FINAL_FASHION_MNIST_WITH_DROPOUT_MODIFIED3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "FINAL_FASHION_MNIST_WITH_DROPOUT_MODIFIED3",
      "version": "0.3.2",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "metadata": {
        "id": "rZFoBK4AIAyi",
        "colab_type": "code",
        "outputId": "cdd8680e-7d0c-40f7-f7b7-4aae496eb93c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        }
      },
      "cell_type": "code",
      "source": [
        "# http://pytorch.org/\n",
        "from os.path import exists\n",
        "from wheel.pep425tags import get_abbr_impl, get_impl_ver, get_abi_tag\n",
        "platform = '{}{}-{}'.format(get_abbr_impl(), get_impl_ver(), get_abi_tag())\n",
        "cuda_output = !ldconfig -p|grep cudart.so|sed -e 's/.*\\.\\([0-9]*\\)\\.\\([0-9]*\\)$/cu\\1\\2/'\n",
        "accelerator = cuda_output[0] if exists('/dev/nvidia0') else 'cpu'\n",
        "\n",
        "!pip install -q http://download.pytorch.org/whl/{accelerator}/torch-0.4.1-{platform}-linux_x86_64.whl torchvision\n",
        "import torch"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "tcmalloc: large alloc 1073750016 bytes == 0x58ece000 @  0x7fbf5dcf22a4 0x591a07 0x5b5d56 0x502e9a 0x506859 0x502209 0x502f3d 0x506859 0x504c28 0x502540 0x502f3d 0x506859 0x504c28 0x502540 0x502f3d 0x506859 0x504c28 0x502540 0x502f3d 0x507641 0x502209 0x502f3d 0x506859 0x504c28 0x502540 0x502f3d 0x507641 0x504c28 0x502540 0x502f3d 0x507641\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "HlflMzNEIJQK",
        "colab_type": "code",
        "outputId": "fdb84880-f45e-4f44-9cb2-6d85bc888357",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        }
      },
      "cell_type": "code",
      "source": [
        "import torch\n",
        "from torchvision import datasets, transforms\n",
        "# Define a transform to normalize the data\n",
        "transform = transforms.Compose([transforms.ToTensor(),\n",
        "                                transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])\n",
        "# Download and load the training data\n",
        "trainset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=True, train=True, transform=transform)\n",
        "trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True)\n",
        "\n",
        "# Download and load the test data\n",
        "testset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=True, train=False, transform=transform)\n",
        "testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=True)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz\n",
            "Processing...\n",
            "Done!\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "t6sN5E4eILaU",
        "colab_type": "code",
        "outputId": "04c478b1-8645-4fea-cf55-731931b6cb06",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 210
        }
      },
      "cell_type": "code",
      "source": [
        "!wget -c https://raw.githubusercontent.com/udacity/deep-learning-v2-pytorch/master/intro-to-pytorch/helper.py"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2018-12-27 13:22:44--  https://raw.githubusercontent.com/udacity/deep-learning-v2-pytorch/master/intro-to-pytorch/helper.py\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 2813 (2.7K) [text/plain]\n",
            "Saving to: ‘helper.py’\n",
            "\n",
            "\rhelper.py             0%[                    ]       0  --.-KB/s               \rhelper.py           100%[===================>]   2.75K  --.-KB/s    in 0s      \n",
            "\n",
            "2018-12-27 13:22:44 (43.7 MB/s) - ‘helper.py’ saved [2813/2813]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "a5FFCtGfIPZS",
        "colab_type": "code",
        "outputId": "0f68d6b4-dbee-4af2-98d5-641cfdfc52d7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 330
        }
      },
      "cell_type": "code",
      "source": [
        "import helper\n",
        "images, labels = next(iter(trainloader))\n",
        "helper.imshow(images[0,:]);"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAATkAAAE5CAYAAADr4VfxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEE3IAdGEHBDtfwAvNL69pA/9aAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7fde2772e668>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "hses2CPXIQ0L",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from torch import nn\n",
        "import torch.nn.functional as F\n",
        "from torch import optim\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "luq5s2rjITql",
        "colab_type": "code",
        "outputId": "604c8a06-e54d-41d4-a590-36c5ab09eda0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        }
      },
      "cell_type": "code",
      "source": [
        "model = nn.Sequential(nn.Linear(784,256),nn.ReLU(),nn.Dropout(p=0.2),nn.Linear(256,128),nn.ReLU(),nn.Dropout(p=0.2),nn.Linear(128,10),nn.LogSoftmax(dim=1))\n",
        "\n",
        "criterion = nn.NLLLoss()\n",
        "\n",
        "optimizer = optim.SGD(model.parameters(),lr=0.003)\n",
        "\n",
        "epochs = 30\n",
        "steps = 0\n",
        "train_losses, test_losses = [], []\n",
        "\n",
        "\n",
        "for e in range(epochs):\n",
        "  running_loss = 0\n",
        "  for images , labels in trainloader:\n",
        "    optimizer.zero_grad()\n",
        "  \n",
        "    images = images.flatten(start_dim=1)\n",
        "  \n",
        "    output = model(images)\n",
        "  \n",
        "    loss = criterion(output,labels)\n",
        "  \n",
        "    loss.backward()\n",
        "  \n",
        "    optimizer.step()\n",
        "    \n",
        "    running_loss += loss.item()\n",
        "  else:\n",
        "    test_loss = 0\n",
        "    accuracy = 0\n",
        "    with torch.no_grad():\n",
        "      model.eval()\n",
        "      for images , labels in testloader:\n",
        "        images = images.flatten(start_dim=1)\n",
        "        log_ps = model(images)\n",
        "        test_loss += criterion(log_ps,labels)\n",
        "        \n",
        "        \n",
        "        ps = torch.exp(log_ps)\n",
        "        top_p , top_class = ps.topk(1,dim=1)\n",
        "        equals = top_class == labels.view(*top_class.shape)\n",
        "        accuracy += torch.mean(equals.type(torch.FloatTensor))    \n",
        "    model.train()\n",
        "    train_losses.append(running_loss/len(trainloader))\n",
        "    test_losses.append(test_loss/len(testloader))    \n",
        "    print(\"Epoch: {}/{}.. \".format(e+1, epochs),\"Training Loss: {:.3f}.. \".format(running_loss/len(trainloader)),\"Test Loss: {:.3f}.. \".format(test_loss/len(testloader)),\"Test Accuracy: {:.3f}\".format(accuracy/len(testloader)))\n",
        "    "
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch: 1/30..  Training Loss: 1.670..  Test Loss: 1.024..  Test Accuracy: 0.698\n",
            "Epoch: 2/30..  Training Loss: 0.911..  Test Loss: 0.745..  Test Accuracy: 0.736\n",
            "Epoch: 3/30..  Training Loss: 0.750..  Test Loss: 0.663..  Test Accuracy: 0.755\n",
            "Epoch: 4/30..  Training Loss: 0.678..  Test Loss: 0.618..  Test Accuracy: 0.771\n",
            "Epoch: 5/30..  Training Loss: 0.636..  Test Loss: 0.585..  Test Accuracy: 0.783\n",
            "Epoch: 6/30..  Training Loss: 0.604..  Test Loss: 0.564..  Test Accuracy: 0.793\n",
            "Epoch: 7/30..  Training Loss: 0.578..  Test Loss: 0.543..  Test Accuracy: 0.802\n",
            "Epoch: 8/30..  Training Loss: 0.556..  Test Loss: 0.526..  Test Accuracy: 0.806\n",
            "Epoch: 9/30..  Training Loss: 0.540..  Test Loss: 0.514..  Test Accuracy: 0.810\n",
            "Epoch: 10/30..  Training Loss: 0.524..  Test Loss: 0.502..  Test Accuracy: 0.815\n",
            "Epoch: 11/30..  Training Loss: 0.512..  Test Loss: 0.491..  Test Accuracy: 0.818\n",
            "Epoch: 12/30..  Training Loss: 0.501..  Test Loss: 0.482..  Test Accuracy: 0.821\n",
            "Epoch: 13/30..  Training Loss: 0.490..  Test Loss: 0.474..  Test Accuracy: 0.825\n",
            "Epoch: 14/30..  Training Loss: 0.481..  Test Loss: 0.469..  Test Accuracy: 0.828\n",
            "Epoch: 15/30..  Training Loss: 0.472..  Test Loss: 0.462..  Test Accuracy: 0.832\n",
            "Epoch: 16/30..  Training Loss: 0.465..  Test Loss: 0.456..  Test Accuracy: 0.834\n",
            "Epoch: 17/30..  Training Loss: 0.457..  Test Loss: 0.450..  Test Accuracy: 0.835\n",
            "Epoch: 18/30..  Training Loss: 0.450..  Test Loss: 0.446..  Test Accuracy: 0.837\n",
            "Epoch: 19/30..  Training Loss: 0.446..  Test Loss: 0.443..  Test Accuracy: 0.838\n",
            "Epoch: 20/30..  Training Loss: 0.440..  Test Loss: 0.436..  Test Accuracy: 0.840\n",
            "Epoch: 21/30..  Training Loss: 0.433..  Test Loss: 0.434..  Test Accuracy: 0.840\n",
            "Epoch: 22/30..  Training Loss: 0.428..  Test Loss: 0.430..  Test Accuracy: 0.843\n",
            "Epoch: 23/30..  Training Loss: 0.424..  Test Loss: 0.426..  Test Accuracy: 0.843\n",
            "Epoch: 24/30..  Training Loss: 0.420..  Test Loss: 0.420..  Test Accuracy: 0.846\n",
            "Epoch: 25/30..  Training Loss: 0.415..  Test Loss: 0.417..  Test Accuracy: 0.847\n",
            "Epoch: 26/30..  Training Loss: 0.410..  Test Loss: 0.418..  Test Accuracy: 0.847\n",
            "Epoch: 27/30..  Training Loss: 0.408..  Test Loss: 0.412..  Test Accuracy: 0.849\n",
            "Epoch: 28/30..  Training Loss: 0.403..  Test Loss: 0.408..  Test Accuracy: 0.851\n",
            "Epoch: 29/30..  Training Loss: 0.399..  Test Loss: 0.407..  Test Accuracy: 0.851\n",
            "Epoch: 30/30..  Training Loss: 0.397..  Test Loss: 0.403..  Test Accuracy: 0.852\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "JM33_3p58ZhP",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "%matplotlib inline\n",
        "%config InlineBackend.figure_format = 'retina'\n",
        "\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "b0ImkWz3-N2u",
        "colab_type": "code",
        "outputId": "444636ae-f69c-4374-8fa7-bfb1d975357e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "cell_type": "code",
      "source": [
        "plt.plot(train_losses, label='Training loss')\n",
        "plt.plot(test_losses, label='Validation loss')\n",
        "plt.legend(frameon=False)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fddc81e5518>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuUAAAHwCAYAAAAByRFLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7fddccc5ccf8>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 370,
              "height": 248
            }
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "imz4F0fkHS56",
        "colab_type": "code",
        "outputId": "a4200b6c-4dc4-4149-dd67-14f6235fa535",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 193
        }
      },
      "cell_type": "code",
      "source": [
        "print(model)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sequential(\n",
            "  (0): Linear(in_features=784, out_features=256, bias=True)\n",
            "  (1): ReLU()\n",
            "  (2): Dropout(p=0.2)\n",
            "  (3): Linear(in_features=256, out_features=128, bias=True)\n",
            "  (4): ReLU()\n",
            "  (5): Dropout(p=0.2)\n",
            "  (6): Linear(in_features=128, out_features=10, bias=True)\n",
            "  (7): LogSoftmax()\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "bPUUAxCcHYky",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "print(model.state_dict().keys())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "osdCchj5I9H_",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}