Skip to content

GitLab

Unverified Commit 794ae1f5 authored by fxmarty-amd's avatar fxmarty-amd Committed by GitHub
Browse files

[rocm] Fix wrong attention log (#18764) 


Signed-off-by: default avatarFelix Marty <felmarty@amd.com>
parent d73a9457
Hide whitespace changes
Inline Side-by-side
Showing with 3 additions and 2 deletions
vllm/platforms/rocm.py
View file @ 794ae1f5
...@@ -194,8 +194,9 @@ class RocmPlatform(Platform): ...@@ -194,8 +194,9 @@ class RocmPlatform(Platform):
f" The selected backend, {selected_backend.name}," f" The selected backend, {selected_backend.name},"
f"is not MLA type while requested for MLA backend.") f"is not MLA type while requested for MLA backend.")
selected_backend = (_Backend.ROCM_FLASH if selected_backend if selected_backend is None or selected_backend == _Backend.FLASH_ATTN:
== _Backend.FLASH_ATTN else selected_backend) selected_backend = _Backend.ROCM_FLASH
if envs.VLLM_USE_V1: if envs.VLLM_USE_V1:
logger.info("Using Triton Attention backend on V1 engine.") logger.info("Using Triton Attention backend on V1 engine.")
return ("vllm.v1.attention.backends." return ("vllm.v1.attention.backends."
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment