q yao
@zhyxun 麻烦提供一下复现方式还有环境信息
@zhyxun 我这里没办法复现。能提供一张确定能复现的图片数据吗？还有大约跑多少数据会发生错误？
@zhyxun https://github.com/InternLM/lmdeploy/pull/3307 试试看
> ``` > lmdeploy - ERROR - async_engine.py:592 - [safe_run] exception caught: GeneratorExit > ``` > 这个一般是请求连接相关的错，和引擎关系不大，@AllentDan 能不能帮忙看下
Thanks for the notification. This should be a bug of a big refactor https://github.com/open-mmlab/mmdeploy/pull/1091. Removing either of the slices should be ok.
> I can create a PR for this patch too if it helpss. Sure, that would be cool.
https://github.com/InternLM/lmdeploy/blob/edcdd8e36520b8bf7dbc99feecd2d2822c4cb5ba/lmdeploy/pytorch/kernels/cuda/pagedattention.py#L35 https://github.com/InternLM/lmdeploy/blob/edcdd8e36520b8bf7dbc99feecd2d2822c4cb5ba/lmdeploy/pytorch/kernels/cuda/pagedattention.py#L583 可以试试看把这些地方的 num_stages 改成 1
如果 adapter 对 vision 的部分做了修改那么暂时还没有支持
`lmdeploy serve api_server` + `profile_restful_api.py` failed with log: ``` 2025-05-14 19:54:29,323 - lmdeploy - ERROR - async_engine.py:599 - [safe_run] exception caught: AttributeError 'NoneType' object has no attribute 'get' ```
We will do some environment checks before starting engine. As the log indicate, the package triton check failed. Our triton check is a simple vector add kernel. https://github.com/InternLM/lmdeploy/blob/05914be1e0ef3ceb9d0ce37ca4912bc3ec2e2864/lmdeploy/pytorch/check_env/triton_custom_add.py#L28-L33 Update your...