{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "07f18fcf-04de-4405-9c15-224becc96b24",
   "metadata": {
    "tags": []
   },
   "source": [
    "# AutoGluon Environment Validation\n",
    "\n",
    "[![Open in SageMaker Studio Lab](https://studiolab.sagemaker.aws/studiolab.svg)](https://studiolab.sagemaker.aws/import/github/aws/studio-lab-examples/blob/main/custom-environments/AutoGluon/env_validation.ipynb)\n",
    "\n",
    "AutoML is a technology that aims to automate the process of applying machine learning to real-world problems, or to automate the process of building machine learning models from data. [AutoGluon](https://github.com/awslabs/autogluon) enables easy-to-use and easy-to-extend AutoML with a focus on automated stack ensembling, deep learning, and real-world applications spanning text, image, and tabular data.\n",
    "\n",
    "In this notebook, we will set up AutoGluon in [Amazon SageMaker Studio Lab](https://studiolab.sagemaker.aws/), a free machine learning development environment, and train machine learning models using AutoGluon on a sample dataset.\n",
    "\n",
    "Run the following cell to generate a YAML file for creating the Anaconda virtual environment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "34ecced8-486e-4f77-b912-6dc95dd0ec62",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%writefile autogluon_cpu.yml\n",
    "# run in terminal: conda env create --file autogluon_cpu.yml\n",
    "# note: This command creates an environment that is larger than 4GB.\n",
    "name: autogluon_cpu\n",
    "dependencies:\n",
    "  - python=3.9\n",
    "  - conda\n",
    "  - pip\n",
    "  - ipykernel\n",
    "  - nodejs\n",
    "  - pip:\n",
    "    - ipywidgets\n",
    "    - setuptools\n",
    "    - wheel\n",
    "    - autogluon\n",
    "    - jupyter_bokeh"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d27d5c92-6601-4d8d-bd7a-7e69a85e8000",
   "metadata": {},
   "source": [
    "After `autogluon_cpu.yml` is generated, run `conda env create -f autogluon_cpu.yml` in your system terminal to create a virtual environment for Anaconda with AutoGluon set up.\n",
    "\n",
    "If the installation is successful, you should be able to select `autogluon_cpu:Python` in the popup that appears when you click the kernel name in the upper right corner of the screen. Switch to the `autogluon_cpu:Python` kernel. You should now be able to use AutoGluon."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f6306047-ab5c-491a-aa07-1a436fd39efb",
   "metadata": {},
   "source": [
    "## Import Packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "13885941-3217-4b72-baf7-ed140ebffabc",
   "metadata": {},
   "outputs": [],
   "source": [
    "from autogluon.tabular import TabularDataset, TabularPredictor\n",
    "from IPython.display import HTML"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4248d97f-a540-4e33-a112-d569860c52a6",
   "metadata": {},
   "source": [
    "## Train on Sample Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0caa3a29-f341-4e6f-ab27-c8bb6cc6dfe6",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_data = TabularDataset('https://autogluon.s3.amazonaws.com/datasets/Inc/train.csv')\n",
    "test_data = TabularDataset('https://autogluon.s3.amazonaws.com/datasets/Inc/test.csv')\n",
    "predictor = TabularPredictor(label='class', path='ag-default').fit(train_data, time_limit=120)  # Fit models for 120s"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "35954ea3-14b2-4a3f-b47a-fb4e57507493",
   "metadata": {},
   "source": [
    "## Check Models Summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "627eab9c-8073-4c88-aec5-48ca3656f510",
   "metadata": {},
   "outputs": [],
   "source": [
    "leaderboard = predictor.leaderboard(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "888d191e-5dc9-4015-b7da-0f409f2d581c",
   "metadata": {},
   "outputs": [],
   "source": [
    "results = predictor.fit_summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e8acc5ec-92e3-4bfe-8b02-413439b7d1c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "HTML(filename='ag-default/SummaryOfModels.html')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "autogluon_cpu:Python",
   "language": "python",
   "name": "conda-env-autogluon_cpu-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}