train.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "train.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyNHeTDTbvvtrkwCP2uDezsJ",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/kevin3010/Python-Plays-Dino/blob/master/train.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BuZkQxe72ICW",
        "colab_type": "text"
      },
      "source": [
        "*This files requires to have **training_data(3ch).npy** and **testing_data(3ch).npy** file which is generated by prepare_data.py file*"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2GOq0l-1kN0M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')\n",
        "\n",
        "%cd gdrive/'My Drive'/'Deep Learning'/Convolution/'python plays dino'\n",
        "!ls"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KxhnYg-VknCw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "training_data = np.load(\"training_data(3ch).npy\", allow_pickle=True)\n",
        "testing_data = np.load(\"testing_data(3ch).npy\", allow_pickle=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jQ0zw7TJkuZh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "xtrain = []\n",
        "ytrain = []\n",
        "\n",
        "for i in range(len(training_data)):\n",
        "  xtrain.append(training_data[i][0])\n",
        "  ytrain.append(training_data[i][1])\n",
        "\n",
        "xtest = []\n",
        "ytest = []\n",
        "\n",
        "for i in range(len(testing_data)):\n",
        "  xtest.append(testing_data[i][0])\n",
        "  ytest.append(testing_data[i][1])\n",
        "\n",
        "xtrain = np.asarray(xtrain,dtype=float)\n",
        "ytrain = np.asarray(ytrain,dtype=float)\n",
        "\n",
        "xtest = np.asarray(xtest,dtype=float)\n",
        "ytest = np.asarray(ytest,dtype=float)\n",
        "\n",
        "xtrain = xtrain.astype('float32')/255\n",
        "xtest = xtest.astype('float32')/255\n",
        "\n",
        "print(xtrain.shape)\n",
        "print(xtrain[0])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-RD8pmX-lfsh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense,Convolution2D,MaxPooling2D,Flatten,Dropout\n",
        "\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Convolution2D(filters=16,kernel_size=3,padding='same',activation='relu',input_shape=(80,80,3)))\n",
        "model.add(MaxPooling2D(pool_size=2))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Convolution2D(filters=32,kernel_size=3,padding='same',activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=2))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Convolution2D(filters=64,kernel_size=3,padding='same',activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=2))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(500,activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Dense(2,activation='softmax'))\n",
        "model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fLbwixSxvOGB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "from keras.callbacks import ModelCheckpoint   \n",
        "\n",
        "checkpointer = ModelCheckpoint(filepath='aug_model.weights.best.hdf5', verbose=1, \n",
        "                               save_best_only=True)\n",
        "\n",
        "\n",
        "# train the model\n",
        "checkpointer = ModelCheckpoint(filepath='aug_model.weights.best.hdf5', verbose=1,save_best_only=True)\n",
        "\n",
        "model.fit(xtrain,ytrain,batch_size=32,epochs=100,validation_split=0.10,verbose=1,shuffle=True,callbacks=[checkpointer])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dsg7D8em2rY6",
        "colab_type": "code",
        "outputId": "e6e0569a-95d9-439b-aafc-dde2da67a485",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "model.load_weights('aug_model.weights.best.hdf5')\n",
        "\n",
        "score = model.evaluate(xtest, ytest, verbose=0)\n",
        "print('\\n', 'Test accuracy:', score[1])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            " Test accuracy: 0.9161490683229814\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jKAoXEkK2wB2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save(\"dino_model\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3LvgnL9g7hPS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "from google.colab import files\n",
        "files.download(\"aug_model.weights.best.hdf5\") "
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}