[Bug] non-async call in async fn

[lz1998]

Checklist

• 1. I have searched related issues but cannot get the expected help.
• 2. The bug has not been fixed in the latest version.
• 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.

Describe the bug

这个 async 是不是假的，里面的 forward 是同步的，CPU 执行到这里，等待 GPU 过程中不会让出执行权。
https://github.com/InternLM/lmdeploy/blob/main/lmdeploy/pytorch/engine/model_agent.py#L260

Reproduction

https://github.com/InternLM/lmdeploy/blob/main/lmdeploy/pytorch/engine/model_agent.py#L260

Environment

https://github.com/InternLM/lmdeploy/blob/main/lmdeploy/pytorch/engine/model_agent.py#L260

Error traceback

[grimoire]

_forward_impl 起 kernel 之后 asyncio.sleep(0) 会让出的。