Search results

Ray::TaskRunner.run out of memory - cpu

Status: Open.

#4995 In volcengine/verl;

· michael-tjuni opened on Jan 20, 2026

assert len(mm_embeddings) == expected_num_items

bugSomething isn't workingSomething isn't working

Status: Open.

#4994 In volcengine/verl;

· yingyukexiansheng opened on Jan 20, 2026

hf rollout

Status: Open.

#4992 In volcengine/verl;

· limitedfxw opened on Jan 20, 2026

Qwen 3 VL 30b moe training fails when save checkpoint on megatron

Status: Open.

#4990 In volcengine/verl;

· Hayak3 opened on Jan 20, 2026

deepeyes的recipe的bug

bugSomething isn't workingSomething isn't working

Status: Open.

#4989 In volcengine/verl;

· liuchaohu opened on Jan 20, 2026

Qwen3 VL training and validation data cannot be truncated correctly.

Status: Open.

#4975 In volcengine/verl;

· B1ueorange opened on Jan 19, 2026

An error occurs when setting the <code>reshard_after_forward=False</code> while offloading the model to the CPU.

bugSomething isn't workingSomething isn't working

Status: Open.

#4974 In volcengine/verl;

· ji-huazhong opened on Jan 19, 2026

If there are plan for integrating muon optimizer

Status: Open.

#4966 In volcengine/verl;

· Kedaya66 opened on Jan 17, 2026

QLoRA with GRPO in verl

Status: Open.

#4963 In volcengine/verl;

· JohnConnor123 opened on Jan 17, 2026

Importing import verl.trainer.sft_trainer_ray requires qwen-vl-utils

bugSomething isn't workingSomething isn't working

Status: Open.

#4958 In volcengine/verl;

· thvasilo opened on Jan 16, 2026

Encounter <code>Exception: Could not find the transformer layer class to wrap in the model. </code> with gemma3 and try to fix

Status: Open.

#4950 In volcengine/verl;

· chiseanchang0727 opened on Jan 16, 2026

Has anyone successfully run GKD?

bugSomething isn't workingSomething isn't working

Status: Open.

#4945 In volcengine/verl;

· kaiyliu opened on Jan 16, 2026