GitHub · Python
DeepGym/deepgym
RL environments with verifiable rewards you can trust. Sandboxed coding and computer-use tasks, verifiers kept apart from the answer, a QA gate against reward hacking, and infra failures kept out of training. Works with verl, TRL, OpenRLHF and Axolotl.
- Stars
- 22
- Forks
- 5
- Stars gained in 24 hours
- −3
- Forks gained in 24 hours
- 0
Topics: ai-agents, axolotl, coding-agents, computer-use, daytona, evaluation, grpo, hacktoberfest, llm, openrlhf
Latest data point: Nov 21, 2026 · 22 stars · 5 forks