hsm207/imdb_data

Script to download IMDb data and convert it to tsv files

Introduction

This repository contains code to create a tsv file of the IMDb dataset using the tensor2tensor library.

Usage

1. Create and switch to a new Python 3.6+ environment.
2. Navigate to the project's root directory.
3. Execute:

   pip install -r requirements.txt

4. Execute:

   python create_imdb_dataset.py --output_dir OUTPUT_DIR

   where OUTPUT_DIR is the path to where you want to save the training and test files.