PyTorch_Playground.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "PyTorch Playground.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.8"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "gXmCHcwKs6rd"
      },
      "source": [
        "# PyTorch playground"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "PzCCniVwNTdp",
        "colab": {}
      },
      "source": [
        "# Setting seeds to try and ensure we have the same results - this is not guaranteed across PyTorch releases.\n",
        "import torch\n",
        "torch.manual_seed(0)\n",
        "torch.backends.cudnn.deterministic = True\n",
        "torch.backends.cudnn.benchmark = False\n",
        "\n",
        "import numpy as np\n",
        "np.random.seed(0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "fQLW-HL7_0pT",
        "outputId": "a2347d9a-bd38-4fa5-b11a-6c2ae5767134",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
        "print(device)\n"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "cuda\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "PCJzXv0OK1Bs",
        "colab": {}
      },
      "source": [
        "from torchvision import datasets, transforms\n",
        "import torch.nn.functional as F\n",
        "from torch import nn\n",
        "\n",
        "mean, std = (0.5,), (0.5,)\n",
        "\n",
        "# Create a transform and normalise data\n",
        "transform = transforms.Compose([transforms.ToTensor(),\n",
        "                                transforms.Normalize(mean, std)\n",
        "                              ])\n",
        "\n",
        "# Download FMNIST training dataset and load training data\n",
        "trainset = datasets.FashionMNIST('~/.pytorch/FMNIST/', download=True, train=True, transform=transform)\n",
        "trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True)\n",
        "\n",
        "# Download FMNIST test dataset and load test data\n",
        "testset = datasets.FashionMNIST('~/.pytorch/FMNIST/', download=True, train=False, transform=transform)\n",
        "testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "rqMqFbIVrbFH",
        "colab": {}
      },
      "source": [
        "class FMNIST(nn.Module):\n",
        "  def __init__(self):\n",
        "    super().__init__()\n",
        "    self.fc1 = nn.Linear(784, 128)\n",
        "    self.fc2 = nn.Linear(128,64)\n",
        "    self.fc3 = nn.Linear(64,10)\n",
        "    \n",
        "  def forward(self, x):\n",
        "    x = x.view(x.shape[0], -1)\n",
        "    \n",
        "    x = F.relu(self.fc1(x))\n",
        "    x = F.relu(self.fc2(x))\n",
        "    x = self.fc3(x)\n",
        "    x = F.log_softmax(x, dim=1)\n",
        "    \n",
        "    return x\n",
        "\n",
        "model = FMNIST()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "67eZUNEM5b7n",
        "outputId": "65d086ff-3a03-4cee-addd-315c55d7b626",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        }
      },
      "source": [
        "model.to(device)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "FMNIST(\n",
              "  (fc1): Linear(in_features=784, out_features=128, bias=True)\n",
              "  (fc2): Linear(in_features=128, out_features=64, bias=True)\n",
              "  (fc3): Linear(in_features=64, out_features=10, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "XPdDu7KfWEfW"
      },
      "source": [
        "- The only change we have made to the code is that we are going to track the training loss, the testing loss and the accuracy across the 30 epochs.\n",
        "- We'll print out the train loss, the test loss and the accuracy after each epoch.\n",
        "- Because we are running this over 30 epochs this will take a bit longer to run - approx 15 minutes."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "VJLzWi0UqGWm",
        "outputId": "77ce05a3-3f82-4cef-cefc-8a43c4ddbafb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 578
        }
      },
      "source": [
        "from torch import optim\n",
        "criterion = nn.NLLLoss()\n",
        "optimizer = optim.SGD(model.parameters(), lr=0.01)\n",
        "\n",
        "num_epochs = 30\n",
        "train_tracker, test_tracker, accuracy_tracker = [], [], []\n",
        "\n",
        "\n",
        "for i in range(num_epochs):\n",
        "    cum_loss = 0\n",
        "    \n",
        "    for batch, (images, labels) in enumerate(trainloader,1):\n",
        "        images = images.to(device)\n",
        "        labels = labels.to(device)\n",
        "        \n",
        "        optimizer.zero_grad()\n",
        "        output = model(images)\n",
        "        loss = criterion(output, labels)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        \n",
        "        cum_loss += loss.item()\n",
        "    \n",
        "    train_tracker.append(cum_loss/len(trainloader))\n",
        "    print(f\"Epoch({i+1}/{num_epochs}) | Training loss: {cum_loss/len(trainloader)} | \",end='')\n",
        "    \n",
        "    test_loss = 0\n",
        "    num_correct = 0\n",
        "    total = 0\n",
        "    \n",
        "    for batch, (images, labels) in enumerate(testloader,1):\n",
        "        images = images.to(device)\n",
        "        labels = labels.to(device)\n",
        "        \n",
        "        logps = model(images)\n",
        "        batch_loss = criterion(logps, labels)        \n",
        "        test_loss += batch_loss.item()\n",
        "        \n",
        "        output = torch.exp(logps)\n",
        "        pred = torch.argmax(output, 1)\n",
        "        total += labels.size(0)\n",
        "        num_correct += (pred == labels).sum().item()\n",
        "    \n",
        "    test_tracker.append(test_loss/len(testloader))\n",
        "    print(f\"Test loss: {test_loss/len(testloader)} | \", end='')\n",
        "    accuracy_tracker.append(num_correct/total)\n",
        "    print(f'Accuracy : {num_correct/total}')        \n",
        "print(f'\\nNumber correct : {num_correct}, Total : {total}')\n",
        "print(f'Accuracy of the model after 30 epochs on the 10000 test images: {num_correct * 100 / total}% ')\n",
        "    \n",
        "    "
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch(1/30) | Training loss: 1.0330898230835828 | Test loss: 0.6479068924287322 | Accuracy : 0.7614\n",
            "Epoch(2/30) | Training loss: 0.5597374950771901 | Test loss: 0.5414081445545148 | Accuracy : 0.8039\n",
            "Epoch(3/30) | Training loss: 0.4903963123049055 | Test loss: 0.4971817100693466 | Accuracy : 0.8194\n",
            "Epoch(4/30) | Training loss: 0.4554869087615501 | Test loss: 0.47312212540845205 | Accuracy : 0.8298\n",
            "Epoch(5/30) | Training loss: 0.4310528090450047 | Test loss: 0.45723548284761467 | Accuracy : 0.835\n",
            "Epoch(6/30) | Training loss: 0.4134255885315348 | Test loss: 0.44864174571766213 | Accuracy : 0.8359\n",
            "Epoch(7/30) | Training loss: 0.399817388893953 | Test loss: 0.4298076530930343 | Accuracy : 0.8478\n",
            "Epoch(8/30) | Training loss: 0.388061835392833 | Test loss: 0.4293592136566806 | Accuracy : 0.8439\n",
            "Epoch(9/30) | Training loss: 0.3774932924427712 | Test loss: 0.41148661741405534 | Accuracy : 0.8523\n",
            "Epoch(10/30) | Training loss: 0.36851061820221354 | Test loss: 0.4274013562566915 | Accuracy : 0.8427\n",
            "Epoch(11/30) | Training loss: 0.360288798094177 | Test loss: 0.4050957746566481 | Accuracy : 0.8553\n",
            "Epoch(12/30) | Training loss: 0.3528367919581277 | Test loss: 0.442640547444866 | Accuracy : 0.842\n",
            "Epoch(13/30) | Training loss: 0.34551923679136265 | Test loss: 0.3942889951786418 | Accuracy : 0.8599\n",
            "Epoch(14/30) | Training loss: 0.3393631616968717 | Test loss: 0.385670344730851 | Accuracy : 0.8614\n",
            "Epoch(15/30) | Training loss: 0.3335536307871723 | Test loss: 0.3839634743275916 | Accuracy : 0.8618\n",
            "Epoch(16/30) | Training loss: 0.32737260671662116 | Test loss: 0.37746300372728114 | Accuracy : 0.867\n",
            "Epoch(17/30) | Training loss: 0.32203575862305506 | Test loss: 0.3799787531043314 | Accuracy : 0.8627\n",
            "Epoch(18/30) | Training loss: 0.3169646824934462 | Test loss: 0.38316315251171207 | Accuracy : 0.8628\n",
            "Epoch(19/30) | Training loss: 0.31197599076957844 | Test loss: 0.3653619491560444 | Accuracy : 0.8674\n",
            "Epoch(20/30) | Training loss: 0.3071116164866795 | Test loss: 0.3709522712572365 | Accuracy : 0.8697\n",
            "Epoch(21/30) | Training loss: 0.3017437636280365 | Test loss: 0.36106638838151456 | Accuracy : 0.8728\n",
            "Epoch(22/30) | Training loss: 0.2983515392075469 | Test loss: 0.36962069514071105 | Accuracy : 0.8668\n",
            "Epoch(23/30) | Training loss: 0.2934731903678573 | Test loss: 0.357472366493219 | Accuracy : 0.8702\n",
            "Epoch(24/30) | Training loss: 0.28921857981412397 | Test loss: 0.3545182277062896 | Accuracy : 0.8733\n",
            "Epoch(25/30) | Training loss: 0.2855971241270555 | Test loss: 0.3524002528684154 | Accuracy : 0.8745\n",
            "Epoch(26/30) | Training loss: 0.2807486147673399 | Test loss: 0.3812956469264 | Accuracy : 0.8648\n",
            "Epoch(27/30) | Training loss: 0.2763775169356927 | Test loss: 0.35617227785906214 | Accuracy : 0.8739\n",
            "Epoch(28/30) | Training loss: 0.2735283151745542 | Test loss: 0.3554404157741814 | Accuracy : 0.8743\n",
            "Epoch(29/30) | Training loss: 0.270145039449432 | Test loss: 0.35572395990988254 | Accuracy : 0.8723\n",
            "Epoch(30/30) | Training loss: 0.266094913233572 | Test loss: 0.35367005076378016 | Accuracy : 0.8778\n",
            "\n",
            "Number correct : 8778, Total : 10000\n",
            "Accuracy of the model after 30 epochs on the 10000 test images: 87.78% \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "IqNpkYO6V9YI"
      },
      "source": [
        "- Has the accuracy of the model increased?\n",
        "- Now plot the training loss vs the test loss over 30 epochs."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "89a8FdTi-cNM",
        "outputId": "82701451-7413-4c00-d4e7-24f0e375a4c2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "plt.plot(train_tracker, label='Training loss')\n",
        "plt.plot(test_tracker, label='Test loss')\n",
        "plt.legend()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f00c8cb33c8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nabgrpVQE0nBXSqkIpOGulFIR6P8D3J8T5j30lfkAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DHtKTHKKjG3r",
        "colab_type": "text"
      },
      "source": [
        "- Now add the accuracy to the mix."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "AJgyMHm2Pvx5",
        "outputId": "881d9ce0-19cd-4991-86a1-2d014ee00735",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "plt.plot(train_tracker, label='Training loss')\n",
        "plt.plot(test_tracker, label='Test loss')\n",
        "plt.plot(accuracy_tracker, label='Test accuracy')\n",
        "plt.legend()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f00c8da8f98>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nF0KIMCThLoQQYUjCXQghwtD/BxdsaWtscTIRAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "APhVglVTk-og",
        "colab_type": "text"
      },
      "source": [
        "## Further challenges and experiments\n",
        "- Can you get better accuracy from a model if you :\n",
        "    - Add more layers?\n",
        "    - Change the number of nodes in the layers?\n",
        "    - Train over fewer/higher epochs?\n",
        "    \n",
        "- Can you improve on your results if you add additional layers like [Dropout](https://pytorch.org/docs/master/nn.html#torch.nn.Dropout)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ex8FzAY7jG3v",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6ImOG4oXjG3x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CwHV3ig7jG3y",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fg3aU2vJjG30",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9UfGdMuCjG32",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xxOzS2-TjG34",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3kdUCpCyjG35",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RMhC9JJujG37",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}