Weight_Initialization.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Weight_Initialization.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/github/pkraison/ex.ai/blob/master/Weight_Initialization.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "WILul7sonVSs",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#  Training your neural network requires specifying an initial value of the weights. A well chosen initialization method will help learning. \n",
        "- how do you choose the initialization for a new neural network? In this notebook, you will see     how different initializations lead to different results. \n",
        "-A well chosen initialization can:\n",
        "- Speed up the convergence of gradient descent\n",
        "- Increase the odds of gradient descent converging to a lower training (and generalization) error \n"
      ]
    },
    {
      "metadata": {
        "id": "jcs_VONSmq7e",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import h5py\n",
        "import sklearn\n",
        "import sklearn.datasets"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "0_kTGzQan5tf",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def sigmoid(x):\n",
        "    \"\"\"\n",
        "    Compute the sigmoid of x\n",
        "\n",
        "    Arguments:\n",
        "    x -- A scalar or numpy array of any size.\n",
        "\n",
        "    Return:\n",
        "    s -- sigmoid(x)\n",
        "    \"\"\"\n",
        "    s = 1/(1+np.exp(-x))\n",
        "    return s\n",
        "\n",
        "def relu(x):\n",
        "    \"\"\"\n",
        "    Compute the relu of x\n",
        "\n",
        "    Arguments:\n",
        "    x -- A scalar or numpy array of any size.\n",
        "\n",
        "    Return:\n",
        "    s -- relu(x)\n",
        "    \"\"\"\n",
        "    s = np.maximum(0,x)\n",
        "    \n",
        "    return s\n",
        "\n",
        "def forward_propagation(X, parameters):\n",
        "    \"\"\"\n",
        "    Implements the forward propagation (and computes the loss) presented in Figure 2.\n",
        "    \n",
        "    Arguments:\n",
        "    X -- input dataset, of shape (input size, number of examples)\n",
        "    Y -- true \"label\" vector (containing 0 if cat, 1 if non-cat)\n",
        "    parameters -- python dictionary containing your parameters \"W1\", \"b1\", \"W2\", \"b2\", \"W3\", \"b3\":\n",
        "                    W1 -- weight matrix of shape ()\n",
        "                    b1 -- bias vector of shape ()\n",
        "                    W2 -- weight matrix of shape ()\n",
        "                    b2 -- bias vector of shape ()\n",
        "                    W3 -- weight matrix of shape ()\n",
        "                    b3 -- bias vector of shape ()\n",
        "    \n",
        "    Returns:\n",
        "    loss -- the loss function (vanilla logistic loss)\n",
        "    \"\"\"\n",
        "        \n",
        "    # retrieve parameters\n",
        "    W1 = parameters[\"W1\"]\n",
        "    b1 = parameters[\"b1\"]\n",
        "    W2 = parameters[\"W2\"]\n",
        "    b2 = parameters[\"b2\"]\n",
        "    W3 = parameters[\"W3\"]\n",
        "    b3 = parameters[\"b3\"]\n",
        "    \n",
        "    # LINEAR -> RELU -> LINEAR -> RELU -> LINEAR -> SIGMOID\n",
        "    z1 = np.dot(W1, X) + b1\n",
        "    a1 = relu(z1)\n",
        "    z2 = np.dot(W2, a1) + b2\n",
        "    a2 = relu(z2)\n",
        "    z3 = np.dot(W3, a2) + b3\n",
        "    a3 = sigmoid(z3)\n",
        "    \n",
        "    cache = (z1, a1, W1, b1, z2, a2, W2, b2, z3, a3, W3, b3)\n",
        "    \n",
        "    return a3, cache\n",
        "\n",
        "def backward_propagation(X, Y, cache):\n",
        "    \"\"\"\n",
        "    Implement the backward propagation presented in figure 2.\n",
        "    \n",
        "    Arguments:\n",
        "    X -- input dataset, of shape (input size, number of examples)\n",
        "    Y -- true \"label\" vector (containing 0 if cat, 1 if non-cat)\n",
        "    cache -- cache output from forward_propagation()\n",
        "    \n",
        "    Returns:\n",
        "    gradients -- A dictionary with the gradients with respect to each parameter, activation and pre-activation variables\n",
        "    \"\"\"\n",
        "    m = X.shape[1]\n",
        "    (z1, a1, W1, b1, z2, a2, W2, b2, z3, a3, W3, b3) = cache\n",
        "    \n",
        "    dz3 = 1./m * (a3 - Y)\n",
        "    dW3 = np.dot(dz3, a2.T)\n",
        "    db3 = np.sum(dz3, axis=1, keepdims = True)\n",
        "    \n",
        "    da2 = np.dot(W3.T, dz3)\n",
        "    dz2 = np.multiply(da2, np.int64(a2 > 0))\n",
        "    dW2 = np.dot(dz2, a1.T)\n",
        "    db2 = np.sum(dz2, axis=1, keepdims = True)\n",
        "    \n",
        "    da1 = np.dot(W2.T, dz2)\n",
        "    dz1 = np.multiply(da1, np.int64(a1 > 0))\n",
        "    dW1 = np.dot(dz1, X.T)\n",
        "    db1 = np.sum(dz1, axis=1, keepdims = True)\n",
        "    \n",
        "    gradients = {\"dz3\": dz3, \"dW3\": dW3, \"db3\": db3,\n",
        "                 \"da2\": da2, \"dz2\": dz2, \"dW2\": dW2, \"db2\": db2,\n",
        "                 \"da1\": da1, \"dz1\": dz1, \"dW1\": dW1, \"db1\": db1}\n",
        "    \n",
        "    return gradients\n",
        "\n",
        "def update_parameters(parameters, grads, learning_rate):\n",
        "    \"\"\"\n",
        "    Update parameters using gradient descent\n",
        "    \n",
        "    Arguments:\n",
        "    parameters -- python dictionary containing your parameters \n",
        "    grads -- python dictionary containing your gradients, output of n_model_backward\n",
        "    \n",
        "    Returns:\n",
        "    parameters -- python dictionary containing your updated parameters \n",
        "                  parameters['W' + str(i)] = ... \n",
        "                  parameters['b' + str(i)] = ...\n",
        "    \"\"\"\n",
        "    \n",
        "    L = len(parameters) // 2 # number of layers in the neural networks\n",
        "\n",
        "    # Update rule for each parameter\n",
        "    for k in range(L):\n",
        "        parameters[\"W\" + str(k+1)] = parameters[\"W\" + str(k+1)] - learning_rate * grads[\"dW\" + str(k+1)]\n",
        "        parameters[\"b\" + str(k+1)] = parameters[\"b\" + str(k+1)] - learning_rate * grads[\"db\" + str(k+1)]\n",
        "        \n",
        "    return parameters"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "fBLy_eTlrzF6",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "\n",
        "def compute_loss(a3, Y):\n",
        "    \n",
        "    \"\"\"\n",
        "    Implement the loss function\n",
        "    \n",
        "    Arguments:\n",
        "    a3 -- post-activation, output of forward propagation\n",
        "    Y -- \"true\" labels vector, same shape as a3\n",
        "    \n",
        "    Returns:\n",
        "    loss - value of the loss function\n",
        "    \"\"\"\n",
        "    \n",
        "    m = Y.shape[1]\n",
        "    logprobs = np.multiply(-np.log(a3),Y) + np.multiply(-np.log(1 - a3), 1 - Y)\n",
        "    loss = 1./m * np.nansum(logprobs)\n",
        "    \n",
        "    return loss"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "wGwim-I7rqKk",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def load_cat_dataset():\n",
        "    train_dataset = h5py.File('datasets/train_catvnoncat.h5', \"r\")\n",
        "    train_set_x_orig = np.array(train_dataset[\"train_set_x\"][:]) # your train set features\n",
        "    train_set_y_orig = np.array(train_dataset[\"train_set_y\"][:]) # your train set labels\n",
        "\n",
        "    test_dataset = h5py.File('datasets/test_catvnoncat.h5', \"r\")\n",
        "    test_set_x_orig = np.array(test_dataset[\"test_set_x\"][:]) # your test set features\n",
        "    test_set_y_orig = np.array(test_dataset[\"test_set_y\"][:]) # your test set labels\n",
        "\n",
        "    classes = np.array(test_dataset[\"list_classes\"][:]) # the list of classes\n",
        "    \n",
        "    train_set_y = train_set_y_orig.reshape((1, train_set_y_orig.shape[0]))\n",
        "    test_set_y = test_set_y_orig.reshape((1, test_set_y_orig.shape[0]))\n",
        "    \n",
        "    train_set_x_orig = train_set_x_orig.reshape(train_set_x_orig.shape[0], -1).T\n",
        "    test_set_x_orig = test_set_x_orig.reshape(test_set_x_orig.shape[0], -1).T\n",
        "    \n",
        "    train_set_x = train_set_x_orig/255\n",
        "    test_set_x = test_set_x_orig/255\n",
        "\n",
        "    return train_set_x, train_set_y, test_set_x, test_set_y, classes"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ZuVhCelArkQn",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def predict(X, y, parameters):\n",
        "    \"\"\"\n",
        "    This function is used to predict the results of a  n-layer neural network.\n",
        "    \n",
        "    Arguments:\n",
        "    X -- data set of examples you would like to label\n",
        "    parameters -- parameters of the trained model\n",
        "    \n",
        "    Returns:\n",
        "    p -- predictions for the given dataset X\n",
        "    \"\"\"\n",
        "    \n",
        "    m = X.shape[1]\n",
        "    p = np.zeros((1,m), dtype = np.int)\n",
        "    \n",
        "    # Forward propagation\n",
        "    a3, caches = forward_propagation(X, parameters)\n",
        "    \n",
        "    # convert probas to 0/1 predictions\n",
        "    for i in range(0, a3.shape[1]):\n",
        "        if a3[0,i] > 0.5:\n",
        "            p[0,i] = 1\n",
        "        else:\n",
        "            p[0,i] = 0\n",
        "\n",
        "    # print results\n",
        "    print(\"Accuracy: \"  + str(np.mean((p[0,:] == y[0,:]))))\n",
        "    \n",
        "    return p"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "aMw8OIlordv_",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def plot_decision_boundary(model, X, y):\n",
        "    # Set min and max values and give it some padding\n",
        "    x_min, x_max = X[0, :].min() - 1, X[0, :].max() + 1\n",
        "    y_min, y_max = X[1, :].min() - 1, X[1, :].max() + 1\n",
        "    h = 0.01\n",
        "    # Generate a grid of points with distance h between them\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    # Predict the function value for the whole grid\n",
        "    Z = model(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    # Plot the contour and training examples\n",
        "    plt.contourf(xx, yy, Z, cmap=plt.cm.Spectral)\n",
        "    plt.ylabel('x2')\n",
        "    plt.xlabel('x1')\n",
        "    plt.scatter(X[0, :], X[1, :], c=y, cmap=plt.cm.Spectral)\n",
        "    plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ZEzYckrHrY6N",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def predict_dec(parameters, X):\n",
        "    \"\"\"\n",
        "    Used for plotting decision boundary.\n",
        "    \n",
        "    Arguments:\n",
        "    parameters -- python dictionary containing your parameters \n",
        "    X -- input data of size (m, K)\n",
        "    \n",
        "    Returns\n",
        "    predictions -- vector of predictions of our model (red: 0 / blue: 1)\n",
        "    \"\"\"\n",
        "    \n",
        "    # Predict using forward propagation and a classification threshold of 0.5\n",
        "    a3, cache = forward_propagation(X, parameters)\n",
        "    predictions = (a3>0.5)\n",
        "    return predictions\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "6Lm7KTxTrVly",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def load_dataset():\n",
        "    np.random.seed(1)\n",
        "    train_X, train_Y = sklearn.datasets.make_circles(n_samples=300, noise=.05)\n",
        "    np.random.seed(2)\n",
        "    test_X, test_Y = sklearn.datasets.make_circles(n_samples=100, noise=.05)\n",
        "    # Visualize the data\n",
        "    plt.scatter(train_X[:, 0], train_X[:, 1], c=train_Y, s=40, cmap=plt.cm.Spectral);\n",
        "    train_X = train_X.T\n",
        "    train_Y = train_Y.reshape((1, train_Y.shape[0]))\n",
        "    test_X = test_X.T\n",
        "    test_Y = test_Y.reshape((1, test_Y.shape[0]))\n",
        "    return train_X, train_Y, test_X, test_Y"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "makYLzjvoKEc",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "get_ipython().magic('matplotlib inline')\n",
        "plt.rcParams['figure.figsize'] = (7.0, 4.0) # set default size of plots\n",
        "plt.rcParams['image.interpolation'] = 'nearest'\n",
        "plt.rcParams['image.cmap'] = 'gray'\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "bizwc-5YoXDj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "e41dbdec-1eac-4621-d437-d053898f7e11"
      },
      "cell_type": "code",
      "source": [
        "train_X, train_Y, test_X, test_Y = load_dataset()"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAa8AAAD4CAYAAABbl2n6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUkq5HQ1eSiml3M7/A5zHV8We5NGfAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f680f40cf98>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "l50a2DWKoduB",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def model(X, Y, learning_rate = 0.01, num_iterations = 30000, print_cost = True, initialization = \"he\"):\n",
        "    \"\"\"\n",
        "    Implements a three-layer neural network: LINEAR->RELU->LINEAR->RELU->LINEAR->SIGMOID.\n",
        "    \n",
        "    Arguments:\n",
        "    X -- input data, of shape (2, number of examples)\n",
        "    Y -- true \"label\" vector (containing 0 for red dots; 1 for blue dots), of shape (1, number of examples)\n",
        "    learning_rate -- learning rate for gradient descent \n",
        "    num_iterations -- number of iterations to run gradient descent\n",
        "    print_cost -- if True, print the cost every 1000 iterations\n",
        "    initialization -- flag to choose which initialization to use (\"zeros\",\"random\" or \"he\")\n",
        "    \n",
        "    Returns:\n",
        "    parameters -- parameters learnt by the model\n",
        "    \"\"\"\n",
        "        \n",
        "    grads = {}\n",
        "    costs = [] # to keep track of the loss\n",
        "    m = X.shape[1] # number of examples\n",
        "    layers_dims = [X.shape[0], 10, 5, 1]\n",
        "    \n",
        "    # Initialize parameters dictionary.\n",
        "    if initialization == \"zeros\":\n",
        "        parameters = initialize_parameters_zeros(layers_dims)\n",
        "    elif initialization == \"random\":\n",
        "        parameters = initialize_parameters_random(layers_dims)\n",
        "    elif initialization == \"he\":\n",
        "        parameters = initialize_parameters_he(layers_dims)\n",
        "\n",
        "    # Loop (gradient descent)\n",
        "\n",
        "    for i in range(0, num_iterations):\n",
        "\n",
        "        # Forward propagation: LINEAR -> RELU -> LINEAR -> RELU -> LINEAR -> SIGMOID.\n",
        "        a3, cache = forward_propagation(X, parameters)\n",
        "        \n",
        "        # Loss\n",
        "        cost = compute_loss(a3, Y)\n",
        "\n",
        "        # Backward propagation.\n",
        "        grads = backward_propagation(X, Y, cache)\n",
        "        \n",
        "        # Update parameters.\n",
        "        parameters = update_parameters(parameters, grads, learning_rate)\n",
        "        \n",
        "        # Print the loss every 1000 iterations\n",
        "        if print_cost and i % 1000 == 0:\n",
        "            print(\"Cost after iteration {}: {}\".format(i, cost))\n",
        "            costs.append(cost)\n",
        "            \n",
        "    # plot the loss\n",
        "    plt.plot(costs)\n",
        "    plt.ylabel('cost')\n",
        "    plt.xlabel('iterations (per hundreds)')\n",
        "    plt.title(\"Learning rate =\" + str(learning_rate))\n",
        "    plt.show()\n",
        "    \n",
        "    return parameters\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "G7gsQjwRokxb",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def initialize_parameters_zeros(layers_dims):\n",
        "    \"\"\"\n",
        "    Arguments:\n",
        "    layer_dims -- python array (list) containing the size of each layer.\n",
        "    \n",
        "    Returns:\n",
        "    parameters -- python dictionary containing your parameters \"W1\", \"b1\", ..., \"WL\", \"bL\":\n",
        "                    W1 -- weight matrix of shape (layers_dims[1], layers_dims[0])\n",
        "                    b1 -- bias vector of shape (layers_dims[1], 1)\n",
        "                    ...\n",
        "                    WL -- weight matrix of shape (layers_dims[L], layers_dims[L-1])\n",
        "                    bL -- bias vector of shape (layers_dims[L], 1)\n",
        "    \"\"\"\n",
        "    \n",
        "    parameters = {}\n",
        "    L = len(layers_dims)            # number of layers in the network\n",
        "    \n",
        "    for l in range(1, L):\n",
        "        \n",
        "        parameters['W' + str(l)] = np.zeros((layers_dims[l],layers_dims[l-1]))\n",
        "        parameters['b' + str(l)] = np.zeros((layers_dims[l],1))\n",
        "        \n",
        "    return parameters\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "gxge6n7Zou_7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "f8c06f44-0fe0-4f2f-ab8d-240967250df3"
      },
      "cell_type": "code",
      "source": [
        "parameters = initialize_parameters_zeros([3,2,1])\n",
        "print(\"W1 = \" + str(parameters[\"W1\"]))\n",
        "print(\"b1 = \" + str(parameters[\"b1\"]))\n",
        "print(\"W2 = \" + str(parameters[\"W2\"]))\n",
        "print(\"b2 = \" + str(parameters[\"b2\"]))"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "W1 = [[0. 0. 0.]\n",
            " [0. 0. 0.]]\n",
            "b1 = [[0.]\n",
            " [0.]]\n",
            "W2 = [[0. 0.]]\n",
            "b2 = [[0.]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "X4m6UI_yo64K",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 872
        },
        "outputId": "82a02737-8c44-4801-dd69-4e2348deb02b"
      },
      "cell_type": "code",
      "source": [
        "parameters = model(train_X, train_Y, initialization = \"zeros\")\n",
        "print (\"On the train set:\")\n",
        "predictions_train = predict(train_X, train_Y, parameters)\n",
        "print (\"On the test set:\")\n",
        "predictions_test = predict(test_X, test_Y, parameters)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cost after iteration 0: 0.6931471805599453\n",
            "Cost after iteration 1000: 0.6931471805599453\n",
            "Cost after iteration 2000: 0.6931471805599453\n",
            "Cost after iteration 3000: 0.6931471805599453\n",
            "Cost after iteration 4000: 0.6931471805599453\n",
            "Cost after iteration 5000: 0.6931471805599453\n",
            "Cost after iteration 6000: 0.6931471805599453\n",
            "Cost after iteration 7000: 0.6931471805599453\n",
            "Cost after iteration 8000: 0.6931471805599453\n",
            "Cost after iteration 9000: 0.6931471805599453\n",
            "Cost after iteration 10000: 0.6931471805599455\n",
            "Cost after iteration 11000: 0.6931471805599453\n",
            "Cost after iteration 12000: 0.6931471805599453\n",
            "Cost after iteration 13000: 0.6931471805599453\n",
            "Cost after iteration 14000: 0.6931471805599453\n",
            "Cost after iteration 15000: 0.6931471805599453\n",
            "Cost after iteration 16000: 0.6931471805599455\n",
            "Cost after iteration 17000: 0.6931471805599453\n",
            "Cost after iteration 18000: 0.6931471805599453\n",
            "Cost after iteration 19000: 0.6931471805599453\n",
            "Cost after iteration 20000: 0.6931471805599453\n",
            "Cost after iteration 21000: 0.6931471805599453\n",
            "Cost after iteration 22000: 0.6931471805599455\n",
            "Cost after iteration 23000: 0.6931471805599453\n",
            "Cost after iteration 24000: 0.6931471805599453\n",
            "Cost after iteration 25000: 0.6931471805599453\n",
            "Cost after iteration 26000: 0.6931471805599453\n",
            "Cost after iteration 27000: 0.6931471805599453\n",
            "Cost after iteration 28000: 0.6931471805599455\n",
            "Cost after iteration 29000: 0.6931471805599453\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb0AAAEVCAYAAAB0YTvHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f680ef51240>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "On the train set:\n",
            "Accuracy: 0.5\n",
            "On the test set:\n",
            "Accuracy: 0.5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "3WTfjxzFpBkE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        },
        "outputId": "0a522bfe-be16-4d6b-965e-798e1c0cedac"
      },
      "cell_type": "code",
      "source": [
        "print (\"predictions_train = \" + str(predictions_train))\n",
        "print (\"predictions_test = \" + str(predictions_test))"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "predictions_train = [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0]]\n",
            "predictions_test = [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "91wgHEoPpI_X",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "#plt.title(\"Model with Zeros initialization\")\n",
        "#axes = plt.gca()\n",
        "#axes.set_xlim([-1.5,1.5])\n",
        "#axes.set_ylim([-1.5,1.5])\n",
        "#plot_decision_boundary(lambda x: predict_dec(parameters, x.T), train_X, train_Y)\n",
        "#plt.scatter(X[:, 0], X[:, 1], s=40, c=T, cmap=plt.cm.Spectral)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "aVqubZsgpPmq",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def initialize_parameters_random(layers_dims):\n",
        "    \"\"\"\n",
        "    Arguments:\n",
        "    layer_dims -- python array (list) containing the size of each layer.\n",
        "    \n",
        "    Returns:\n",
        "    parameters -- python dictionary containing your parameters \"W1\", \"b1\", ..., \"WL\", \"bL\":\n",
        "                    W1 -- weight matrix of shape (layers_dims[1], layers_dims[0])\n",
        "                    b1 -- bias vector of shape (layers_dims[1], 1)\n",
        "                    ...\n",
        "                    WL -- weight matrix of shape (layers_dims[L], layers_dims[L-1])\n",
        "                    bL -- bias vector of shape (layers_dims[L], 1)\n",
        "    \"\"\"\n",
        "    \n",
        "    np.random.seed(3)               # This seed makes sure your \"random\" numbers will be the as ours\n",
        "    parameters = {}\n",
        "    L = len(layers_dims)            # integer representing the number of layers\n",
        "    \n",
        "    for l in range(1, L):\n",
        "        \n",
        "        parameters['W' + str(l)] = np.random.randn(layers_dims[l],layers_dims[l-1])*10\n",
        "        parameters['b' + str(l)] = np.zeros((layers_dims[l],1))\n",
        "        \n",
        "\n",
        "    return parameters"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "j7cREyPKsNuj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "e83a49ae-825a-4527-a580-f3ab7f29b813"
      },
      "cell_type": "code",
      "source": [
        "parameters = initialize_parameters_random([3, 2, 1])\n",
        "print(\"W1 = \" + str(parameters[\"W1\"]))\n",
        "print(\"b1 = \" + str(parameters[\"b1\"]))\n",
        "print(\"W2 = \" + str(parameters[\"W2\"]))\n",
        "print(\"b2 = \" + str(parameters[\"b2\"]))"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "W1 = [[ 17.88628473   4.36509851   0.96497468]\n",
            " [-18.63492703  -2.77388203  -3.54758979]]\n",
            "b1 = [[0.]\n",
            " [0.]]\n",
            "W2 = [[-0.82741481 -6.27000677]]\n",
            "b2 = [[0.]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "vhb9tukOsTQl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 940
        },
        "outputId": "16675377-b7ac-4af1-f468-0fca1eb1ee82"
      },
      "cell_type": "code",
      "source": [
        "parameters = model(train_X, train_Y, initialization = \"random\")\n",
        "print (\"On the train set:\")\n",
        "predictions_train = predict(train_X, train_Y, parameters)\n",
        "print (\"On the test set:\")\n",
        "predictions_test = predict(test_X, test_Y, parameters)"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:16: RuntimeWarning: divide by zero encountered in log\n",
            "  app.launch_new_instance()\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:16: RuntimeWarning: invalid value encountered in multiply\n",
            "  app.launch_new_instance()\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Cost after iteration 0: inf\n",
            "Cost after iteration 1000: 0.6247924745506072\n",
            "Cost after iteration 2000: 0.5980258056061102\n",
            "Cost after iteration 3000: 0.5637539062842213\n",
            "Cost after iteration 4000: 0.5501256393526495\n",
            "Cost after iteration 5000: 0.5443826306793814\n",
            "Cost after iteration 6000: 0.5373895855049121\n",
            "Cost after iteration 7000: 0.47157999220550006\n",
            "Cost after iteration 8000: 0.39770475516243037\n",
            "Cost after iteration 9000: 0.3934560146692851\n",
            "Cost after iteration 10000: 0.3920227137490125\n",
            "Cost after iteration 11000: 0.38913700035966736\n",
            "Cost after iteration 12000: 0.3861358766546214\n",
            "Cost after iteration 13000: 0.38497629552893475\n",
            "Cost after iteration 14000: 0.38276694641706693\n",
            "Cost after iteration 15000: 0.38105053248956466\n",
            "Cost after iteration 16000: 0.3799490392460345\n",
            "Cost after iteration 17000: 0.37912933245750763\n",
            "Cost after iteration 18000: 0.37446376042642815\n",
            "Cost after iteration 19000: 0.371619817460682\n",
            "Cost after iteration 20000: 0.36967732765298966\n",
            "Cost after iteration 21000: 0.3630311607367288\n",
            "Cost after iteration 22000: 0.3577474056696152\n",
            "Cost after iteration 23000: 0.35535853789078925\n",
            "Cost after iteration 24000: 0.35396171772377055\n",
            "Cost after iteration 25000: 0.3530157873456813\n",
            "Cost after iteration 26000: 0.3523809777045673\n",
            "Cost after iteration 27000: 0.35196829829403226\n",
            "Cost after iteration 28000: 0.3516478821464336\n",
            "Cost after iteration 29000: 0.35139657535109703\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb4AAAEVCAYAAACfVoDEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niEikKPhERCRSFHwiIhIp/x/cesekAIyb/wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f680efa9550>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "On the train set:\n",
            "Accuracy: 0.85\n",
            "On the test set:\n",
            "Accuracy: 0.84\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "5wwd1erVsYo4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        },
        "outputId": "135790da-342d-45f7-f268-fa6d8232e100"
      },
      "cell_type": "code",
      "source": [
        "print (predictions_train)\n",
        "print (predictions_test)"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1 0 1 1 0 0 1 1 1 1 1 0 1 0 0 1 0 1 1 0 0 0 1 0 1 0 1 1 1 1 0 1 1 0 0 1\n",
            "  1 1 1 1 1 1 1 0 1 1 1 1 0 1 0 1 1 1 1 0 0 1 1 1 1 1 1 1 0 1 0 1 1 1 1 0\n",
            "  0 0 0 0 1 0 1 0 1 1 1 0 0 1 1 1 1 1 1 0 0 1 1 1 0 1 1 0 1 0 1 1 0 1 1 0\n",
            "  1 0 1 1 0 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 0 1 1 1 1 1 1 1 0 1 1 0 0 1 1 0\n",
            "  0 0 1 0 1 0 1 0 1 1 1 0 0 1 1 1 1 0 1 1 0 1 0 1 1 0 1 0 1 1 1 1 0 1 1 1\n",
            "  1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 1 0 1 1 0 1 0 1 1 1 0 1 1 1 0 1 0 1 0 0 1\n",
            "  0 1 1 0 1 1 1 1 1 0 1 1 1 1 1 1 0 1 0 1 0 0 1 1 0 1 1 1 0 0 0 1 1 0 1 1\n",
            "  1 1 0 1 1 0 1 1 0 0 0 1 0 0 0 1 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 1 1 1 1\n",
            "  1 1 1 1 0 0 0 1 1 1 1 1]]\n",
            "[[1 1 1 1 0 1 0 1 1 0 1 1 1 0 0 0 1 1 0 1 0 0 1 0 1 0 1 1 1 1 1 0 1 0 0 1\n",
            "  0 1 1 0 0 1 1 1 1 1 0 1 1 1 0 1 0 1 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 0 1 0\n",
            "  1 1 1 1 1 0 1 0 0 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 0 0]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "1MoJ6cabs1_U",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "#plt.title(\"Model with large random initialization\")\n",
        "#axes = plt.gca()\n",
        "#axes.set_xlim([-1.5,1.5])\n",
        "#axes.set_ylim([-1.5,1.5])\n",
        "#plot_decision_boundary(lambda x: predict_dec(parameters, x.T), train_X, train_Y)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "LiuekNHotBDg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def initialize_parameters_he(layers_dims):\n",
        "    \"\"\"\n",
        "    Arguments:\n",
        "    layer_dims -- python array (list) containing the size of each layer.\n",
        "    \n",
        "    Returns:\n",
        "    parameters -- python dictionary containing your parameters \"W1\", \"b1\", ..., \"WL\", \"bL\":\n",
        "                    W1 -- weight matrix of shape (layers_dims[1], layers_dims[0])\n",
        "                    b1 -- bias vector of shape (layers_dims[1], 1)\n",
        "                    ...\n",
        "                    WL -- weight matrix of shape (layers_dims[L], layers_dims[L-1])\n",
        "                    bL -- bias vector of shape (layers_dims[L], 1)\n",
        "    \"\"\"\n",
        "    \n",
        "    np.random.seed(3)\n",
        "    parameters = {}\n",
        "    L = len(layers_dims) - 1 # integer representing the number of layers\n",
        "     \n",
        "    for l in range(1, L + 1):\n",
        "        \n",
        "        parameters['W' + str(l)] = np.random.randn(layers_dims[l],layers_dims[l-1])*np.sqrt(2/layers_dims[l-1])\n",
        "        parameters['b' + str(l)] = np.zeros((layers_dims[l],1))\n",
        "        \n",
        "        \n",
        "    return parameters"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "7hCZ6Kw4tpZG",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 187
        },
        "outputId": "d882518f-840c-4e19-c8c1-4583078ab12b"
      },
      "cell_type": "code",
      "source": [
        "parameters = initialize_parameters_he([2, 4, 1])\n",
        "print(\"W1 = \" + str(parameters[\"W1\"]))\n",
        "print(\"b1 = \" + str(parameters[\"b1\"]))\n",
        "print(\"W2 = \" + str(parameters[\"W2\"]))\n",
        "print(\"b2 = \" + str(parameters[\"b2\"]))"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "W1 = [[ 1.78862847  0.43650985]\n",
            " [ 0.09649747 -1.8634927 ]\n",
            " [-0.2773882  -0.35475898]\n",
            " [-0.08274148 -0.62700068]]\n",
            "b1 = [[0.]\n",
            " [0.]\n",
            " [0.]\n",
            " [0.]]\n",
            "W2 = [[-0.03098412 -0.33744411 -0.92904268  0.62552248]]\n",
            "b2 = [[0.]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "R5kvp1oktt_s",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 872
        },
        "outputId": "588c55a0-62b9-4310-fd2d-356e3a2d211d"
      },
      "cell_type": "code",
      "source": [
        "parameters = model(train_X, train_Y, initialization = \"he\")\n",
        "print (\"On the train set:\")\n",
        "predictions_train = predict(train_X, train_Y, parameters)\n",
        "print (\"On the test set:\")\n",
        "predictions_test = predict(test_X, test_Y, parameters)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cost after iteration 0: 0.8830537463419761\n",
            "Cost after iteration 1000: 0.6879825919728063\n",
            "Cost after iteration 2000: 0.6751286264523371\n",
            "Cost after iteration 3000: 0.6526117768893805\n",
            "Cost after iteration 4000: 0.6082958970572938\n",
            "Cost after iteration 5000: 0.5304944491717495\n",
            "Cost after iteration 6000: 0.4138645817071794\n",
            "Cost after iteration 7000: 0.3117803464844441\n",
            "Cost after iteration 8000: 0.23696215330322562\n",
            "Cost after iteration 9000: 0.1859728720920684\n",
            "Cost after iteration 10000: 0.15015556280371808\n",
            "Cost after iteration 11000: 0.12325079292273551\n",
            "Cost after iteration 12000: 0.09917746546525937\n",
            "Cost after iteration 13000: 0.08457055954024283\n",
            "Cost after iteration 14000: 0.07357895962677366\n",
            "Cost after iteration 15000: 0.06481018741251149\n",
            "Cost after iteration 16000: 0.057499933022302545\n",
            "Cost after iteration 17000: 0.05193920333459785\n",
            "Cost after iteration 18000: 0.04743514438939104\n",
            "Cost after iteration 19000: 0.04365730226129789\n",
            "Cost after iteration 20000: 0.04021926106254285\n",
            "Cost after iteration 21000: 0.03589621301166325\n",
            "Cost after iteration 22000: 0.03302741192159168\n",
            "Cost after iteration 23000: 0.030898260658900396\n",
            "Cost after iteration 24000: 0.029111460479265286\n",
            "Cost after iteration 25000: 0.027579120292478002\n",
            "Cost after iteration 26000: 0.026246663878183217\n",
            "Cost after iteration 27000: 0.02507797292390439\n",
            "Cost after iteration 28000: 0.02404267615918573\n",
            "Cost after iteration 29000: 0.023119531689312776\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbgAAAEVCAYAAACSSPCDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f680ee72f98>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "On the train set:\n",
            "Accuracy: 0.9966666666666667\n",
            "On the test set:\n",
            "Accuracy: 0.96\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ryyqsG2VuKNP",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "- Different initializations lead to different results\n",
        "- Random initialization is used to break symmetry and make sure different hidden units can learn different things\n",
        "- Don't intialize to values that are too large\n",
        "- He initialization works well for networks with ReLU activations. "
      ]
    },
    {
      "metadata": {
        "id": "GunmZf33t2fp",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}