Skip to content


Folders and files

Last commit message
Last commit date

Latest commit


Repository files navigation

Probabilistic Neural-symbolic Models

Code for our ICML 2019 paper:

Probabilistic Neural-Symbolic Models for Interpretable Visual Question Answering
Ramakrishna Vedantam, Karan Desai, Stefan Lee, Marcus Rohrbach, Dhruv Batra, Devi Parikh

Checkout our package documentation at!


If you find this code useful, please consider citing:

  title={Probabilistic Neural-symbolic Models for Interpretable Visual Question Answering},
  author={Ramakrishna Vedantam and Karan Desai and Stefan Lee and Marcus Rohrbach and Dhruv Batra and Devi Parikh},

Usage Instructions

  1. How to setup this codebase?
  2. How to train your ProbNMN?
  3. How to evaluate or run inference?

Pre-trained Checkpoint

Pre-trained checkpoints and corresponding config files (with all the hyper-parameters) for all training phases is available with v1.0 release of this repository. Check out the Releases!


We thank the developers of:

  1. @davidmascharka/tbd-nets for providing a very clean implementation of our core Neural Module Network.

  2. @allenai/allennlp for providing an awesome framework which indeed takes masking and padding seriously.

  3. @rbgirshick/yacs for providing an efficient package-wide configuration management.

  4. @pytorch/pytorch, this needs no reason.