsize=15)\n",
        "plt.subplot(2,1,2)\n",
        "plt.boxplot([test_acc_vgg16p.flatten(), test_acc_r18sc.flatten(),\n",
        "             test_acc_r18.flatten(), [19.78], test_acc_r18noda.flatten(),\n",
        "             test_acc_r18da.flatten(), test_acc_vgg16.flatten(),\n",
        "             test_acc_vgg16noda.flatten(), test_acc_vgg16da.flatten()], vert=False)\n",
        "plt.yticks([1, 2, 3, 4, 5, 6, 7, 8, 9], ['Pretrained VGG16', 'ResNet18+scattering',\n",
        "                                         'ResNet18+DA1', 'DCGAN', 'ResNet18 no DA',\n",
        "                                       'ResNet18+DA2', 'VGG16+DA1',\n",
        "                                         'VGG16 no DA', 'VGG16+DA2'])\n",
        "plt.title('Testing accuracies for the 9 models', fontsize=15)\n",
        "plt.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABHkAAAK9CAYAAACwxbBAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 1296x864 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "rwnq4YQ0rqqi",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> | Model | Number of  epochs  | Train accuracy | Test accuracy | Test accuracy (full data)\n",
        "|------|------|------|------|------|\n",
        "|   ResNet18 no DA  | 20 | 97.30 ± 2.238%  |23.165 ± 0.524% |?|\n",
        "|   VGG16 no DA | 20 | 14.900 ± 4.277% |  12.528 ± 0.990%|?|\n",
        "|   ResNet18  DA1 | 20 | 35.500 ± 3.828% | 24.672 ± 0.777% |93.02% (a)|\n",
        "|   VGG16  DA1 | 20 | 21.300 ± 3.466% | 18.465 ± 1.940% |92.64% (a)|\n",
        "|   ResNet18  DA2 | 20 | 18.100 ± 3.390% |  18.837 ± 1.187%|?|\n",
        "|   VGG16  DA2 | 20 | 12.900 ± 4.867% |12.239 ± 1.193%|?|\n",
        "| DCGAN | 200 (no training) |100.000 ± 0.000% | 19.780 ± 0.000% |82.80% (b)|\n",
        "|   **Pretrained VGG16**  | 40 | 69.400 ± 5.219% | **37.909 ± 0.603%** |92.64% (a) |\n",
        "| Scattering ResNet18 | 30 | 68.200 ± 6.046% | 29.976 ± 1.107%| 93.10% (b)|\n",
        "\n",
        ">Sources:\n",
        ">- (a): Kuang Liu's github:  https://github.com/kuangliu/pytorch-cifar\n",
        ">- (b): \"Scaling the Scattering Transform: Deep Hybrid Networks\", Oyallon et al. (2017), https://arxiv.org/pdf/1703.08961.pdf"
      ]
    },
    {
      "metadata": {
        "id": "dtZ_Y_TGruL1",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "> To conclude, I choose to follow the chronology of the assignment. The first architectures implemented here as part of a classical supervised classification, were ResNet 18 and VGG 16. Those architectures were trained from scratch with 20 epochs. The figure above shows the distribution of the accuracies obtained for each of those 20 epochs, repeated 10 times (10 independent simulations). The table indicates, on the basis of those 10 simulations, the mean and standard deviations of the best accuracy obtained (during say, 20 epochs) on the test set, and the corresponding accuracy on train set. I will first focus on the accuracies obtained on the test set composed of 10 000 images. To come back to ResNet 18 and VGG16, one first trained the models using some techniques of data augmentation, which I called \"+DA1\". The latter correspond to horizontal flipping and random cropping. ResNet18 gave better results than VGG16, with less parameters to train. Both networks behave well in terms of overfitting. In fact, questions 3 and 4 were, I think, dedicated to trying no data augmentation at all for both ResNet and VGG architectures, which I tried later on, question 8. This approach reveals that, still, ResNet18 performs well on test set with an average accuracy of 23.165% (24.672% with DA1). We can't say the same thing for VGG16, which dramatically falls to 12.528%. For ResNet18, no data augmentation implies a huge overfit (97.30% on train set), which is not found for the 2 other data augmentation strategies, DA1 and DA2. Interestingly, VGG16 do not suffer from such overfit in any of the data augmentation strategy, which would be valuable if it didn't give such poor results. As mentioned just above, the goal of this assignment was also to try some more complex data augmentation strategy beyond translation. It's interesting to see how DA2 (where transformations are various) significantly reduces the overfit observed with the first DA strategy (10.828 to -0.737 pts of accuracy). Nevertheless, DA2 also diminishes the accuracies on test set for both architectures ! This is linked to question 7 on the drawback of data augmentation with small data: it can significantly impacts the data set in terms of outliers and the outcome of some transformations can lead to non-sensical images that, with small data, corrupt the modelling. In fact, it would be more valuable to train the parameters controlling the geometric transformations instead, as evoked question 7. This is clear here: combining cropping, flipping, shearing and/or some others \"unsupervised\"/naive transformations becomes too risky and has a huge impact on the performance. \n",
        ">\n",
        "> Using VGG16 network through transfer learning led to the best accuracy on test set. This approach can be costly if some fully connected layers have parameters to be learned but remains very easy to implement. Here, ImageNet features are relevant for our classification task. However, some cases where we deal with specific data sets can't be handled with transfer learning. For example, it's really hard to do transfer learning in medical imaging which suffers from very small amount of data! \n",
        ">\n",
        "> Despite it's wonderful potential, DCGAN was quite deceptive. In fact, when talking with other MVA students, I heard that some of them reached 30% of accuracy with DCGAN, so I will qualify my analysis. In fact, using DCGAN weights is a kind of transfer learning too, from which comes the same advantages and drawbacks. Moreover, I used here weights obtained with a 200 epochs training on CIFAR-10; which has no sense for a real-life application. As advised, I should have used DCGAN trained on ImageNet but I was faced up to many difficulties to achieve it. In any case, DCGAN (when loading pretrained weights) is particularly fast, as training only consists in tuning a SVM. It is particularly robust to overfitting when constant $C$ is well tuned. Here I used the smallest $C$ that gave the best accuracy on test set, and in fact led to a perfect training accuracy... Recall that here $n << p$ which explains such overfitting (the system is over-determined). DCGAN is a very elegant unsupervised learning technique which certainly leads to better results than that of my implementation. I tried lasso logistic regression but it was really not satisfying. \n",
        ">\n",
        "> To finish with, ResNet18 with scattering technique led to amazing results compared to DA1 (for ResNet). Indeed, test accuracy jumped from 24.672% to 29.976% only using a single transformation: scattering. Oyallon et al. obtained 38.9 ± 1.2% with a Wide Residual Network + Scattering on the same test set while training on 100 samples. The also achieve state-of-the-art performance on the whole data set. This inventive transformation is more prone to overfitting as I didn't add some others  transformations. \n"
      ]
    }
  ]
}