CLEVR Graph Parser Library

CLEVR-Parser is a graph parser for the CLEVR dataset. Code currrently under review stage - soon to be released.

Table of Contents

• Introduction
• Getting Started
• Getting Help
• Demo
• Installation
  • Install ClevrParser using PyPI
  • Install ClevrParser in Anaconda Python
  • Install ClevrParser from Github source
• Citing
• References

Introduction

Getting Started

Getting Help

If you get stuck or have a problem, there's many ways to make progress and get help or support:

• Read the documentation
• [Consult the examples][demos]
• Contact us:
  • Ask questions and discuss problems on the ClevrParser Discourse forum
  • File an issue
  • Send us an email at asd|rsaqur@cs.princeton.edu

Demo

Run the download.sh script form the root directory. clevr_parser |-data/raw

Installation

ClevrParser is a Python 3 library and we recommend using Python version 3.7. The required Python version can be downloaded and installed from python.org. Alternatively, use the Anaconda Python environment, available from anaconda.com.

The ClevrParser library can be installed from PyPI, from Anaconda Cloud, or directly from GitHub, as described below.

Install ClevrParser using PyPI:

To install ClevrParser library from PyPI using pip, execute the following command:

$ pip install clevr-parser

[Some of the examples][demos] require installing additional dependencies as well as clevr-parser. To install these dependencies as well as ClevrParser using pip execute the following command:

$ pip install -r requirements.txt 

Install ClevrParser in Anaconda Python:

The ClevrParser library is available an Anaconda Cloud and can be installed in Anaconda Python using the command line conda tool, execute the following command:

$ conda install -c raeidsaqur clevr-parser

Install ClevrParser from Github source:

First, clone the ClevrParser repository using git:

$ git clone https://github.com/raeidsaqur/clevr-parser.git

Then, cd to the ClevrParser folder, and install the library by executing the following commands:

$ cd clevr-parser
$ pip install .

Prerequisites

Select the backend to use:

-- Spacy --

The default backend uses spacy for language parsing and pretrained LM models used for embeddings.

Please see spacy's doc for installation instructions.

Spacy language models (LM) can be downloaded following instructions here. N.b. the spacy-transformers package (homepage, github), can be used to download SotA transformer based (BERT, XLNet, RoBerTa) LMs - including the popular HuggingFace implementations.

The very basic installation entails:

$ pip install spacy
$ python -m spacy download en_core_web_sm 

Once installed, validate the available LMs using: python -m spacy info and python -m spacy validate.

-- Stanford NLP --

Any other standard NLP parsers can be used as a backend. A concrete parser class should extend the ParserBackend class, satisfying the Parser interface signature (e.g. parse(..)).

Citing

If you use any part of this library in your research, please cite it using the following BibTex entry

@article{saqur2020clevr,
  title={CLEVR Parser: A Graph Parser Library for Geometric Learning on Language Grounded Image Scenes},
  author={Saqur, Raeid and Deshpande, Ameet},
  journal={arXiv preprint arXiv:2009.09154},
  year={2020}
}