-
Notifications
You must be signed in to change notification settings - Fork 2.7k
Pull requests: volcengine/verl
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[rollout, vllm] fix: add compatibility shim for vllm process_weights_after_loading
#4298
opened Nov 26, 2025 by
Kidand
Loading…
7 tasks done
[algo] fix: correctly aggregate metrics in actor/critic
#4294
opened Nov 26, 2025 by
dongju-2
Loading…
4 of 7 tasks
[rollout] fix: some compatibility changes in agent loop and reward
#4293
opened Nov 25, 2025 by
pengwu22
Loading…
5 tasks done
Fix: Pass attn_implementation override to CriticWorker
#4292
opened Nov 25, 2025 by
arde171
Loading…
3 of 7 tasks
Fix FSDP checkpoint loading with world_size=1
#4290
opened Nov 25, 2025 by
shilogic0929
Loading…
7 tasks
[data] feat: TransferQueue - Support AgentLoop performance metrics & minor fix
#4289
opened Nov 25, 2025 by
0oshowero0
Loading…
6 of 7 tasks
[Recipe, Prime] fix: prime compatibility issues
#4287
opened Nov 25, 2025 by
lsy641
Loading…
4 of 7 tasks
[BREAKING][worker, rollout, vllm] feat: implement vLLM co-located training-inference rollout with process separation
#4280
opened Nov 25, 2025 by
jianjunzhong
•
Draft
7 tasks
[fsdp] fix: shard LoRA modules separately to prevent dtype mismatch errors with FSDP2
#4275
opened Nov 24, 2025 by
xxnpark
Loading…
7 tasks done
[WIP][megatron] feat: support discrete profiling for mindspeed
#4271
opened Nov 24, 2025 by
tardis-key
Loading…
3 of 7 tasks
why are there multiple settings for
actor_rollout_ref.model.enable_gradient_checkpointing? Is this a deliberate design choice?
#4263
opened Nov 24, 2025 by
khazic
Loading…
[Draft,Don't reivew][fsdp_workers] fix: Skip FSDP loading in async rollout mode to save GPU memory
#4257
opened Nov 24, 2025 by
JobQiu
Loading…
4 tasks done
[trainer] fix: use unscaled logits for accurate pearson correlation metric
#4252
opened Nov 23, 2025 by
JobQiu
Loading…
[vllm, rollout] fix: lazy initialize ZMQ to avoid event loop error
#4246
opened Nov 23, 2025 by
JobQiu
Loading…
[BREAKING][trainer, worker] feat: multi generative reward models
#4241
opened Nov 22, 2025 by
alek6kun
Loading…
5 tasks done
[env] fix: Move uvloop==0.21.0 dependency to requirement files for avoiding task errors
#4231
opened Nov 21, 2025 by
duesdues
Loading…
6 of 7 tasks
[recipe] feat: support reward_loop for recipe/fully_async_policy
#4224
opened Nov 21, 2025 by
sl-1314
Loading…
7 tasks done
[RFC] A Fully decoupled and auto-scaled rollout engine using AWS Bedrock AgentCore Runtime
#4216
opened Nov 20, 2025 by
luyuzhe111
Loading…
3 of 5 tasks
[megatron] fix: Pass optimizer config betas to Megatron optimizer config
#4214
opened Nov 20, 2025 by
guillemgt
Loading…
6 of 7 tasks
[tool] fix: Add missing __init__.py to mcp_clients for package inclusion & use timeout param in call_tool of McpClientManager.py
#4212
opened Nov 20, 2025 by
brzqbrzq
Loading…
6 of 7 tasks
[vllm] fix: Fix issues that occur during the ACLGraph initialization process in the NPU.
#4209
opened Nov 20, 2025 by
chengminhua
Loading…
3 of 7 tasks
[Security] Fix HIGH vulnerability: python.lang.security.audit.eval-detected.eval-detected
#4204
opened Nov 20, 2025 by
orbisai0security
Loading…
[trainer] feat: make max_colocate_count configurable in ResourcePoolManager
#4197
opened Nov 19, 2025 by
JobQiu
Loading…
3 tasks done
Previous Next
ProTip!
Exclude everything labeled
bug with -label:bug.