google-deepmind / google-deepmind/zipfian_environments
Reproduce Results
Open
- Dominant language
- Python
- Stars
- 28
- Forks
- 1
- PR merge metrics
- No merged PRs in 30d
Description
I tried training IMPALA on gridworld using the implementation from [torchbeast](https://github.com/facebookresearch/torchbeast). However, the results (uniform + uniform rare accuracy) I'm getting are not exactly matching the results in paper. Could you please share the implementation of IMPALA that was used in the paper.
Contributor guide
Assessment
This issue has not been assessed yet.