Projeto_aula5.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "view-in-github"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/github/cassiass/keras-tensorflow/blob/master/Projeto_aula5.ipynb)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "sNc3NouWYrN7"
      },
      "source": [
        "Imports"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {},
        "colab_type": "code",
        "id": "4bQCahYjYdUB"
      },
      "outputs": [],
      "source": [
        "import tensorflow\n",
        "from tensorflow import keras\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "f8uuzUjVYvkf"
      },
      "source": [
        "Carregando o dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {},
        "colab_type": "code",
        "id": "QoIys3wMYirH"
      },
      "outputs": [],
      "source": [
        "dataset = keras.datasets.fashion_mnist\n",
        "((imagens_treino, identificacoes_treino), (imagens_teste, identificacoes_teste)) = dataset.load_data()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "oFr0526ZY5E_"
      },
      "source": [
        "Exploração dos dados"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "colab_type": "code",
        "id": "fPIT36hpYl5A",
        "outputId": "c1115b73-38f8-44ee-d16b-9ed85fcea1be"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "9"
            ]
          },
          "execution_count": 8,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(imagens_treino)\n",
        "imagens_treino.shape\n",
        "imagens_teste.shape\n",
        "len(identificacoes_teste)\n",
        "identificacoes_treino.min()\n",
        "identificacoes_treino.max()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "yTWqT9DIY-iB"
      },
      "source": [
        "Exibição dos dados"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 364
        },
        "colab_type": "code",
        "id": "bGESm49JVahh",
        "outputId": "11546b53-773f-453e-ce77-5105fe2cee6f"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.colorbar.Colorbar at 0x7f665618a588>"
            ]
          },
          "execution_count": 3,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f66561dfa20>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "total_de_classificacoes = 10\n",
        "nomes_de_classificacoes = ['Camiseta', 'Calça', 'Pullover', \n",
        "                           'Vestido', 'Casaco', 'Sandália', 'Camisa',\n",
        "                           'Tênis', 'Bolsa', 'Bota']\n",
        "'''\n",
        "plt.imshow(imagens_treino[0])\n",
        "plt.title(identificacoes_treino[0])\n",
        "\n",
        "for imagem in range(10):\n",
        "  plt.subplot(2, 5, imagem+1)\n",
        "  plt.imshow(imagens_treino[imagem])\n",
        "  plt.title(nomes_de_classificacoes[identificacoes_treino[imagem]])\n",
        "'''  \n",
        "plt.imshow(imagens_treino[0])\n",
        "plt.colorbar()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "#normalização\n",
        "imagens_treino = imagens_treino/float(255)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 222
        },
        "colab_type": "code",
        "id": "Ral_hdl9ulGG",
        "outputId": "34eaa022-a1a6-446c-9fe2-c39692c4ae3d"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/5\n",
            "48000/48000 [==============================] - 10s 200us/step - loss: 2.3028 - acc: 0.0990 - val_loss: 2.3028 - val_acc: 0.0983\n",
            "Epoch 2/5\n",
            "48000/48000 [==============================] - 9s 190us/step - loss: 2.3028 - acc: 0.0996 - val_loss: 2.3028 - val_acc: 0.0983\n",
            "Epoch 3/5\n",
            "48000/48000 [==============================] - 9s 193us/step - loss: 2.3028 - acc: 0.0985 - val_loss: 2.3029 - val_acc: 0.0958\n",
            "Epoch 4/5\n",
            "48000/48000 [==============================] - 10s 205us/step - loss: 2.3028 - acc: 0.0980 - val_loss: 2.3028 - val_acc: 0.0958\n",
            "Epoch 5/5\n",
            "48000/48000 [==============================] - 9s 196us/step - loss: 2.3028 - acc: 0.1004 - val_loss: 2.3027 - val_acc: 0.0995\n"
          ]
        }
      ],
      "source": [
        "\n",
        "\n",
        "modelo = keras.Sequential([ \n",
        "    keras.layers.Flatten(input_shape=(28, 28)),\n",
        "    keras.layers.Dense(256, activation=tensorflow.nn.relu),\n",
        "    keras.layers.Dense(128, activation=tensorflow.nn.relu),\n",
        "    keras.layers.Dense(64, activation=tensorflow.nn.relu),\n",
        "    keras.layers.Dense(10, activation=tensorflow.nn.softmax)\n",
        "])\n",
        "\n",
        "modelo.compile(optimizer='adam', \n",
        "               loss='sparse_categorical_crossentropy',\n",
        "               metrics=['accuracy'])\n",
        "\n",
        "historico = modelo.fit(imagens_treino, identificacoes_treino, epochs=5, validation_split=0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "colab_type": "code",
        "id": "I2kLEHPYUdhd",
        "outputId": "2cffc8b5-101a-4a45-9490-f8cf3d1d0f53"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "resultado teste: 1\n",
            "número da imagem de teste: 2\n"
          ]
        }
      ],
      "source": [
        "testes = modelo.predict(imagens_teste)\n",
        "print('resultado teste:', np.argmax(testes[1]))\n",
        "print('número da imagem de teste:', identificacoes_teste[1])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        },
        "colab_type": "code",
        "id": "wsbr9WqyXQ2G",
        "outputId": "eedc6940-270e-436a-da5a-93e86d6ea372"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "10000/10000 [==============================] - 0s 39us/step\n",
            "Perda do teste: 13.729463269042968\n",
            "Acurácia do teste: 0.1105\n"
          ]
        }
      ],
      "source": [
        "perda_teste, acuracia_teste = modelo.evaluate(imagens_teste, identificacoes_teste)\n",
        "print('Perda do teste:', perda_teste)\n",
        "print('Acurácia do teste:', acuracia_teste)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "include_colab_link": true,
      "name": "Projeto_aula5.ipynb",
      "provenance": [],
      "version": "0.3.2"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}