Created using Colaboratory

TensorFlow Deployment/Course 3 - TensorFlow Datasets/Week 2/Examples/Week2ExerciseQ.ipynb

@@ -0,0 +1,271 @@
+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "name": "Week 2 Exercise Answer.ipynb",
+      "provenance": [],
+      "collapsed_sections": [],
+      "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/lmoroney/dlaicourse/blob/master/TensorFlow%20Deployment/Course%203%20-%20TensorFlow%20Datasets/Week%202/Examples/Week2ExerciseQ.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "zX4Kg8DUTKWO",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "# https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab_type": "code",
+        "id": "ioLbtB3uGKPX",
+        "colab": {}
+      },
+      "source": [
+        "try:\n",
+        "  # %tensorflow_version only exists in Colab.\n",
+        "  %tensorflow_version 2.x\n",
+        "except Exception:\n",
+        "  pass"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "HLSSb7Qly6xf",
+        "colab_type": "text"
+      },
+      "source": [
+        "# TRANSFER LEARNING\n",
+        "The next code block will download the mobilenet model from TensorFlow Hub, and\n",
+        "use its learned features, extracted as feature_extractor and set to be \n",
+        "fine tuned by making them trainable.\n"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "tSW2AcBLuiHv",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "import tensorflow_hub as hub\n",
+        "model_selection = (\"mobilenet_v2\", 224, 1280) \n",
+        "\n",
+        "handle_base, pixels, FV_SIZE = model_selection\n",
+        "\n",
+        "IMAGE_SIZE = (pixels, pixels)\n",
+        "\n",
+        "MODULE_HANDLE =\"https://tfhub.dev/google/tf2-preview/{}/feature_vector/4\".format(handle_base)\n",
+        "\n",
+        "feature_extractor = hub.KerasLayer(MODULE_HANDLE,\n",
+        "                                   input_shape=IMAGE_SIZE + (3,))\n",
+        "\n",
+        "feature_extractor.trainable = True  "
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "jZ2qZGehzPgM",
+        "colab_type": "text"
+      },
+      "source": [
+        "## Import libraries and set up the splits"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "QGWOsReCW451",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "import tensorflow as tf\n",
+        "import tensorflow_datasets as tfds\n"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "QUSLZO8IuEtt",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "# Here is where you will write your code\n",
+        "# You need to use subsets of the original data, which is entirely in the 'train'\n",
+        "# split. I.E. 'train' contains 25000 records.\n",
+        "# Split this up so that you get\n",
+        "# The first 10% is your 'new' training set\n",
+        "# The last 10% is your validation and test sets, split down the middle \n",
+        "# (i.e. the first half of the last 10% is validation, the second half is test)\n",
+        "# These 3 recordsets should be called\n",
+        "# train_examples, validation_examples and test_examples respectively\n",
+        "\n",
+        "splits = ['train[:10%]', 'train[90%:95%]', 'train[95%:]']\n",
+        "splits, info = tfds.load('cats_vs_dogs', with_info=True, as_supervised=True, split=splits)\n",
+        "(train_examples, validation_examples, test_examples) = splits"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "LtSJorjivpS9",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "num_examples = 2500\n",
+        "num_classes = 2"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "nkh5t21-uZFs",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "# This will turn the 3 sets into batches\n",
+        "# so we can train\n",
+        "# This code should not be changed\n",
+        "\n",
+        "def format_image(image, label):\n",
+        "    image = tf.image.resize(image, IMAGE_SIZE) / 255.0\n",
+        "    return  image, label\n",
+        "    \n",
+        "BATCH_SIZE =  32\n",
+        "\n",
+        "train_batches = train_examples.shuffle(num_examples).map(format_image).batch(BATCH_SIZE)\n",
+        "validation_batches = validation_examples.map(format_image).batch(BATCH_SIZE)\n",
+        "test_batches = test_examples.map(format_image).batch(BATCH_SIZE)"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "rmyQ207suyGY",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "# The new model will take the features from the mobilenet via transfer learning\n",
+        "# And add a new dense layer at the bottom, with 2 classes -- for cats and dogs\n",
+        "\n",
+        "model = tf.keras.Sequential([\n",
+        "        feature_extractor,\n",
+        "        tf.keras.layers.Dense(2, activation='softmax')\n",
+        "])\n",
+        "\n",
+        "model.summary()"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "yVFjq8GHu9iN",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "# Compile the model with adam optimizer and sparse categorical crossentropy, \n",
+        "# and track the accuracy metric\n",
+        "    \n",
+        "model.compile(optimizer='adam',\n",
+        "              loss='sparse_categorical_crossentropy',\n",
+        "              metrics=['accuracy'])"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "dLIwqtilvBcN",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "# Train it for a number of epochs. You should not need many\n",
+        "# Train on the train_Batches, and validation on the validation_batches\n",
+        "EPOCHS = 5\n",
+        "\n",
+        "history = model.fit(train_batches,\n",
+        "                    epochs=EPOCHS,\n",
+        "                    validation_data=validation_batches)"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "3jkG0zBHvEnP",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "# Evaluate the model on the test batches\n",
+        "eval_results = model.evaluate(test_batches, verbose=0)\n",
+        "\n",
+        "# And print the results. You should have >93% accuracy\n",
+        "for metric, value in zip(model.metrics_names, eval_results):\n",
+        "    print(metric + ': {:.4}'.format(value))"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    }
+  ]
+}