tutorialKeras.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TensorFlow e Keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.0.0\n"
     ]
    }
   ],
   "source": [
    "# Import do TF e da ferramentas usadas\n",
    "from __future__ import absolute_import, division, print_function, unicode_literals\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers\n",
    "\n",
    "# Import de outras bibliotecas que serão usada\n",
    "import numpy as np\n",
    "import datetime\n",
    "import os\n",
    "\n",
    "# Imprimindo versão do TensorFlow\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exemplo 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Carregando base de dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Passando base de dados para one hot encoding\n",
    "mapping = np.identity(10, dtype=int)\n",
    "\n",
    "y_train = np.array([mapping[y] for y in y_train])\n",
    "y_test = np.array([mapping[y] for y in y_test])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Montando modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "modelo = tf.keras.Sequential()\n",
    "\n",
    "modelo.add(layers.Flatten())\n",
    "modelo.add(layers.Dense(800, kernel_initializer=\"random_uniform\", bias_initializer=\"random_uniform\", activation=\"sigmoid\"))\n",
    "modelo.add(layers.Dense(10, kernel_initializer=\"random_uniform\", bias_initializer=\"random_uniform\", activation=\"sigmoid\"))\n",
    "\n",
    "modelo.compile(optimizer=\"sgd\", loss=\"categorical_crossentropy\", metrics=[\"binary_accuracy\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Treinando o modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "log_dir = os.path.join( \"logs\", \"fit\", datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\"))\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=log_dir, histogram_freq=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples\n",
      "Epoch 1/99\n",
      "60000/60000 [==============================] - 2s 41us/sample - loss: 2.3973 - binary_accuracy: 0.4837\n",
      "Epoch 2/99\n",
      "60000/60000 [==============================] - 2s 26us/sample - loss: 2.3790 - binary_accuracy: 0.4818\n",
      "Epoch 3/99\n",
      "60000/60000 [==============================] - 2s 25us/sample - loss: 2.3623 - binary_accuracy: 0.4799\n",
      "Epoch 4/99\n",
      "60000/60000 [==============================] - 2s 26us/sample - loss: 2.3470 - binary_accuracy: 0.4781\n",
      "[...]\n",
      "Epoch 96/99\n",
      "60000/60000 [==============================] - 2s 27us/sample - loss: 1.2007 - binary_accuracy: 0.9089\n",
      "Epoch 97/99\n",
      "60000/60000 [==============================] - 2s 25us/sample - loss: 1.1912 - binary_accuracy: 0.9087\n",
      "Epoch 98/99\n",
      "60000/60000 [==============================] - 2s 27us/sample - loss: 1.1817 - binary_accuracy: 0.9086\n",
      "Epoch 99/99\n",
      "60000/60000 [==============================] - 2s 27us/sample - loss: 1.1725 - binary_accuracy: 0.9084\n"
     ]
    }
   ],
   "source": [
    "results = modelo.fit(x_train, y_train, batch_size = 60000, epochs=99)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exemplo 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Carregando base de dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.cifar10.load_data()\n",
    "# Normalizando os valores dos pixel para serem entre 0 e 1\n",
    "train_images, test_images = train_images / 255.0, test_images / 255.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Montando modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "modelo = tf.keras.Sequential()\n",
    "\n",
    "modelo.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(32, 32, 3)))\n",
    "modelo.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "modelo.add(layers.MaxPooling2D((2, 2)))\n",
    "modelo.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "modelo.add(layers.Flatten())\n",
    "modelo.add(layers.Dense(64, activation='relu'))\n",
    "modelo.add(layers.Dense(10, activation='softmax'))\n",
    "\n",
    "modelo.compile(optimizer=\"adam\", loss=\"sparse_categorical_crossentropy\", metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Treinando o modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "log_dir = os.path.join( \"logs\", \"fit\", datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\"))\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=log_dir, histogram_freq=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 50000 samples\n",
      "Epoch 1/20\n",
      "50000/50000 [==============================] - 138s 3ms/sample - loss: 1.4302 - accuracy: 0.4832\n",
      "Epoch 2/20\n",
      "50000/50000 [==============================] - 142s 3ms/sample - loss: 1.0061 - accuracy: 0.6466\n",
      "Epoch 3/20\n",
      "50000/50000 [==============================] - 151s 3ms/sample - loss: 0.8440 - accuracy: 0.7072\n",
      "Epoch 4/20\n",
      "50000/50000 [==============================] - 144s 3ms/sample - loss: 0.7344 - accuracy: 0.7447\n",
      "[...]\n",
      "Epoch 18/20\n",
      "50000/50000 [==============================] - 136s 3ms/sample - loss: 0.1187 - accuracy: 0.9574\n",
      "Epoch 19/20\n",
      "50000/50000 [==============================] - 137s 3ms/sample - loss: 0.1227 - accuracy: 0.9569\n",
      "Epoch 20/20\n",
      "50000/50000 [==============================] - 138s 3ms/sample - loss: 0.1079 - accuracy: 0.9612\n"
     ]
    }
   ],
   "source": [
    "results = modelo.fit(train_images, train_labels, epochs=20, callbacks=[tensorboard_callback])"
   ]
  }
 ],
 "metadata": {
  "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}