Skip to content

Issues: volcengine/verl

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

40 Open 53 Closed
40 Open 53 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

kl is applied twice if kl_loss_coef >0 and kl_coef for GRPO
#265 opened Feb 13, 2025 by vermouth1992
FSDP model Parallel
#263 opened Feb 13, 2025 by AIRobotZhang
1
How long will it be before support for DeepSeek R1 is available?
#253 opened Feb 12, 2025 by zhangyipin
Flash Attention 2.0 only supports torch.float16 and torch.bfloat16 dtypes, but the current dype in Qwen2ForCausalLM is torch.float32.
#252 opened Feb 12, 2025 by OKC13
3
When will verl support deepseekv3
#251 opened Feb 12, 2025 by jameswu2014
Load checkpoint from default_local_dir & Save hdfs checkpoints
#250 opened Feb 12, 2025 by cliangyu
2
Does Verl Support for Running PPO or GRPO Algorithms on QWEN2.5 72B Model
#249 opened Feb 11, 2025 by none0663
1
CUDA Error Persists in Qwen GRPO Training Despite Setting VLLM_ATTENTION_BACKEND=XFORMERS
#246 opened Feb 11, 2025 by AIBionics
2
Having issues with vLLM for GRPO
#245 opened Feb 11, 2025 by 3rdAT
1
Running the GRPO program on multiple nodes causes it to hang.
#242 opened Feb 10, 2025 by zhilizju
2
CUDA error: an illegal memory access was encountered while increasing the maximum prompt length (max_prompt_length)
#241 opened Feb 10, 2025 by AIBionics
Support Generative Reward Model (GenRM) enhancement New feature or request good first issue Good for newcomers
#229 opened Feb 9, 2025 by maksimstw
@vermouth1992
5
Will veRL support deepspeed?
#221 opened Feb 7, 2025 by albertcity
3
0.5B model OOM when Initializing actor_rollout
#217 opened Feb 6, 2025 by MiaoLu3
2
Support Training with Both Function-Based Reward and DPO Reward Simultaneously
#214 opened Feb 6, 2025 by lianghsun
Question about kl_penalty question Further information is requested
#211 opened Feb 6, 2025 by StarDewXXX
@PeterSH6
1
Can we use VeRL to train the reward models enhancement New feature or request
#197 opened Feb 4, 2025 by YSLIU627
@vermouth1992
2
Integrate Verl with 🤗 hub
#190 opened Feb 3, 2025 by NielsRogge
1
1
if rollout.n is doubled, will the samples used for training doubled too? question Further information is requested
#180 opened Feb 1, 2025 by StarDewXXX
@vermouth1992
3
[Question] Is vLLMRollout.generate_sequences the right place to implement tool calling? enhancement New feature or request question Further information is requested vllm related
#176 opened Jan 31, 2025 by accupham
7
Add a format issue template
#172 opened Jan 30, 2025 by vermouth1992
Add nightly ci to ensure accuracy
#171 opened Jan 30, 2025 by vermouth1992
Mulit-modal rl training support? enhancement New feature or request
#168 opened Jan 30, 2025 by lucasjinreal
@vermouth1992 @PeterSH6
4
Add assertion to ensure the reward in GRPO is generated by ORM
#160 opened Jan 29, 2025 by vermouth1992
@PeterSH6
Any support for peft methods on PPO?
#159 opened Jan 29, 2025 by HaochenZhao
7
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.