RL Integration
OpenRLHF verl TRL
Topic | Replies | Views | Activity | |
---|---|---|---|---|
About the RL Integration category |
![]() |
0 | 26 | March 26, 2025 |
Can the AsyncEngine be compatible with the backend of external_launcher? |
![]() ![]() ![]() |
3 | 48 | October 8, 2025 |
Degraded RL training with v0.10.2 |
![]() ![]() ![]() ![]() |
8 | 816 | September 16, 2025 |
Running vLLM multi-node data parallel with SLURM |
![]() ![]() |
1 | 217 | August 12, 2025 |
Intergate async-llm to openrlfh, when tensor_parallel_size > 1 will cause ray-error |
![]() ![]() |
1 | 106 | April 27, 2025 |
Is is possible to initialize an AsyncLLMEngine inside the LLM object? |
![]() ![]() ![]() |
4 | 243 | April 12, 2025 |
No HIP GPUs are available for VeRL |
![]() ![]() ![]() ![]() |
4 | 388 | April 4, 2025 |
Numerical Difference between vLLM logprobs and huggingface logprobs |
![]() ![]() ![]() ![]() |
7 | 3908 | April 4, 2025 |
RL Training with vLLM Rollout: How to Mitigate Load Imbalance from Variable Response Lengths |
![]() ![]() ![]() ![]() |
4 | 501 | April 1, 2025 |