A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 18:15 UTC
[Bugfix][Harmless] Fix float16 dtype for model_is_embedding (#7566)
Michael Goin committed
Aug 15, 2024 at 21:26 UTC
9c8e2d11611ca5313710df06f215b41f05dde1b4
1 file changed
+1
-1
vllm/entrypoints/openai/api_server.py
+1
-1
@@ -66,7 +66,7 @@ def model_is_embedding(model_name: str, trust_remote_code: bool) -> bool:
66
tokenizer_mode="auto",
67
trust_remote_code=trust_remote_code,
68
seed=0,
69
- dtype="float16").embedding_mode
69
+ dtype="auto").embedding_mode
70
71
72
@asynccontextmanager