google-deepmind / google-deepmind/reverb
How to use multiple clients - for multiple envs in parallel settings ?
Open
- Dominant language
- C++
- Stars
- 793
- Forks
- 113
- PR merge metrics
- No merged PRs in 30d
Description
Hi,
How can we use Reverb for multiple environments in a parallel setting - where we have multiple workers - each for one environment - to speed up the data collection part?
Thank you !
Contributor guide
Research direction
The issue mentions Reverb, multiple clients, workers, environments, and parallel data collection, but names no files or tests. Start by reviewing the existing Reverb usage documentation and client setup guidance for multi-environment operation. Done means documenting whether this configuration is supported and explaining the supported setup or limitation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp, machine-learning
- Domain
- data-engineering, distributed-systems
- Issue type
- Documentation
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100