1-NN-Introduction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "joI9Ck_HEEfQ"
   },
   "source": [
    "# Neural Net : Introduction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "eoM32EDzEEfR"
   },
   "source": [
    "For classifying MNIST digits. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "lgNqpXcGEEfS"
   },
   "source": [
    "#### Intall Tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 661
    },
    "colab_type": "code",
    "id": "Qr6RAnvpFzae",
    "outputId": "1af7f9c1-e13e-44bb-c84a-ba5a2172abe2"
   },
   "outputs": [],
   "source": [
    "#pip install tensorflow==2.0.0-beta0\n",
    "#pip install --upgrade tensorflow==2.0.0-beta0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "d3sEySBYEEfS"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras  # tf.keras\n",
    "import seaborn as sns\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "import sklearn\n",
    "import sys\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The tensorboard extension is already loaded. To reload it, use:\n",
      "  %reload_ext tensorboard\n"
     ]
    }
   ],
   "source": [
    "sns.set()\n",
    "%matplotlib inline\n",
    "%load_ext tensorboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "python 3.7.1 (default, Dec 14 2018, 13:28:58) \n",
      "[Clang 4.0.1 (tags/RELEASE_401/final)]\n",
      "matplotlib 3.0.2\n",
      "numpy 1.15.4\n",
      "pandas 0.23.4\n",
      "sklearn 0.20.1\n",
      "tensorflow 2.0.0-beta0\n",
      "tensorflow.python.keras.api._v2.keras 2.2.4-tf\n"
     ]
    }
   ],
   "source": [
    "print(\"python\", sys.version)\n",
    "for module in mpl, np, pd, sklearn, tf, keras:\n",
    "    print(module.__name__, module.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "colab_type": "code",
    "id": "tEARWdhNEEfW",
    "outputId": "1a1a2b0f-cab5-40b1-8bc0-6e036256e2c6"
   },
   "outputs": [],
   "source": [
    "(X_train, y_train), (X_valid, y_valid) = keras.datasets.mnist.load_data()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### What are the shape of the data?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28) (60000,) (10000, 28, 28) (10000,)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape, y_train.shape,X_valid.shape, y_valid.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- There are 60,000 Train data which are 28 by 28 pixcel images and \n",
    "- There are 10,000 Train data which are also 28 by 28 pixcel images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Data Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 50 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "plt.figure(figsize = [10,8])\n",
    "for i in range(1,51):\n",
    "    plt.subplot(5,10,i)\n",
    "    plt.imshow(X_train[i].reshape(28,28))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 4, 1, 9, 2, 1, 3, 1, 4, 3, 5, 3, 6, 1, 7, 2, 8, 6, 9, 4, 0, 9,\n",
       "       1, 1, 2], dtype=uint8)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[1:26]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "hPqUUAWKEEfY"
   },
   "source": [
    "#### Preprocess data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We flaten the data (from 28 by 28 to 784) before it is feed into the model. We can also create a layer as a front layer to perform this flattening step."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "s5sd9UQeEEfZ"
   },
   "outputs": [],
   "source": [
    "X_train = X_train.reshape(60000, 784).astype('float32')\n",
    "X_valid = X_valid.reshape(10000, 784).astype('float32')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Data Normalization\n",
    "This step converts the pixcel value ranging from 0 to 255 into 0 to 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "A5RfNSUfEEfb"
   },
   "outputs": [],
   "source": [
    "X_train /= 255\n",
    "X_valid /= 255"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Prepare Labels\n",
    "We convert label 1,2,3,4,5,6,7,8,9 in label vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "rgktXQNiEEfc"
   },
   "outputs": [],
   "source": [
    "n_classes = 10\n",
    "y_train = keras.utils.to_categorical(y_train, n_classes)\n",
    "y_valid = keras.utils.to_categorical(y_valid, n_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],\n",
       "       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
       "       [0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 0., 1., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
       "       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 0., 0., 0., 1., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[1:16]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "JAU1Cyy1EEfe"
   },
   "source": [
    "#### Design neural network architecture"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![img](https://datafiction.github.io/docs/ml/perceptron/pic1/perceptron.jpg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Demo: https://datafiction.github.io/docs/ml/perceptron/dlnd-your-first-neural-network/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After the pixels are flattened, the network consists of a sequence of two ```tf.keras.layers.Dense``` layers. These are densely connected, or fully connected, neural layers. The first Dense layer has 64 nodes (or neurons). The second (and last) layer is a 10-node softmax layer that returns an array of 10 probability scores that sum to 1. Each node contains a score that indicates the probability that the current image belongs to one of the 10 classes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "model\n",
    "   = keras.models.sequential([\n",
    "       keras.layer.layerType(parameters...),\n",
    "       keras.layer.layerType(parameters...),\n",
    "       keras.layer.layerType(parameters...),\n",
    "       keras.layer.layerType(parameters...)\n",
    "   ])  \n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GIVzBGIfEEfh"
   },
   "outputs": [],
   "source": [
    "model = keras.models.Sequential([\n",
    "    \n",
    "    \n",
    "    keras.layers.Dense(64,\\\n",
    "                       activation='sigmoid',\\\n",
    "                       input_shape=(784,)),\n",
    "    \n",
    "    keras.layers.Dense(10,\\\n",
    "                       activation='softmax')\n",
    "    \n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 330
    },
    "colab_type": "code",
    "id": "OJIwNRLNEEfj",
    "outputId": "91f81122-7305-4d82-93c3-6942432381a0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_2 (Dense)              (None, 64)                50240     \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                650       \n",
      "=================================================================\n",
      "Total params: 50,890\n",
      "Trainable params: 50,890\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(50240, 650)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(784*64)+ 64, (64*10)+10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "o2hYUMTzEEfp"
   },
   "source": [
    "#### Configure model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before the model is ready for training, it needs a few more settings. These are added during the model's compile step:\n",
    "\n",
    "- ```Loss function``` —This measures how accurate the model is during training. You want to minimize this function to \"steer\" the model in the right direction.\n",
    "- ```Optimizer``` —This is how the model is updated based on the data it sees and its loss function.\n",
    "- ```Metrics``` —Used to monitor the training and testing steps. The following example uses accuracy, the fraction of the images that are correctly classified."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZANJM3AeEEfq"
   },
   "outputs": [],
   "source": [
    "model.compile(loss='categorical_crossentropy',\\\n",
    "              optimizer='adam',\\\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "zq4UBoc5EEfs"
   },
   "source": [
    "#### Train!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now your model is ready to be trained. Call its `fit()` method, passing it the input features (`X_train`) and the target classes (`y_train`). Set `epochs=10` (or else it will just run for a single epoch). You can also (optionally) pass the validation data by setting `validation_data=(X_valid, y_valid)`. If you do, Keras will compute the loss and the additional metrics (the accuracy in this case) on the validation set at the end of each epoch. If the performance on the training set is much better than on the validation set, your model is probably overfitting the training set (or there is a bug, such as a mismatch between the training set and the validation set).\n",
    "**Note**: the `fit()` method will return a `History` object containing training stats. Make sure to preserve it (`history = model.fit(...)`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 471
    },
    "colab_type": "code",
    "id": "ln4J2UMLEEfu",
    "outputId": "9fb80236-f579-45d7-ac9a-39f85be0ee0f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/10\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.1051 - accuracy: 0.9707 - val_loss: 0.1200 - val_accuracy: 0.9641\n",
      "Epoch 2/10\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.0973 - accuracy: 0.9732 - val_loss: 0.1133 - val_accuracy: 0.9649\n",
      "Epoch 3/10\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.0904 - accuracy: 0.9752 - val_loss: 0.1092 - val_accuracy: 0.9672\n",
      "Epoch 4/10\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.0843 - accuracy: 0.9773 - val_loss: 0.1047 - val_accuracy: 0.9689\n",
      "Epoch 5/10\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.0786 - accuracy: 0.9789 - val_loss: 0.1019 - val_accuracy: 0.9680\n",
      "Epoch 6/10\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.0737 - accuracy: 0.9800 - val_loss: 0.0997 - val_accuracy: 0.9697\n",
      "Epoch 7/10\n",
      "60000/60000 [==============================] - 1s 16us/sample - loss: 0.0689 - accuracy: 0.9821 - val_loss: 0.0976 - val_accuracy: 0.9702\n",
      "Epoch 8/10\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0648 - accuracy: 0.9831 - val_loss: 0.0957 - val_accuracy: 0.9706\n",
      "Epoch 9/10\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0609 - accuracy: 0.9845 - val_loss: 0.0930 - val_accuracy: 0.9713\n",
      "Epoch 10/10\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0574 - accuracy: 0.9849 - val_loss: 0.0922 - val_accuracy: 0.9723\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, y_train,\\\n",
    "          batch_size=128,\\\n",
    "          epochs=10,\\\n",
    "          verbose=1,\\\n",
    "          validation_data=(X_valid, y_valid))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Plot History"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Try running `pd.DataFrame(history.history).plot()` to plot the learning curves. To make the graph more readable, you can also set `figsize=(8, 5)`, call `plt.grid(True)` and `plt.gca().set_ylim(0, 1)`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_learning_curves(history):\n",
    "    pd.DataFrame(history.history).plot(figsize=(8, 5))\n",
    "    plt.grid(True)\n",
    "    plt.gca().set_ylim(0, 1)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_learning_curves(history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "yFJm3TV_mrmH"
   },
   "source": [
    "#### Performing Inference"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Call the model's `evaluate()` method, passing it the test set (`X_test` and `y_test`). This will compute the loss (cross-entropy) on the test set, as well as all the additional metrics (in this case, the accuracy). Your model should achieve over 80% accuracy on the test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000/60000 [==============================] - 1s 18us/sample - loss: 0.0519 - accuracy: 0.9872\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.051851865920548634, 0.98723334]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "sosuJg5cmnaU"
   },
   "outputs": [],
   "source": [
    "valid_0 = X_valid[0].reshape(1, 784)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 69
    },
    "colab_type": "code",
    "id": "V2yQfSRBmxIk",
    "outputId": "6f016984-2559-43a8-99e6-7bdb6d73cf9a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2.8813774e-09, 3.8345047e-10, 1.1818415e-04, 2.0551306e-06,\n",
       "        2.2122413e-11, 1.7986340e-09, 1.0651025e-13, 9.9987960e-01,\n",
       "        1.7287803e-08, 1.0190855e-07]], dtype=float32)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(valid_0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "0MOL4f-smzKw",
    "outputId": "41dbb6fa-3fbe-420a-abf7-44c47869d14d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([7])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict_classes(valid_0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "X3wvk6OWm2oz"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1a478e6278>"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(valid_0.reshape(28,28))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define `X_new` as the first 10 instances of the test set. Call the model's `predict()` method to estimate the probability of each class for each instance (for better readability, you may use the output array's `round()` method):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.  , 0.  , 0.  , 0.01, 0.  , 0.99, 0.  , 0.  , 0.  , 0.  ],\n",
       "       [1.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 0.  , 0.01, 0.  , 0.99, 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 1.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.99],\n",
       "       [0.  , 0.  , 1.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 1.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 0.  , 0.  , 1.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 1.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  , 0.  ],\n",
       "       [0.  , 0.  , 0.  , 0.  , 1.  , 0.  , 0.  , 0.  , 0.  , 0.  ]],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_new = 10\n",
    "X_new = X_train[:n_new]\n",
    "y_proba = model.predict(X_new)\n",
    "y_proba.round(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Often, you may only be interested in the most likely class. Use `np.argmax()` to get the class ID of the most likely class for each instance. **Tip**: you want to set `axis=1`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5, 0, 4, 1, 9, 2, 1, 3, 1, 4])"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = y_proba.argmax(axis=1)\n",
    "y_pred"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Call the model's `predict_classes()` method for `X_new`. You should get the same result as above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5, 0, 4, 1, 9, 2, 1, 3, 1, 4])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = model.predict_classes(X_new)\n",
    "y_pred"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(Optional) It is often useful to know how confident the model is for each prediction. Try finding the estimated probability for each predicted class using `np.max()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.99, 1.  , 0.99, 1.  , 0.99, 1.  , 1.  , 1.  , 1.  , 1.  ],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_proba.max(axis=1).round(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(Optional) It is frequent to want the top k classes and their estimated probabilities rather just the most likely class. You can use `np.argsort()` for this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5, 3, 1],\n",
       "       [0, 5, 2],\n",
       "       [4, 2, 7],\n",
       "       [1, 2, 7],\n",
       "       [9, 4, 7],\n",
       "       [2, 8, 9],\n",
       "       [1, 8, 3],\n",
       "       [3, 9, 8],\n",
       "       [1, 9, 7],\n",
       "       [4, 5, 6]])"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "k = 3\n",
    "top_k = np.argsort(-y_proba, axis=1)[:, :k]\n",
    "top_k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.99, 0.01, 0.  ],\n",
       "       [1.  , 0.  , 0.  ],\n",
       "       [0.99, 0.01, 0.  ],\n",
       "       [1.  , 0.  , 0.  ],\n",
       "       [0.99, 0.  , 0.  ],\n",
       "       [1.  , 0.  , 0.  ],\n",
       "       [1.  , 0.  , 0.  ],\n",
       "       [1.  , 0.  , 0.  ],\n",
       "       [1.  , 0.  , 0.  ],\n",
       "       [1.  , 0.  , 0.  ]], dtype=float32)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "row_indices = np.tile(np.arange(len(top_k)), [k, 1]).T\n",
    "y_proba[row_indices, top_k].round(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "TPU",
  "colab": {
   "include_colab_link": true,
   "name": "deep_net.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}