{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## mnist-hpo\n",
    "This notebook uses Keras to build a model for the mnist dataset with the network's learning rate configurable as a hyperparameter in order to publish in Watson Studios' Deep Learning as a Service\n",
    "\n",
    "First, libraries, including Keras, are imported"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import json\n",
    "import keras\n",
    "import os.path\n",
    "from os import environ\n",
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.optimizers import RMSprop\n",
    "from keras.layers import Dense, Dropout"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Watson Studio will send a json for the Keras model with information on how to vary the models' learning rate. More information on how this is performed is in the Guide of the DLaaSWorkshop repo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "###############################################################################\n",
    "# Set up HPO.\n",
    "###############################################################################\n",
    "\n",
    "config_file = \"config.json\"\n",
    "\n",
    "if os.path.exists(config_file):\n",
    "    with open(config_file, 'r') as f:\n",
    "        json_obj = json.load(f)\n",
    "    learning_rate = json_obj[\"learning_rate\"]\n",
    "else:\n",
    "    learning_rate = 0.001\n",
    "\n",
    "\n",
    "###############################################################################"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#loading the dataset, setting parameters\n",
    "batch_size = 128\n",
    "num_classes = 10\n",
    "epochs = 1\n",
    "\n",
    "# the data, split between train and test sets\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "x_train = x_train.reshape(60000, 784)\n",
    "x_test = x_test.reshape(10000, 784)\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "x_train /= 255\n",
    "x_test /= 255\n",
    "print(x_train.shape[0], 'train samples')\n",
    "print(x_test.shape[0], 'test samples')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert class vectors to binary class matrices, \n",
    "# each class has a node in the last layer of the network, which calculates the strength of a prediction for each class\n",
    "print(y_train[0])\n",
    "y_train = keras.utils.to_categorical(y_train, num_classes)\n",
    "y_test = keras.utils.to_categorical(y_test, num_classes)\n",
    "print(y_train[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model building with Keras\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Dense(512, activation='relu', input_shape=(784,)))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(512, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Code to execute Keras training, observe model statistics\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=RMSprop(lr=learning_rate),\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "\n",
    "history = model.fit(x_train, y_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    validation_data=(x_test, y_test))\n",
    "\n",
    "\n",
    "\n",
    "print(\"Training history:\" + str(history.history))\n",
    "\n",
    "score = model.evaluate(x_test, y_test, verbose=0)\n",
    "print('Test loss:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}