AIfordeeplearning_03.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/plannermango/AI-for-deep-learning/blob/main/AIfordeeplearning_03.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F0CUifyM0NMn"
      },
      "source": [
        "This is a companion notebook for the book [Deep Learning with Python, Second Edition](https://www.manning.com/books/deep-learning-with-python-second-edition?a_aid=keras&a_bid=76564dff). For readability, it only contains runnable code blocks and section titles, and omits everything else in the book: text paragraphs, figures, and pseudocode.\n",
        "\n",
        "**If you want to be able to follow what's going on, I recommend reading the notebook side by side with your copy of the book.**\n",
        "\n",
        "This notebook was generated for TensorFlow 2.6."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QsQHC30L0NMu"
      },
      "source": [
        "# Introduction to Keras and TensorFlow"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9WVo-ibt0NMv"
      },
      "source": [
        "## What's TensorFlow?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IKVpgA_p0NMw"
      },
      "source": [
        "## What's Keras?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BI3VBtDv0NMx"
      },
      "source": [
        "## Keras and TensorFlow: A brief history"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i1yGh19b0NMy"
      },
      "source": [
        "## Setting up a deep-learning workspace"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oKTIuEce0NMy"
      },
      "source": [
        "### Jupyter notebooks: The preferred way to run deep-learning experiments"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jbQQ0lx_0NMz"
      },
      "source": [
        "### Using Colaboratory"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QvMyk1Mi0NM0"
      },
      "source": [
        "#### First steps with Colaboratory"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xbzSHvpT0NM0"
      },
      "source": [
        "#### Installing packages with pip"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uxD2_Inr0NM1"
      },
      "source": [
        "#### Using the GPU runtime"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nkXXH58W0NM1"
      },
      "source": [
        "## First steps with TensorFlow"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6XRS0ZP60NM2"
      },
      "source": [
        "#### Constant tensors and variables"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3cKgJqMY0NM2"
      },
      "source": [
        "**All-ones or all-zeros tensors**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "L4uoAaZi0NM3",
        "outputId": "f4430fcc-8573-459f-ae86-cf6dd3d0ce64",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tf.Tensor(\n",
            "[[1.]\n",
            " [1.]], shape=(2, 1), dtype=float32)\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "x = tf.ones(shape=(2, 1))\n",
        "print(x)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "xfb057A00NM4",
        "outputId": "b2cb3e59-fe6e-47a0-9770-1509d5224ffc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tf.Tensor(\n",
            "[[0.]\n",
            " [0.]], shape=(2, 1), dtype=float32)\n"
          ]
        }
      ],
      "source": [
        "x = tf.zeros(shape=(2, 1))\n",
        "print(x)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r5CO8tco0NM4"
      },
      "source": [
        "**Random tensors**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "xzbGVlGf0NM4",
        "outputId": "8ea7d6d2-4c11-4767-bb55-39d53ff0f0b4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tf.Tensor(\n",
            "[[-0.564222  ]\n",
            " [-0.14145334]\n",
            " [ 2.3589604 ]], shape=(3, 1), dtype=float32)\n"
          ]
        }
      ],
      "source": [
        "x = tf.random.normal(shape=(3, 1), mean=0., stddev=1.)\n",
        "print(x)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "jPpUb03C0NM5",
        "outputId": "9a8ffc35-a422-4259-e94b-ee3141feb97b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tf.Tensor(\n",
            "[[0.26191115]\n",
            " [0.14196444]\n",
            " [0.04371297]], shape=(3, 1), dtype=float32)\n"
          ]
        }
      ],
      "source": [
        "x = tf.random.uniform(shape=(3, 1), minval=0., maxval=1.)\n",
        "print(x)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hONlX7G0NM5"
      },
      "source": [
        "**NumPy arrays are assignable**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "LrKhPcsG0NM5"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "x = np.ones(shape=(2, 2))\n",
        "x[0, 0] = 0."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gzuPz5lp0NM6"
      },
      "source": [
        "**Creating a TensorFlow variable**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "2kM0FKsQ0NM6",
        "outputId": "d1028fb2-aa08-4a75-b7f3-cdda98a629bd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<tf.Variable 'Variable:0' shape=(3, 1) dtype=float32, numpy=\n",
            "array([[-0.24071167],\n",
            "       [-0.15315333],\n",
            "       [ 0.31576157]], dtype=float32)>\n"
          ]
        }
      ],
      "source": [
        "v = tf.Variable(initial_value=tf.random.normal(shape=(3, 1)))\n",
        "print(v)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lEfO_Ppm0NM9"
      },
      "source": [
        "**Assigning a value to a TensorFlow variable**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "MQcy-AM90NM-",
        "outputId": "1ef1a13f-2ca1-44c6-bfb9-e41733e647ba",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Variable 'UnreadVariable' shape=(3, 1) dtype=float32, numpy=\n",
              "array([[1.],\n",
              "       [1.],\n",
              "       [1.]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ],
      "source": [
        "v.assign(tf.ones((3, 1)))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WjQn2Wr30NM_"
      },
      "source": [
        "**Assigning a value to a subset of a TensorFlow variable**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "qJr85FzQ0NM_",
        "outputId": "e1239500-bbca-453f-8b15-c730206fa923",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Variable 'UnreadVariable' shape=(3, 1) dtype=float32, numpy=\n",
              "array([[3.],\n",
              "       [1.],\n",
              "       [1.]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ],
      "source": [
        "v[0, 0].assign(3.)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7-OthSXm0NM_"
      },
      "source": [
        "**Using `assign_add`**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "TP5hDU2s0NM_",
        "outputId": "0f110ab9-8639-4dce-a229-cd9b02ab5e44",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Variable 'UnreadVariable' shape=(3, 1) dtype=float32, numpy=\n",
              "array([[4.],\n",
              "       [2.],\n",
              "       [2.]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ],
      "source": [
        "v.assign_add(tf.ones((3, 1)))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XiHqpZ9u0NNA"
      },
      "source": [
        "#### Tensor operations: Doing math in TensorFlow"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7KiC0I8Y0NNA"
      },
      "source": [
        "**A few basic math operations**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "vo22C92G0NNA"
      },
      "outputs": [],
      "source": [
        "a = tf.ones((2, 2))\n",
        "b = tf.square(a)\n",
        "c = tf.sqrt(a)\n",
        "d = b + c\n",
        "e = tf.matmul(a, b)\n",
        "e *= d"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Zv4RIE_F0NNA"
      },
      "source": [
        "#### A second look at the GradientTape API"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fVV_aXpd0NNB"
      },
      "source": [
        "**Using the `GradientTape`**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "l9d9LXL_0NNB"
      },
      "outputs": [],
      "source": [
        "input_var = tf.Variable(initial_value=3.)\n",
        "with tf.GradientTape() as tape:\n",
        "   result = tf.square(input_var)\n",
        "gradient = tape.gradient(result, input_var)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I4oko2Cd0NNB"
      },
      "source": [
        "**Using `GradientTape` with constant tensor inputs**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "bDVjwxpc0NNB"
      },
      "outputs": [],
      "source": [
        "input_const = tf.constant(3.)\n",
        "with tf.GradientTape() as tape:\n",
        "   tape.watch(input_const)\n",
        "   result = tf.square(input_const)\n",
        "gradient = tape.gradient(result, input_const)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hkXsfM3W0NNC"
      },
      "source": [
        "**Using nested gradient tapes to compute second-order gradients**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "7jsDj9Za0NNC"
      },
      "outputs": [],
      "source": [
        "time = tf.Variable(0.)\n",
        "with tf.GradientTape() as outer_tape:\n",
        "    with tf.GradientTape() as inner_tape:\n",
        "        position =  4.9 * time ** 2\n",
        "    speed = inner_tape.gradient(position, time)\n",
        "acceleration = outer_tape.gradient(speed, time)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W4geguL60NNC"
      },
      "source": [
        "#### An end-to-end example: A linear classifier in pure TensorFlow"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AChL1uJc0NNC"
      },
      "source": [
        "**Generating two classes of random points in a 2D plane**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "9q3E6WYo0NNC"
      },
      "outputs": [],
      "source": [
        "num_samples_per_class = 1000\n",
        "negative_samples = np.random.multivariate_normal(\n",
        "    mean=[0, 3],\n",
        "    cov=[[1, 0.5],[0.5, 1]],\n",
        "    size=num_samples_per_class)\n",
        "positive_samples = np.random.multivariate_normal(\n",
        "    mean=[3, 0],\n",
        "    cov=[[1, 0.5],[0.5, 1]],\n",
        "    size=num_samples_per_class)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jPei1b5c0NND"
      },
      "source": [
        "**Stacking the two classes into an array with shape (2000, 2)**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "dZcHAdXY0NND"
      },
      "outputs": [],
      "source": [
        "inputs = np.vstack((negative_samples, positive_samples)).astype(np.float32)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tOmRa9xM0NND"
      },
      "source": [
        "**Generating the corresponding targets (0 and 1)**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "iVfzVy800NND"
      },
      "outputs": [],
      "source": [
        "targets = np.vstack((np.zeros((num_samples_per_class, 1), dtype=\"float32\"),\n",
        "                     np.ones((num_samples_per_class, 1), dtype=\"float32\")))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "54eMh6I70NNE"
      },
      "source": [
        "**Plotting the two point classes**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "XUVnL78j0NNE",
        "outputId": "ec1e0c9d-25b0-45aa-9db1-a7b310635984",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.scatter(inputs[:, 0], inputs[:, 1], c=targets[:, 0])\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TcO4rRmB0NNE"
      },
      "source": [
        "**Creating the linear classifier variables**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "n8GETXhS0NNE"
      },
      "outputs": [],
      "source": [
        "input_dim = 2\n",
        "output_dim = 1\n",
        "W = tf.Variable(initial_value=tf.random.uniform(shape=(input_dim, output_dim)))\n",
        "b = tf.Variable(initial_value=tf.zeros(shape=(output_dim,)))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3HzEOhUV0NNE"
      },
      "source": [
        "**The forward pass function**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "Jf_iISi60NNE"
      },
      "outputs": [],
      "source": [
        "def model(inputs):\n",
        "    return tf.matmul(inputs, W) + b"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fkdE2p7I0NNF"
      },
      "source": [
        "**The mean squared error loss function**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "PUqy5-170NNF"
      },
      "outputs": [],
      "source": [
        "def square_loss(targets, predictions):\n",
        "    per_sample_losses = tf.square(targets - predictions)\n",
        "    return tf.reduce_mean(per_sample_losses)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2SZAOKmw0NNF"
      },
      "source": [
        "**The training step function**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "4AqyLJhA0NNI"
      },
      "outputs": [],
      "source": [
        "learning_rate = 0.1\n",
        "\n",
        "def training_step(inputs, targets):\n",
        "    with tf.GradientTape() as tape:\n",
        "        predictions = model(inputs)\n",
        "        loss = square_loss(targets, predictions)\n",
        "    grad_loss_wrt_W, grad_loss_wrt_b = tape.gradient(loss, [W, b])\n",
        "    W.assign_sub(grad_loss_wrt_W * learning_rate)\n",
        "    b.assign_sub(grad_loss_wrt_b * learning_rate)\n",
        "    return loss"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iRNN5Zsa0NNJ"
      },
      "source": [
        "**The batch training loop**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "Fapl5pjP0NNJ",
        "outputId": "344a1c47-ee50-4b45-efe6-ea2794a774a2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Loss at step 0: 0.8518\n",
            "Loss at step 1: 0.1870\n",
            "Loss at step 2: 0.1125\n",
            "Loss at step 3: 0.0970\n",
            "Loss at step 4: 0.0895\n",
            "Loss at step 5: 0.0835\n",
            "Loss at step 6: 0.0782\n",
            "Loss at step 7: 0.0733\n",
            "Loss at step 8: 0.0689\n",
            "Loss at step 9: 0.0649\n",
            "Loss at step 10: 0.0613\n",
            "Loss at step 11: 0.0580\n",
            "Loss at step 12: 0.0549\n",
            "Loss at step 13: 0.0522\n",
            "Loss at step 14: 0.0497\n",
            "Loss at step 15: 0.0475\n",
            "Loss at step 16: 0.0454\n",
            "Loss at step 17: 0.0435\n",
            "Loss at step 18: 0.0418\n",
            "Loss at step 19: 0.0403\n",
            "Loss at step 20: 0.0389\n",
            "Loss at step 21: 0.0376\n",
            "Loss at step 22: 0.0364\n",
            "Loss at step 23: 0.0353\n",
            "Loss at step 24: 0.0344\n",
            "Loss at step 25: 0.0335\n",
            "Loss at step 26: 0.0327\n",
            "Loss at step 27: 0.0320\n",
            "Loss at step 28: 0.0313\n",
            "Loss at step 29: 0.0307\n",
            "Loss at step 30: 0.0302\n",
            "Loss at step 31: 0.0297\n",
            "Loss at step 32: 0.0292\n",
            "Loss at step 33: 0.0288\n",
            "Loss at step 34: 0.0284\n",
            "Loss at step 35: 0.0281\n",
            "Loss at step 36: 0.0278\n",
            "Loss at step 37: 0.0275\n",
            "Loss at step 38: 0.0273\n",
            "Loss at step 39: 0.0270\n"
          ]
        }
      ],
      "source": [
        "for step in range(40):\n",
        "    loss = training_step(inputs, targets)\n",
        "    print(f\"Loss at step {step}: {loss:.4f}\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "id": "FLXioBq60NNJ",
        "outputId": "4d976594-3f4b-4b0e-db5f-624658c0f2e7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "predictions = model(inputs)\n",
        "plt.scatter(inputs[:, 0], inputs[:, 1], c=predictions[:, 0] > 0.5)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "id": "PPSygT9R0NNJ",
        "outputId": "543d9028-b527-423c-81cb-ce7cf448f86a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7fdbcb5ed850>"
            ]
          },
          "metadata": {},
          "execution_count": 24
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "x = np.linspace(-1, 4, 100)\n",
        "y = - W[0] /  W[1] * x + (0.5 - b) / W[1]\n",
        "plt.plot(x, y, \"-r\")\n",
        "plt.scatter(inputs[:, 0], inputs[:, 1], c=predictions[:, 0] > 0.5)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "82jYfUnJ0NNJ"
      },
      "source": [
        "## Anatomy of a neural network: Understanding core Keras APIs"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YIj_PoHi0NNK"
      },
      "source": [
        "### Layers: The building blocks of deep learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VbhpeMgP0NNK"
      },
      "source": [
        "#### The base Layer class in Keras"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wh8yKtYH0NNK"
      },
      "source": [
        "**A `Dense` layer implemented as a `Layer` subclass**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "id": "S_b3aywY0NNK"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "\n",
        "class SimpleDense(keras.layers.Layer):\n",
        "\n",
        "    def __init__(self, units, activation=None):\n",
        "        super().__init__()\n",
        "        self.units = units\n",
        "        self.activation = activation\n",
        "\n",
        "    def build(self, input_shape):\n",
        "        input_dim = input_shape[-1]\n",
        "        self.W = self.add_weight(shape=(input_dim, self.units),\n",
        "                                 initializer=\"random_normal\")\n",
        "        self.b = self.add_weight(shape=(self.units,),\n",
        "                                 initializer=\"zeros\")\n",
        "\n",
        "    def call(self, inputs):\n",
        "        y = tf.matmul(inputs, self.W) + self.b\n",
        "        if self.activation is not None:\n",
        "            y = self.activation(y)\n",
        "        return y"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "JgqBB_8w0NNK",
        "outputId": "8b51a7f0-0704-4d99-81ea-ea65f418ff85",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(2, 32)\n"
          ]
        }
      ],
      "source": [
        "my_dense = SimpleDense(units=32, activation=tf.nn.relu)\n",
        "input_tensor = tf.ones(shape=(2, 784))\n",
        "output_tensor = my_dense(input_tensor)\n",
        "print(output_tensor.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K3uwsU5s0NNL"
      },
      "source": [
        "#### Automatic shape inference: Building layers on the fly"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "0ZA2emec0NNL"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras import layers\n",
        "layer = layers.Dense(32, activation=\"relu\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "id": "Yieup32n0NNL"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras import models\n",
        "from tensorflow.keras import layers\n",
        "model = models.Sequential([\n",
        "    layers.Dense(32, activation=\"relu\"),\n",
        "    layers.Dense(32)\n",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "id": "CAIPFsMK0NNL"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    SimpleDense(32, activation=\"relu\"),\n",
        "    SimpleDense(64, activation=\"relu\"),\n",
        "    SimpleDense(32, activation=\"relu\"),\n",
        "    SimpleDense(10, activation=\"softmax\")\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HRIrprh40NNM"
      },
      "source": [
        "### From layers to models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "snNldC7J0NNN"
      },
      "source": [
        "### The \"compile\" step: Configuring the learning process"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "nLNtCVJl0NNN"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([keras.layers.Dense(1)])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"mean_squared_error\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "ZjGQGZCU0NNN"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=keras.optimizers.RMSprop(),\n",
        "              loss=keras.losses.MeanSquaredError(),\n",
        "              metrics=[keras.metrics.BinaryAccuracy()])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HfUswfYJ0NNN"
      },
      "source": [
        "### Picking a loss function"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ko1rh-vR0NNO"
      },
      "source": [
        "### Understanding the fit() method"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ke6NbzC90NNO"
      },
      "source": [
        "**Calling `fit()` with NumPy data**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "y2h8rTaS0NNO",
        "outputId": "9f4d21a1-98d1-42d8-e30f-287070ed5600",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "16/16 [==============================] - 0s 2ms/step - loss: 8.6786 - binary_accuracy: 0.3555\n",
            "Epoch 2/5\n",
            "16/16 [==============================] - 0s 2ms/step - loss: 8.3127 - binary_accuracy: 0.3475\n",
            "Epoch 3/5\n",
            "16/16 [==============================] - 0s 2ms/step - loss: 8.0077 - binary_accuracy: 0.3420\n",
            "Epoch 4/5\n",
            "16/16 [==============================] - 0s 2ms/step - loss: 7.7173 - binary_accuracy: 0.3345\n",
            "Epoch 5/5\n",
            "16/16 [==============================] - 0s 2ms/step - loss: 7.4369 - binary_accuracy: 0.3285\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(\n",
        "    inputs,\n",
        "    targets,\n",
        "    epochs=5,\n",
        "    batch_size=128\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "id": "qf9yIMLu0NNO",
        "outputId": "88bf4f39-04b5-40ef-eded-b7967a0d8b59",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'binary_accuracy': [0.3555000126361847,\n",
              "  0.3474999964237213,\n",
              "  0.34200000762939453,\n",
              "  0.3345000147819519,\n",
              "  0.32850000262260437],\n",
              " 'loss': [8.67861557006836,\n",
              "  8.312700271606445,\n",
              "  8.007737159729004,\n",
              "  7.717260837554932,\n",
              "  7.436948776245117]}"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ],
      "source": [
        "history.history"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V99OpEQb0NNO"
      },
      "source": [
        "### Monitoring loss and metrics on validation data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NFvPq_fY0NNO"
      },
      "source": [
        "**Using the `validation_data` argument**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "id": "z0rGAFo80NNO",
        "outputId": "2ffe5fa4-7707-41b2-c2da-d02f945cb1dd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "88/88 [==============================] - 1s 4ms/step - loss: 0.2607 - binary_accuracy: 0.9636 - val_loss: 0.0384 - val_binary_accuracy: 0.9950\n",
            "Epoch 2/5\n",
            "88/88 [==============================] - 0s 4ms/step - loss: 0.0758 - binary_accuracy: 0.9536 - val_loss: 0.0787 - val_binary_accuracy: 0.9767\n",
            "Epoch 3/5\n",
            "88/88 [==============================] - 0s 4ms/step - loss: 0.0725 - binary_accuracy: 0.9629 - val_loss: 0.2562 - val_binary_accuracy: 0.7000\n",
            "Epoch 4/5\n",
            "88/88 [==============================] - 0s 5ms/step - loss: 0.0734 - binary_accuracy: 0.9557 - val_loss: 0.0326 - val_binary_accuracy: 0.9967\n",
            "Epoch 5/5\n",
            "88/88 [==============================] - 0s 4ms/step - loss: 0.0722 - binary_accuracy: 0.9564 - val_loss: 0.0401 - val_binary_accuracy: 1.0000\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fdb4376e2d0>"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ],
      "source": [
        "model = keras.Sequential([keras.layers.Dense(1)])\n",
        "model.compile(optimizer=keras.optimizers.RMSprop(learning_rate=0.1),\n",
        "              loss=keras.losses.MeanSquaredError(),\n",
        "              metrics=[keras.metrics.BinaryAccuracy()])\n",
        "\n",
        "indices_permutation = np.random.permutation(len(inputs))\n",
        "shuffled_inputs = inputs[indices_permutation]\n",
        "shuffled_targets = targets[indices_permutation]\n",
        "\n",
        "num_validation_samples = int(0.3 * len(inputs))\n",
        "val_inputs = shuffled_inputs[:num_validation_samples]\n",
        "val_targets = shuffled_targets[:num_validation_samples]\n",
        "training_inputs = shuffled_inputs[num_validation_samples:]\n",
        "training_targets = shuffled_targets[num_validation_samples:]\n",
        "model.fit(\n",
        "    training_inputs,\n",
        "    training_targets,\n",
        "    epochs=5,\n",
        "    batch_size=16,\n",
        "    validation_data=(val_inputs, val_targets)\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "65alla3x0NNP"
      },
      "source": [
        "### Inference: Using a model after training"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "metadata": {
        "id": "HqWNtORn0NNP",
        "outputId": "81d901b5-47be-4dea-f02d-1e3b2aadccd5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.0646559 ]\n",
            " [0.91452205]\n",
            " [0.3773205 ]\n",
            " [0.7363893 ]\n",
            " [0.18536568]\n",
            " [0.22297367]\n",
            " [0.666006  ]\n",
            " [0.81582135]\n",
            " [0.9631009 ]\n",
            " [0.18809557]]\n"
          ]
        }
      ],
      "source": [
        "predictions = model.predict(val_inputs, batch_size=128)\n",
        "print(predictions[:10])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w3EALP_u0NNP"
      },
      "source": [
        "## Summary"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "AIfordeeplearning_03",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "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": "3.7.0"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}