[V1] Support any head size for FlexAttention backend (#20467)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-07-06 09:54:36 -07:00
committed by GitHub
parent e202dd2736
commit 9fb52e523a
20 changed files with 202 additions and 118 deletions
+2 -2
View File
@@ -12,8 +12,8 @@ from vllm.logger import init_logger
from vllm.model_executor.model_loader import get_model
from vllm.model_executor.models import supports_multimodal
from vllm.model_executor.models.llama_eagle3 import Eagle3LlamaForCausalLM
from vllm.v1.attention.backends.flash_attn import (CommonAttentionMetadata,
FlashAttentionMetadata)
from vllm.v1.attention.backends.flash_attn import FlashAttentionMetadata
from vllm.v1.attention.backends.utils import CommonAttentionMetadata
from vllm.v1.kv_cache_interface import KVCacheConfig
from vllm.v1.sample.metadata import SamplingMetadata
from vllm.v1.spec_decode.utils import prepare_eagle_input_kernel