TowardsTransaparentAI_vGHC2020

This repository contains content shared in Towards Transparent AI: Understanding Text-Based Model Predictions workshop at Grace Hopper Conference 2020.

Getting Started

This section shows different ways to get setup for the workshop Use one of the 3 approaches to get started with the content. The recommened way to get setup is to follow the local setup section.

Python

For best results please setup this repository on your local machine.

1. Clone this repository to access the sample notebook

    git clone https://github.com/janhavi13/TowardsTransaparentAI_vGHC2020.git

2. Easiest way to run the sample notebook is to create a conda environment

    conda create -n myenv python=3.6.8
    conda activate myenv
    # Find the appropriate torch for your machine from this link: https://pytorch.org/get-started/locally/
    pip install torch===1.6.0 torchvision===0.7.0 -f https://download.pytorch.org/whl/torch_stable.html # for windows
    pip install datasets
    pip install interpret-text
    pip install jupyter

Note: If installing packages in requirements.txt one by one, make sure torch and torchvision is installed before the interpret-text package.

3. Run the jupyter notebook using the command from TowardsTransaparentAI_vGHC2020 location on your laptop/desktop.

    jupyter notebook

Docker

Another way to run the repository content easily is via our pre-built Docker container. To do so, run the following command:

docker run -it -p 8888:8888 -m=5g janhavim13/interprettext:latest

Navigate to http://localhost:8888/ in your web browser to run the sample notebooks.

The getting started guide on Docker has detailed instructions for setting up Docker on Mac, Linux and Windows. See the documentation for more on Docker use.

Binder

Click on the binder badge above to run the sample notebook in your browser