working_with_custom_loss_function_573.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.datasets import fetch_california_housing\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "housing = fetch_california_housing()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_full, X_test, y_train_full, y_test = train_test_split( housing.data,\n",
    "         housing.target.reshape(-1, 1), random_state=42)\n",
    "\n",
    "X_train, X_valid, y_train, y_valid =  train_test_split( X_train_full, \n",
    "        y_train_full, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_scaled = scaler.fit_transform(X_train)\n",
    "X_valid_scaled = scaler.transform(X_valid)\n",
    "X_test_scaled = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Custom Loss Function - Huber Loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def huber_fn(y_true, y_pred):\n",
    "    error = y_true - y_pred\n",
    "    is_small_error = tf.abs(error) < 1\n",
    "    squared_loss = tf.square(error) / 2\n",
    "    linear_loss = tf.abs(error) - 0.5\n",
    "    return tf.where(is_small_error, squared_loss, linear_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = keras.models.Sequential([\n",
    "    keras.layers.Dense(30, activation= \"selu\", \n",
    "       kernel_initializer= \"lecun_normal\", \n",
    "       input_shape=X_train.shape[1:]),\n",
    "    keras.layers.Dense(1),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 30)                270       \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 31        \n",
      "=================================================================\n",
      "Total params: 301\n",
      "Trainable params: 301\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.compile(loss=huber_fn, optimizer=\"nadam\", metrics=[\"mae\"])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(11610, 8) (11610, 1) (3870, 8) (3870, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_train_scaled.shape, y_train.shape, X_valid_scaled.shape, y_valid.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.5696 - mae: 0.9316 - val_loss: 0.2076 - val_mae: 0.4994\n",
      "Epoch 2/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.2081 - mae: 0.5026 - val_loss: 0.1944 - val_mae: 0.4767\n",
      "Epoch 3/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1995 - mae: 0.4886 - val_loss: 0.2197 - val_mae: 0.5022\n",
      "Epoch 4/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1962 - mae: 0.4830 - val_loss: 0.1990 - val_mae: 0.4844\n",
      "Epoch 5/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1922 - mae: 0.4776 - val_loss: 0.1930 - val_mae: 0.4724\n",
      "Epoch 6/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1895 - mae: 0.4737 - val_loss: 0.1822 - val_mae: 0.4591\n",
      "Epoch 7/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1867 - mae: 0.4690 - val_loss: 0.1739 - val_mae: 0.4523\n",
      "Epoch 8/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1841 - mae: 0.4648 - val_loss: 0.1738 - val_mae: 0.4503\n",
      "Epoch 9/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1826 - mae: 0.4627 - val_loss: 0.1860 - val_mae: 0.4630\n",
      "Epoch 10/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1808 - mae: 0.4596 - val_loss: 0.1765 - val_mae: 0.4522\n",
      "Epoch 11/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1794 - mae: 0.4568 - val_loss: 0.1789 - val_mae: 0.4508\n",
      "Epoch 12/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1778 - mae: 0.4547 - val_loss: 0.1688 - val_mae: 0.4469\n",
      "Epoch 13/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1768 - mae: 0.4534 - val_loss: 0.1828 - val_mae: 0.4539\n",
      "Epoch 14/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1762 - mae: 0.4517 - val_loss: 0.1677 - val_mae: 0.4426\n",
      "Epoch 15/15\n",
      "363/363 [==============================] - 1s 2ms/step - loss: 0.1748 - mae: 0.4501 - val_loss: 0.1754 - val_mae: 0.4465\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f87691b5ac8>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train_scaled, y_train, epochs=15,\n",
    "    validation_data=(X_valid_scaled, y_valid))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluate Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "162/162 [==============================] - 0s 1ms/step - loss: 0.1762 - mae: 0.4521\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.17624062299728394, 0.45205244421958923]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_test_scaled, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tensorflow2",
   "language": "python",
   "name": "tensorflow2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}