Bug when enable lora as true

[abcdvzz]

Checklist / 检查清单

• I have searched existing issues, and this is a new bug report. / 我已经搜索过现有的 issues，确认这是一个新的 bug report。

Bug Description / Bug 描述

num_cache_lines, _, state_len = conv_state.size()

When I ran this: https://github.com/modelscope/twinkle/blob/main/cookbook/rl/grpo.py, the above issue happend with enable_lora_true.

How to Reproduce / 如何复现

N/A

Additional Information / 补充信息

No response

[tastelikefeet]

Which model and which version of vLLM you are using? And may you share the error stack for better locating the root cause?