File tree Expand file tree Collapse file tree 1 file changed +6
-1
lines changed
vllm/v1/attention/backends Expand file tree Collapse file tree 1 file changed +6
-1
lines changed Original file line number Diff line number Diff line change 99from vllm .attention .ops .paged_attn import PagedAttention
1010from vllm .attention .ops .prefix_prefill import context_attention_fwd
1111from vllm .logger import init_logger
12- from vllm .v1 .attention .backends .flash_attn import FlashAttentionMetadata
12+ from vllm .v1 .attention .backends .flash_attn import (
13+ FlashAttentionMetadata , FlashAttentionMetadataBuilder )
1314
1415logger = init_logger (__name__ )
1516
@@ -49,6 +50,10 @@ def get_kv_cache_shape(
4950 def use_cascade_attention (* args , ** kwargs ) -> bool :
5051 return False
5152
53+ @staticmethod
54+ def get_builder_cls () -> Type ["FlashAttentionMetadataBuilder" ]:
55+ return FlashAttentionMetadataBuilder
56+
5257
5358class ROCmAttentionImpl (AttentionImpl ):
5459
You can’t perform that action at this time.
0 commit comments