Search results

请问在RL阶段，训练时模型每次的rollout都是实际调用了工具并且把结果返回给下一轮的推理吗？还是只模拟tool response，没有实际获取工具调用的实际结果呢？

Status: Open.

#16 In EvolvingLMMs-Lab/LongVT;

· alanshaoTT opened on Feb 8, 2026

LongVT-RL在videomme上tool_call频率非常低

Status: Open.

#15 In EvolvingLMMs-Lab/LongVT;

· HiFiChang opened on Feb 5, 2026

*** ValueError: The features can't be aligned because the key messages of features 问题

Status: Open.

#14 In EvolvingLMMs-Lab/LongVT;

· fazhdo opened on Jan 30, 2026

Question about metrics when evaluating LVBench

Status: Open.

#11 In EvolvingLMMs-Lab/LongVT;

· Yusux opened on Dec 29, 2025

Json Error in evaluation process

Status: Open.

#10 In EvolvingLMMs-Lab/LongVT;

· wkfdb opened on Dec 23, 2025