engine_config = TurbomindEngineConfig(tp=2, quant_policy=0, cache_max_entry_count=0.2, session_len=4096)# quant_policy=8, self.pipe = pipeline("InternVL-Chat-V1-5", backend_config=engine_config) 其他配置参数不变，改变quant_policy=8，0，4 ，显存占用和推理速度没有任何改变是为什么呢？
Open
- Dominant language
- Python
- Stars
- 8.1k
- Forks
- 748
- Avg merge
- 6d 2h
- Merged PRs (30d)
- 54
Description
这是因为 lmdeploy 采用了"激进"的 kv cache mem分配策略
https://lmdeploy.readthedocs.io/en/latest/inference/pipeline.html#usage
可以参考上面文档的说明
_Originally posted by @lvhan028 in https://github.com/InternLM/lmdeploy/issues/1626#issuecomment-2122040558_
Contributor guide
Assessment
This issue has not been assessed yet.