Search results

GRPO advantage computation uses random UUIDs instead of group_ids, effectively disabling group normalization

Status: Open.

#36 In modelscope/AgentEvolver;

· yueyiwen-create opened on Feb 9, 2026

Questions about BFCL dataset split strategy, multi-turn variants, and model context length

Status: Open.

#34 In modelscope/AgentEvolver;

· lazymieie opened on Jan 27, 2026

Could you provide the script for training on bfcl?

Status: Open.

#32 In modelscope/AgentEvolver;

· gitrxh opened on Jan 16, 2026

[Bug]: EnvWorkerWithPrompt in LlmDedupSamplingExploreStrategy is not defined

bugSomething isn't working Something isn't working

Status: Open.

#31 In modelscope/AgentEvolver;

· xuanzhaopeng opened on Jan 7, 2026

Unexpected high memory usage in AppWorld Env leading to Ray OOM

bugSomething isn't working Something isn't working

Status: Open.

#30 In modelscope/AgentEvolver;

· jingyi-xioa opened on Dec 28, 2025

[Feature]: Add conversion script from training checkpoints to final Hugging Face format

enhancementNew feature or request New feature or request

Status: Open.

#29 In modelscope/AgentEvolver;

· zhanglingzhe0820 opened on Dec 26, 2025

Release CuES Generated Tasks and Trajectories on Hugging Face

Status: Open.

#19 In modelscope/AgentEvolver;

· NielsRogge opened on Dec 4, 2025

Launcher crashes when <code>REME_PATH</code>/<code>REME_SCRIPT</code> are unset

Status: Open.

#9 In modelscope/AgentEvolver;

· arjuntheprogrammer opened on Nov 20, 2025

Document concrete system requirements for running AgentEvolver

Status: Open.

#7 In modelscope/AgentEvolver;

· arjuntheprogrammer opened on Nov 20, 2025