search

successar / Eraser-Benchmark-Baseline-Models

Baseline for ERASER benchmark
https://www.eraserbenchmark.com
17 stars 8 forks source link

Setup instructions are missing #2

Closed gagb closed 4 years ago

gagb commented 4 years ago

Hi Jay/Sarthak,

Thank you for releasing this codebase. Our lab is trying to use your re-implementation of Lei et. al.'s model. Unfortunately we've had little success getting this code base to even setup. For example, pip install -r requirements.txt fails. Further, there seem to be many dependency issues: It seems like you are using a custom fork of AllenNLP which installs pytorch 1.4.0, which conflicts with the version you specify in requirements.txt(1.3.0). Also, it seems that your custom AllenNLP fork has been updated since this codebase was released?? What version of it should we use?

Are you able to setup and run the code with the instructions in your README.md? Can you please update them?

Again, thank you for releasing this codebase and we look forward towards your response.

-Gagan

successar commented 4 years ago

Hi Gagan,

We will soon be releasing (by next week) new updates to this codebase that will have the correct allennlp code in it.

successar commented 4 years ago

I have updated the codebase with the allennlp within this repository. Please let me know if you are still facing any problems.