biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[ROCm][AITER][Bugfix] Switch AITER to use PIECEWISE_AND_FULL compilation (#25104)

Browse Source 
Signed-off-by: Rohan138 <rohanpotdar138@gmail.com>
This commit is contained in:
Rohan Potdar
2025-09-18 12:46:47 -05:00
committed by GitHub
parent dc34059360
commit bbdc0f2366
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/v1/attention/backends/rocm_aiter_fa.py
View File

@@ -232,7 +232,7 @@ class AiterFlashAttentionMetadata:
class AiterFlashAttentionMetadataBuilder(
AttentionMetadataBuilder[AiterFlashAttentionMetadata]):
cudagraph_support = AttentionCGSupport.ALWAYS
cudagraph_support = AttentionCGSupport.UNIFORM_SINGLE_TOKEN_DECODE
def __init__(self, kv_cache_spec: AttentionSpec, layer_names: list[str],
vllm_config: VllmConfig, device: torch.device):