02-tf-xor-revisited.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "view-in-github"
   },
   "source": [
    "<a href=\"https://colab.research.google.com/github/changsin/DeepLearning-101/blob/master/02-tf-xor-revisited.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "dsJ78nDzn5_W"
   },
   "source": [
    "The code is a slight variation of the article [How to solve XOR problem in Keras](https://medium.com/analytics-vidhya/understanding-basics-of-deep-learning-by-solving-xor-problem-cb3ff6a18a06)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "id": "wCtLvEk8n3dq"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "from tensorflow import keras\n",
    "from keras.layers import Dense\n",
    "from keras.models import Sequential\n",
    "from keras.utils import plot_model\n",
    "from matplotlib import pyplot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "zp22hvCroXmA",
    "outputId": "bb302580-6f87-43dc-88db-266711c73bbd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_6 (Dense)              (None, 2)                 6         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 1)                 3         \n",
      "=================================================================\n",
      "Total params: 9\n",
      "Trainable params: 9\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "[array([[-0.60407144, -0.6535538 ],\n",
      "       [ 1.1836365 , -1.0246705 ]], dtype=float32), array([0., 0.], dtype=float32), array([[0.5840899 ],\n",
      "       [0.60741436]], dtype=float32), array([0.], dtype=float32)]\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(units=2,activation='relu', input_dim=2))\n",
    "model.add(Dense(units=1,activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
    "print(model.summary())\n",
    "print(model.get_weights())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "gWmjPERNje-_"
   },
   "source": [
    "Dense() means fully connected layers where every node in one lalyers is connected to all the nodes in the next layer. The model has one hidden layer with nine learnable parameters. Other than model.summary(), Keras has a handy utility to visualize the model.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "id": "VoX8TL0xaIHQ",
    "outputId": "3b644d8a-56f1-46b9-a151-833bc3ed7d61"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 22,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "keras.utils.plot_model(model, to_file='xor_model.png', show_shapes=True, show_layer_names=True, rankdir='TB', expand_nested=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "F8cG98I3ocng"
   },
   "outputs": [],
   "source": [
    "x = np.array([[0.,0.],[0.,1.],[1.,0.],[1.,1.]])\n",
    "y = np.array([0.,1.,1.,0.])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "rX820xqIofcq"
   },
   "outputs": [],
   "source": [
    "history = model.fit(x,y,epochs=1000,batch_size=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "LFYvBpwInat7",
    "outputId": "3aaabea3-a5c7-47f7-b6d7-447e3c5da99b"
   },
   "outputs": [],
   "source": [
    "model.save('./xor-good-weights.hd5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "VtAcwdUEUick",
    "outputId": "74df5043-5251-4210-dc80-0e9c3591d5ce"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([[-1.3081621, -0.6535538],\n",
      "       [ 1.3078191, -1.0246705]], dtype=float32), array([-0.0001756,  0.       ], dtype=float32), array([[1.6605862 ],\n",
      "       [0.60741436]], dtype=float32), array([-0.4481035], dtype=float32)]\n",
      "[[0.38981178]\n",
      " [0.8485601 ]\n",
      " [0.38981178]\n",
      " [0.38981178]]\n"
     ]
    }
   ],
   "source": [
    "print(model.get_weights())\n",
    "print(model.predict(x,batch_size=4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "WpAXtxOTnDeq"
   },
   "source": [
    "  [array([[-1.3081621, -0.6535538],\n",
    "       [ 1.3078191, -1.0246705]], dtype=float32), array([-0.0001756,  0.       ], dtype=float32), array([[1.6605862 ],\n",
    "       [0.60741436]], dtype=float32), array([-0.4481035], dtype=float32)]\n",
    "\n",
    "[[0.38981178]\n",
    " [0.8485601 ]\n",
    " [0.38981178]\n",
    " [0.38981178]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sjQWWLN1rO6_"
   },
   "source": [
    "1/1 [==============================] - 0s 1ms/step - loss: 0.4892 - accuracy: 0.7500\n",
    "\n",
    "\n",
    "[array([[ 0.24800655,  1.10779   ],\n",
    "       [-1.0030258 ,  1.107669  ]], dtype=float32), array([-0.26201358, -1.1084698 ], dtype=float32), array([[-0.26150066],\n",
    "       [-3.3309755 ]], dtype=float32), array([0.62593275], dtype=float32)]\n",
    "[[0.6515666 ]\n",
    " [0.6515666 ]\n",
    " [0.6515666 ]\n",
    " [0.04472694]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 281
    },
    "id": "rn_N_UFbT3O2",
    "outputId": "a0f1e222-2484-4721-a4f7-d594c98bf1d2"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot loss during training\n",
    "pyplot.subplot(211)\n",
    "pyplot.title('Loss')\n",
    "pyplot.plot(history.history['loss'], label='train')\n",
    "pyplot.legend()\n",
    "\n",
    "# plot accuracy during training\n",
    "pyplot.subplot(212)\n",
    "pyplot.title('Accuracy')\n",
    "pyplot.plot(history.history['accuracy'], label='train')\n",
    "pyplot.legend()\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "DDqShqenmPqo"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyNiA5VlAI3XDJ+kl5haYASh",
   "include_colab_link": true,
   "name": "Untitled0.ipynb",
   "provenance": []
  },
  "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}