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

144 Open 120 Closed
144 Open 120 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

[bugfix] Error for preprocessed dataset ['prompt_key'] and ['response_key']
#600 opened Mar 14, 2025 by jiahaolu97
1
install error
#598 opened Mar 14, 2025 by xiaoooobai
Stuck after logging NCCL version, before logging model size
#597 opened Mar 14, 2025 by JunzheShen
Support for freezing the vision tower in Qwen 2.5 VL
#596 opened Mar 14, 2025 by langfengQ
Use ray submit for GRPO training, encountering errors during multi-node training.
#595 opened Mar 14, 2025 by bestpredicts
1
about grpo loss computation
#593 opened Mar 14, 2025 by gameofdimension
@tongyx361
Setting CUDA_VISIBLE_DEVICES for Ray Actors
#589 opened Mar 14, 2025 by jinqinn
assert len(seqs) == 1, AssertionError
#584 opened Mar 13, 2025 by none0663
maximum recursion depth exceeded
#582 opened Mar 13, 2025 by awdrgyjilplij
1
[Bug] A bug in fsdp_worker.py ?
#580 opened Mar 13, 2025 by SportsSan
1
Suggestions for Refactoring the recompute_old_log_prob Feature
#576 opened Mar 13, 2025 by Tshiyao
How to switch on the multi-GPU for GRPO Training?
#574 opened Mar 13, 2025 by tjoymeed
Issues with Megatron install
#573 opened Mar 13, 2025 by sjelassi
Problem with multi-node PPO/GRPO - OwnerDiedError
#572 opened Mar 13, 2025 by juneodie
2
Problems related to the parameter settings of ulysses_sequence_parallel_size
#571 opened Mar 13, 2025 by nomadlx
2
Install veRL in container images
#567 opened Mar 12, 2025 by kevin85421
[Ray][K8s] Container image doesn't include Ray dashboard dependencies
#566 opened Mar 12, 2025 by kevin85421
2
Why do the computed grad_norm differ when using the same loss calculation?
#550 opened Mar 12, 2025 by yuecheng-ma
1
[Bug] A bug in the slurm script example. It will cause "srun: error: Unable to create step for job xxxxx: Requested node configuration is not available"
#548 opened Mar 11, 2025 by SkyFishMoon
Why is the GRPO training code implemented within the ppo package?
#546 opened Mar 11, 2025 by BaiqingL
1
[Bug] NotImplementedError: sequence_length=1081 is larger than max_length=1024
#544 opened Mar 11, 2025 by caaatch22
2
AttributeError: 'Parameter' object has no attribute 'weight_loader' when using FP8 quantization in vLLM
#540 opened Mar 11, 2025 by Wraythh
2
[BUG] _check_resource_available of verl/trainer/ppo/ray_trainer.py causes error
#539 opened Mar 11, 2025 by heyzude
5
Reproduce the issue of R1 zero in the 32B model, where the reward does not increase after a certain step
#538 opened Mar 11, 2025 by xiaojiangzhang
2
actor和rollout分离应该如何实现
#537 opened Mar 11, 2025 by a4zhangfei
@wuxibin89
4
ProTip! Exclude everything labeled bug with -label:bug.