model_3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of model_3.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.5"
    },
    "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/JPedro2/ImageClassificationWorkshop/blob/master/model_3.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VENKPm9zoIwr",
        "colab_type": "code",
        "outputId": "cbd30038-32a0-4b4a-a9d0-6464415f8556",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from __future__ import absolute_import, division, print_function\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "\n",
        "# Helper libraries\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "print(tf.__version__)"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1.13.1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ss2PCb_GoIwv",
        "colab_type": "code",
        "outputId": "00a4f2ba-04e0-4b1e-d837-aa273e6fddc1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        }
      },
      "source": [
        "fashion_mnist = keras.datasets.fashion_mnist\n",
        "\n",
        "(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()\n",
        "\n",
        "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat', \n",
        "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "8192/5148 [===============================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CVf5b7P7oIwx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_images = train_images / 255.0\n",
        "\n",
        "test_images = test_images / 255.0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dnXpd6IAoIwz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_images = train_images.reshape(train_images.shape[0], train_images.shape[1], train_images.shape[2], 1)\n",
        "train_labels = train_labels.reshape(train_labels.shape[0], 1)\n",
        "\n",
        "test_images = test_images.reshape(test_images.shape[0], test_images.shape[1], test_images.shape[2], 1)\n",
        "test_labels = test_labels.reshape(test_labels.shape[0], 1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JTLKc2iWoIw1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        },
        "outputId": "7b3bd1ab-5ac2-4425-94be-c68cf07f8169"
      },
      "source": [
        "model = keras.Sequential([\n",
        "        keras.layers.Conv2D(filters = 32, kernel_size = (5,5),activation ='relu', padding='same', \n",
        "                            input_shape = (28,28,1)),\n",
        "        keras.layers.MaxPooling2D(pool_size=(2,2)),       \n",
        "\n",
        "        keras.layers.Flatten(),\n",
        "        keras.layers.Dense(512, activation = \"relu\"),\n",
        "        keras.layers.Dense(10, activation = \"softmax\")])"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Colocations handled automatically by placer.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V8HmCF-hoIw5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam', \n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "40juzS_5pVAo",
        "colab_type": "code",
        "outputId": "8c34243f-e267-4fd7-b727-59951ad7fb3b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 714
        }
      },
      "source": [
        "from time import time\n",
        "\n",
        "tensorboard = tf.keras.callbacks.TensorBoard(log_dir=\"logs/{}\".format(time()))\n",
        "\n",
        "history = model.fit(train_images, train_labels, epochs=20, validation_split=0.2, callbacks=[tensorboard])"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/20\n",
            "48000/48000 [==============================] - 11s 239us/sample - loss: 0.3807 - acc: 0.8641 - val_loss: 0.2948 - val_acc: 0.8921\n",
            "Epoch 2/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.2553 - acc: 0.9076 - val_loss: 0.2622 - val_acc: 0.9007\n",
            "Epoch 3/20\n",
            "48000/48000 [==============================] - 8s 171us/sample - loss: 0.2041 - acc: 0.9229 - val_loss: 0.2545 - val_acc: 0.9082\n",
            "Epoch 4/20\n",
            "48000/48000 [==============================] - 8s 166us/sample - loss: 0.1661 - acc: 0.9376 - val_loss: 0.2700 - val_acc: 0.9062\n",
            "Epoch 5/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.1346 - acc: 0.9500 - val_loss: 0.2798 - val_acc: 0.9062\n",
            "Epoch 6/20\n",
            "48000/48000 [==============================] - 8s 164us/sample - loss: 0.1051 - acc: 0.9605 - val_loss: 0.2749 - val_acc: 0.9162\n",
            "Epoch 7/20\n",
            "48000/48000 [==============================] - 8s 165us/sample - loss: 0.0866 - acc: 0.9676 - val_loss: 0.3041 - val_acc: 0.9121\n",
            "Epoch 8/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0684 - acc: 0.9749 - val_loss: 0.3354 - val_acc: 0.9141\n",
            "Epoch 9/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0557 - acc: 0.9796 - val_loss: 0.3694 - val_acc: 0.9096\n",
            "Epoch 10/20\n",
            "48000/48000 [==============================] - 8s 162us/sample - loss: 0.0457 - acc: 0.9839 - val_loss: 0.3813 - val_acc: 0.9127\n",
            "Epoch 11/20\n",
            "48000/48000 [==============================] - 8s 162us/sample - loss: 0.0374 - acc: 0.9868 - val_loss: 0.3890 - val_acc: 0.9137\n",
            "Epoch 12/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0332 - acc: 0.9884 - val_loss: 0.4470 - val_acc: 0.9148\n",
            "Epoch 13/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0283 - acc: 0.9908 - val_loss: 0.4532 - val_acc: 0.9103\n",
            "Epoch 14/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0244 - acc: 0.9912 - val_loss: 0.5170 - val_acc: 0.9098\n",
            "Epoch 15/20\n",
            "48000/48000 [==============================] - 8s 162us/sample - loss: 0.0218 - acc: 0.9923 - val_loss: 0.5036 - val_acc: 0.9119\n",
            "Epoch 16/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0198 - acc: 0.9933 - val_loss: 0.5024 - val_acc: 0.9119\n",
            "Epoch 17/20\n",
            "48000/48000 [==============================] - 8s 165us/sample - loss: 0.0170 - acc: 0.9940 - val_loss: 0.5286 - val_acc: 0.9104\n",
            "Epoch 18/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0186 - acc: 0.9942 - val_loss: 0.5477 - val_acc: 0.9117\n",
            "Epoch 19/20\n",
            "48000/48000 [==============================] - 8s 162us/sample - loss: 0.0157 - acc: 0.9947 - val_loss: 0.5976 - val_acc: 0.9050\n",
            "Epoch 20/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.0155 - acc: 0.9952 - val_loss: 0.5467 - val_acc: 0.9094\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YhtgQp8hqazX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "826bea3c-6a8d-417b-aa50-8a17aabaa340"
      },
      "source": [
        "test_loss, test_acc = model.evaluate(test_images, test_labels)\n",
        "\n",
        "print('Test accuracy:', test_acc)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "10000/10000 [==============================] - 1s 54us/sample - loss: 0.5509 - acc: 0.9075\n",
            "Test accuracy: 0.9075\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cj_52tvlorav",
        "colab_type": "text"
      },
      "source": [
        "# Plot Accuracy over time for comparison"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a-8mbcXLoc4B",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "outputId": "a44c500c-5341-4c08-e1bc-c2ae74a29735"
      },
      "source": [
        "#model.fit() returns a History object that contains a dictionary with everything that happened during training:\n",
        "history_dict = history.history\n",
        "history_dict.keys()\n",
        "\n",
        "# There are four entries: one for each monitored metric during training and validation. \n",
        "# We can use these to plot the training and validation loss and accuracy for comparison.\n",
        "acc = history_dict['accuracy']\n",
        "val_acc = history_dict['val_accuracy']\n",
        "loss = history_dict['loss']\n",
        "val_loss = history_dict['val_loss']\n",
        "\n",
        "#epochs = range(1, len(acc) + 1)\n",
        "epochs = range(len(acc))\n",
        "\n",
        "\n",
        "# \"r\" is for \"solid red line\"\n",
        "plt.plot(epochs, loss, 'r', label='Training loss')\n",
        "# b is for \"solid blue line\"\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend(loc=0)\n",
        "plt.show()\n",
        "\n",
        "plt.plot(epochs, acc, 'r', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation Accuracy')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend(loc=0)\n",
        "plt.show()\n",
        "\n"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKTjnnMvkScE551wmTwrOOecy/X+283mmk1UVZgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nS/8PPCFstzvONHgAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ureUhH0roIw7",
        "colab_type": "text"
      },
      "source": [
        "# Define EarlyStopping Callbacks"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tf1VdLasoIw8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = keras.Sequential([\n",
        "        keras.layers.Conv2D(filters = 32, kernel_size = (5,5),activation ='relu', padding='same', \n",
        "                            input_shape = (28,28,1)),\n",
        "        keras.layers.MaxPooling2D(pool_size=(2,2)),       \n",
        "\n",
        "        keras.layers.Flatten(),\n",
        "        keras.layers.Dense(512, activation = \"relu\"),\n",
        "        keras.layers.Dense(10, activation = \"softmax\")])\n",
        "\n",
        "model.compile(optimizer='adam', \n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "earlystop = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=2) # Try with 3"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EHkOjSs4oIw-",
        "colab_type": "code",
        "outputId": "f1c8b402-4d46-45a8-b77f-edf1f870abf1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        }
      },
      "source": [
        "from time import time\n",
        "\n",
        "tensorboard = tf.keras.callbacks.TensorBoard(log_dir=\"logs/{}\".format(time()))\n",
        "\n",
        "history = model.fit(train_images, train_labels, epochs=20, validation_split=0.2, callbacks=[earlystop, tensorboard])"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/20\n",
            "48000/48000 [==============================] - 8s 174us/sample - loss: 0.3832 - acc: 0.8614 - val_loss: 0.2958 - val_acc: 0.8950\n",
            "Epoch 2/20\n",
            "48000/48000 [==============================] - 8s 171us/sample - loss: 0.2557 - acc: 0.9061 - val_loss: 0.2672 - val_acc: 0.9001\n",
            "Epoch 3/20\n",
            "48000/48000 [==============================] - 8s 170us/sample - loss: 0.2057 - acc: 0.9232 - val_loss: 0.2434 - val_acc: 0.9103\n",
            "Epoch 4/20\n",
            "48000/48000 [==============================] - 8s 163us/sample - loss: 0.1669 - acc: 0.9380 - val_loss: 0.2378 - val_acc: 0.9152\n",
            "Epoch 5/20\n",
            "48000/48000 [==============================] - 8s 166us/sample - loss: 0.1339 - acc: 0.9496 - val_loss: 0.2624 - val_acc: 0.9137\n",
            "Epoch 6/20\n",
            "48000/48000 [==============================] - 8s 165us/sample - loss: 0.1089 - acc: 0.9595 - val_loss: 0.2742 - val_acc: 0.9108\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w8MHR5-EoIxB",
        "colab_type": "code",
        "outputId": "ee95ab4c-4659-4787-fbde-fcf1a250e619",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "test_loss, test_acc = model.evaluate(test_images, test_labels)\n",
        "\n",
        "print('Test accuracy:', test_acc)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "10000/10000 [==============================] - 1s 56us/sample - loss: 0.2868 - acc: 0.9097\n",
            "Test accuracy: 0.9097\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xlaR6Eh0oIxD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "outputId": "1ed053a5-a163-4026-a214-ba4055f36b3b"
      },
      "source": [
        "#model.fit() returns a History object that contains a dictionary with everything that happened during training:\n",
        "history_dict = history.history\n",
        "history_dict.keys()\n",
        "\n",
        "# There are four entries: one for each monitored metric during training and validation. \n",
        "# We can use these to plot the training and validation loss and accuracy for comparison.\n",
        "acc = history_dict['accuracy']\n",
        "val_acc = history_dict['val_accuracy']\n",
        "loss = history_dict['loss']\n",
        "val_loss = history_dict['val_loss']\n",
        "\n",
        "#epochs = range(1, len(acc) + 1)\n",
        "epochs = range(len(acc))\n",
        "\n",
        "\n",
        "# \"r\" is for \"solid red line\"\n",
        "plt.plot(epochs, loss, 'r', label='Training loss')\n",
        "# b is for \"solid blue line\"\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend(loc=0)\n",
        "plt.show()\n",
        "\n",
        "plt.plot(epochs, acc, 'r', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation Accuracy')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend(loc=0)\n",
        "plt.show()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nY4wxQVmyMMYYE9T/A4nZKjNdQ5jpAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "source": [
        "# Next Steps\n",
        "\n",
        "By playing with the _early stop_ method we can see that the loss is significantly decreased, together with the model training duration without affecting the accuracy. Now, from the graphs above it does look like there is still room to improve the model, as we should aim for a smaller gap between both the _tranning_ & _validation loss_, as well as _trainning_ & _validation accuracy_. For that we need to find a way to make the model better at generalising. To achieve that, on the next and final section, we will be looking at adding a _Dropout layer_, which randomly and temporarily drops out nodes during the tranning phase of the model. \n",
        "\n",
        "[Prev](https://github.com/JPedro2/ImageClassificationWorkshop/blob/master/model_2.ipynb) - [Next](https://github.com/JPedro2/ImageClassificationWorkshop/blob/master/model_4.ipynb)"
      ],
      "cell_type": "markdown",
      "metadata": {}
    }
  ]
}