{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Optimization with Masterful\n",
    "\n",
    "**Author:** [Nikhil Gajendrakumar](mailto:nikhil@masterfulai.com)  \n",
    "**Date created:** 2022/04/27  \n",
    "**Last modified:** 2022/05/04  \n",
    "**Description:** Optimization with Masterful\n",
    "\n",
    "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)][1]        [![Download](images/download.png)][2][Download this Notebook][2]\n",
    "\n",
    "[1]:https://colab.research.google.com/github/masterfulai/masterful-docs/blob/main/notebooks/guide_batch_lr_finder.ipynb\n",
    "[2]:http://docs.masterfulai.com/0.4.1/notebooks/guide_batch_lr_finder.ipynb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is an introduction to Masterful's automatic solution to tune algorithmic hyperparameters. Hyperparameter such as learning rate and batch size control the training process of a model. Given these hyperparameters, the training algorithm learns the model parameters from the data. Picking the right values for these hyperparameters can be very painful. You take this ardous journey of trial-and-error where you try different values, run the training for few epochs, and pick the values that achieve the best validation accuracy. This process is very tedious and if you modify the model architecture, you will have to repeat the entire process again.\n",
    "\n",
    "This guide will demonstrate how Masterful, algorithmically, picks the right values for these hyperparameters, for the given model and dataset, that can achieve the best performance in shortest training time."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prerequisites\n",
    "\n",
    "Please follow the Masterful installation instructions [here](../tutorials/tutorial_installation.md)\n",
    "in order to run this Quickstart."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports\n",
    "\n",
    "First, import the necessary libraries and register the Masterful package."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MASTERFUL: Your account has been successfully registered. Masterful v0.4.1 is loaded.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import tensorflow_addons as tfa\n",
    "import tensorflow_datasets as tfds\n",
    "import matplotlib.pyplot as plt\n",
    "from time import time\n",
    "import masterful\n",
    "masterful = masterful.register()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare the Data\n",
    "\n",
    "For this guide, you will use SVHN (Street View House Numbers) dataset which is an image digit recognition dataset of digit images coming from real world data. Images are cropped to 32x32. The dataset consists of 73,257 images for training and 26,032 for testing.\n",
    "\n",
    "You can verify that the data was downloaded correctly by displaying a few examples from the dataset. The first step in any ML project is some exploratory data analysis (EDA). The bare minimum is to visualize the images and understand their contents."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_dataset = tfds.load(\"svhn_cropped\",\n",
    "                          split=\"train[:5%]\",\n",
    "                          as_supervised=True)\n",
    "test_dataset = tfds.load(\"svhn_cropped\",\n",
    "                         split=\"test[:5%]\",\n",
    "                         as_supervised=True)\n",
    "\n",
    "f, axarr = plt.subplots(4, 4, figsize=(10,10))\n",
    "\n",
    "for i,(x,y) in enumerate(train_dataset.take(4*4)):\n",
    "    row = i // 4\n",
    "    col = i % 4\n",
    "    axarr[row, col].imshow(x.numpy())\n",
    "    axarr[row, col].title.set_text(f'{y.numpy()}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# set image range to [0-1]\n",
    "train_dataset = train_dataset.map(\n",
    "    lambda image, label:\n",
    "    (tf.cast(image, tf.float32) / 255.0, label),\n",
    "    num_parallel_calls=tf.data.AUTOTUNE\n",
    ")\n",
    "\n",
    "test_dataset = test_dataset.map(\n",
    "    lambda image, label:\n",
    "    (tf.cast(image, tf.float32) / 255.0, label),\n",
    "    num_parallel_calls=tf.data.AUTOTUNE\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the Model\n",
    "For this example, we will use a simple model inspired by simple CNN used in\n",
    "[this tensorflow tutorial](https://www.tensorflow.org/tutorials/images/cnn). This is a toy model for demonstration purposes only, and should not be used in a production environment. It has a few convolutional layers and outputs logits directly, rather than using a softmax layer at the end."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "NUM_CLASSES = 10\n",
    "\n",
    "def get_model():\n",
    "\n",
    "  model = tf.keras.models.Sequential([\n",
    "      tf.keras.layers.Input(shape=(32, 32, 3)),\n",
    "      tf.keras.layers.Conv2D(64, 3, padding='same', activation='relu'),\n",
    "      tf.keras.layers.MaxPooling2D(),\n",
    "      tf.keras.layers.Conv2D(128, 3, padding='same', activation='relu'),\n",
    "      tf.keras.layers.MaxPooling2D(),\n",
    "      tf.keras.layers.Conv2D(256, 3, padding='same', activation='relu'),\n",
    "      tf.keras.layers.MaxPooling2D(),\n",
    "      tf.keras.layers.Conv2D(512, 3, padding='same', activation='relu'),\n",
    "      tf.keras.layers.MaxPooling2D(),\n",
    "      tf.keras.layers.Conv2D(1024, 3, padding='same', activation='relu'),\n",
    "      tf.keras.layers.MaxPooling2D(),\n",
    "      tf.keras.layers.Flatten(),\n",
    "      tf.keras.layers.Dense(256, activation='relu'),\n",
    "      tf.keras.layers.Dense(128, activation='relu'),\n",
    "      tf.keras.layers.Dense(NUM_CLASSES)\n",
    "  ])\n",
    "\n",
    "  return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Baseline Training\n",
    "\n",
    "In order to measure the performance improvements from Masterful,\n",
    "you should measure the performance of your model after training\n",
    "with a standard training loop.\n",
    "\n",
    "The hyperparameter values learning rate = 0.001 and batch size = 32\n",
    "were picked since they are the most popular starting values that\n",
    "have produced reasonable results across different model architectures\n",
    "and datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_lr_metric(optimizer):\n",
    "    def lr(y_true, y_pred):\n",
    "        return optimizer.lr\n",
    "    return lr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# setup callback to log the total training time\n",
    "class TrainingTime(tf.keras.callbacks.Callback):\n",
    "    def __init__(self, batch_size, lr, logs=None):\n",
    "        self.time_per_epoch = []\n",
    "        self.batch_size = batch_size\n",
    "        self.learning_rate = lr\n",
    "\n",
    "    def on_epoch_begin(self, epoch, logs=None):\n",
    "        self.starttime = time()\n",
    "\n",
    "    def on_epoch_end(self, epoch, logs=None):\n",
    "        self.time_per_epoch.append(time() - self.starttime)\n",
    "        \n",
    "    def on_train_end(self, logs=None):\n",
    "        total_train_time = sum(self.time_per_epoch)\n",
    "        msg = (f\"Training time with batch size = {self.batch_size} and \"\n",
    "               f\"learning rate = {self.learning_rate} is {total_train_time:.2f} s.\")\n",
    "        print(msg)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# set few common parameters for both experiments\n",
    "\n",
    "epochs = 100\n",
    "shuffle_buffer_size = 1024\n",
    "optimizer = tfa.optimizers.LAMB(learning_rate=0.001)\n",
    "lr_metric = get_lr_metric(optimizer)\n",
    "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(monitor='val_loss',\n",
    "                                                 factor=0.5,\n",
    "                                                 patience=8)\n",
    "early_stop = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training time with batch size = 32 and learning rate = 0.001 is 122.04 s.\n",
      "Baseline model Train accuracy: 0.991264, and Test accuracy: 0.791091\n"
     ]
    }
   ],
   "source": [
    "batch_size = 32\n",
    "learning_rate = 0.001\n",
    "\n",
    "tf.keras.backend.set_value(optimizer.lr, learning_rate)\n",
    "\n",
    "base_model = get_model()\n",
    "base_model.compile(optimizer=optimizer,\n",
    "                   loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "                   metrics=['accuracy', lr_metric])\n",
    "\n",
    "\n",
    "base_history = base_model.fit(train_dataset.shuffle(shuffle_buffer_size)\n",
    "                              .batch(batch_size),\n",
    "                              epochs=epochs,\n",
    "                              validation_data=test_dataset.batch(64),\n",
    "                              callbacks=[\n",
    "                                  reduce_lr,\n",
    "                                  early_stop,\n",
    "                                  TrainingTime(batch_size,learning_rate)\n",
    "                              ],\n",
    "                              verbose=0)\n",
    "\n",
    "msg = (f\"Baseline model Train accuracy: {max(base_history.history['accuracy']):4f}, \"\n",
    "       f\"and Test accuracy: {max(base_history.history['val_accuracy']):4f}\")\n",
    "print(msg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Automatic hyperparameter value selection\n",
    "\n",
    "The Masterful AutoML platform learns how to train your model by\n",
    "focusing on five core organizational principles in deep\n",
    "learning: architecture, data, optimization, regularization,\n",
    "and semi-supervision.\n",
    "\n",
    "In this guide we are only demostrating Masterful's optimizer and automatic\n",
    "hyperparameter value selection, which is part of our optimization bucket,\n",
    "and the performance benefit we acheive from the right hyperparameter values\n",
    "alone in terms of training time and accuracy. Therefore we only run\n",
    "Masterful apis that  are necessary to get the hyperparameter values and\n",
    "train the model with a standard tensorflow training loop.\n",
    "\n",
    "Masterful's meta-learner for optimization hyperparameters are tailored\n",
    "to your model architecture and data. So the first thing to create\n",
    "[ArchitectureParams](../api/api_architecture.rst#masterful.architecture.ArchitectureParams)\n",
    "and [DataParams](../api/api_data.rst#masterful.data.DataParams) via their\n",
    "respective learner functions. \n",
    "\n",
    "In the code below, you are telling Masterful that your model is \n",
    "performing a classification task (`masterful.enums.Task.CLASSIFICATION`) \n",
    "with 10 labels (`num_classes=NUM_CLASSES`), and that the \n",
    "input range of the image features going into your model are \n",
    "in the range [0,1] (`input_range=masterful.enums.ImageRange.ZERO_ONE`). \n",
    "Also, the model outputs logits rather than a softmax classification\n",
    " (`prediction_logits=True`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Start fresh with a new model\n",
    "tf.keras.backend.clear_session()\n",
    "model = get_model()\n",
    "\n",
    "model_params = masterful.architecture.learn_architecture_params(\n",
    "    model=model,\n",
    "    task=masterful.enums.Task.CLASSIFICATION,\n",
    "    input_range=masterful.enums.ImageRange.ZERO_ONE,\n",
    "    num_classes=NUM_CLASSES,\n",
    "    prediction_logits=True,\n",
    ")\n",
    "training_dataset_params = masterful.data.learn_data_params(\n",
    "    dataset=train_dataset,\n",
    "    task=masterful.enums.Task.CLASSIFICATION,\n",
    "    image_range=masterful.enums.ImageRange.ZERO_ONE,\n",
    "    num_classes=NUM_CLASSES,\n",
    "    sparse_labels=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Meta-Learning Optimization Hyperparameters\n",
    "\n",
    "Now, call the learner for optimization hyperparameters. \n",
    "\n",
    "For more details on the optmization parameters, please see the [OptimizationParams](../api/api_optimization.rst#masterful.optimization.OptimizationParams) API specification.\n",
    "\n",
    "To try our platform to train the model end to end, refer to our [Quick start guide](https://masterful-public.s3.us-west-1.amazonaws.com/933013963/0.4.1/notebooks/tutorial_quickstart.html#)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MASTERFUL: Learning optimal batch size.\n",
      "MASTERFUL: Learning optimal initial learning rate for batch size 256.\n"
     ]
    }
   ],
   "source": [
    "optimization_params = masterful.optimization.learn_optimization_params(\n",
    "    model,\n",
    "    model_params,\n",
    "    train_dataset,\n",
    "    training_dataset_params,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training time with batch size = 256 and learning rate = 0.0035355337895452976 is 64.27 s.\n",
      "Masterful model Train accuracy: 0.998635, and Test accuracy: 0.803379\n"
     ]
    }
   ],
   "source": [
    "batch_size = optimization_params.batch_size\n",
    "learning_rate = optimization_params.learning_rate\n",
    "\n",
    "# set starting optimizer learning rate\n",
    "tf.keras.backend.set_value(optimizer.lr, learning_rate)\n",
    "lr_metric = get_lr_metric(optimizer)\n",
    "\n",
    "model.compile(optimizer=optimizer,\n",
    "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "              metrics=['accuracy', lr_metric])\n",
    "\n",
    "history = model.fit(train_dataset.shuffle(shuffle_buffer_size)\n",
    "                    .batch(batch_size),\n",
    "                    epochs=epochs,\n",
    "                    validation_data=test_dataset.batch(64),\n",
    "                    callbacks=[\n",
    "                        reduce_lr,\n",
    "                        early_stop,\n",
    "                        TrainingTime(batch_size, learning_rate)\n",
    "                    ],\n",
    "                    verbose=0)\n",
    "\n",
    "msg = (f\"Masterful model Train accuracy: {max(history.history['accuracy']):4f}, \"\n",
    "       f\"and Test accuracy: {max(history.history['val_accuracy']):4f}\")\n",
    "print(msg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, model trained with Masterful's hyperparameter achieved better training and test accuracy in 50% less training time, just by selecting the right values for batch size and learning rate. To learn more theory on how we automatically find these hyperparameter values, please refer [this blog post](https://www.masterfulai.com/blog/stop-burning-money-on-the-wrong-batch-size)\n",
    "\n",
    "Masterful's full optimization suite can reduce training time and increase accuracy by many fold by customizing your training loop. To learn more on that, please check out our [docs](https://masterful-public.s3.us-west-1.amazonaws.com/933013963/0.4.1/index.html).\n",
    "\n",
    "We would love to hear your thoughts on this guide or the entire Masterful platform. Join our [slack community]( https://www.masterfulai.com/community) and let us know what you think."
   ]
  }
 ],
 "metadata": {
  "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": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}