Search results

How to Integrate Process Rewards in LangGraph Agents and Handle Node Execution Failures?

Status: Open.

#387 In microsoft/agent-lightning;

· dalek-who opened on Dec 9, 2025

[Feature] Add data validation steps to <code>trainer.dev</code> for early error detection

waiting for reply

Status: Open.

#385 In microsoft/agent-lightning;

· dalek-who opened on Dec 8, 2025

BUG: Compatibility with verl 0.6.0 and LoRA

bugSomething isn't workingSomething isn't working

help wantedExtra hands from community will be appreciatedExtra hands from community will be appreciated

verl

Status: Open.

#383 In microsoft/agent-lightning;

· dalek-who opened on Dec 8, 2025

Redundant logic after using aiologic

enhancementNew feature or requestNew feature or request

store

Status: Open.

#372 In microsoft/agent-lightning;

· ultmaster opened on Dec 5, 2025

Clarify the recommended value and impact of n_workers when instantiating Trainer

documentationImprovements or additions to documentationImprovements or additions to documentation

enhancementNew feature or requestNew feature or request

Status: Open.

#366 In microsoft/agent-lightning;

· flyflypeng opened on Dec 4, 2025

How to assign different LoRA to different sub-agents?

questionQuestion about a feature or some usageQuestion about a feature or some usage

verl

Status: Open.

#365 In microsoft/agent-lightning;

· dalek-who opened on Dec 4, 2025

Can I use https://github.com/microsoft/llm-as-judge as the search reward for GRPO training by combining it in?

questionQuestion about a feature or some usageQuestion about a feature or some usage

topic/rag

Status: Open.

#362 In microsoft/agent-lightning;

· johnson7788 opened on Dec 4, 2025

[feature request] OneStepOff Trainer

enhancementNew feature or requestNew feature or request

verl

Status: Open.

#360 In microsoft/agent-lightning;

· syeehyn opened on Dec 3, 2025

Completed rollouts &gt; total_tasks_queued

help wantedExtra hands from community will be appreciatedExtra hands from community will be appreciated

verl

Status: Open.

#355 In microsoft/agent-lightning;

· chenhr18thu opened on Dec 3, 2025

How can I use Agent Lightning for fine-tuning an agent’s system prompt using OpenAI models (gpt-4o)? Can VERL be used for this?

apo

questionQuestion about a feature or some usageQuestion about a feature or some usage

Status: Open.

#351 In microsoft/agent-lightning;

· rutvik-jaiswal-deeplearning opened on Dec 2, 2025

How to turn off the thinking mode

questionQuestion about a feature or some usageQuestion about a feature or some usage

verl

Status: Open.

#347 In microsoft/agent-lightning;

· ccp123456789 opened on Dec 1, 2025

Verl Rollout is just blocked, GPU utils is always 0

examples

topic/rag

verl

Status: Open.

#339 In microsoft/agent-lightning;

· reallm opened on Nov 27, 2025