xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-04-02 17:07:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix pre-commit

Browse Source 
Signed-off-by: yuantao <2422264527@qq.com>
This commit is contained in:
yuantao 2025-12-17 11:03:13 +08:00
parent 93a7afcab3
commit 378e20833f
4 changed files with 10 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/v1/core/single_type_kv_cache_manager.py
View File

@ -855,7 +855,8 @@ class SinkFullAttentionManager(FullAttentionManager):
req_blocks = self.req_to_blocks[request_id]
assert len(req_blocks) == 0
# Append both sink blocks and hitted prefix cache blocks
req_blocks.extend(self.sink_blocks + new_computed_blocks)
req_blocks.extend(self.sink_blocks)
req_blocks.extend(new_computed_blocks)
self.num_cached_block[request_id] = len(new_computed_blocks)
else:
# A running request. Should not have new computed blocks.

4
vllm/v1/worker/gpu/attn_utils.py
View File

@ -102,8 +102,8 @@ def _reshape_kv_cache(
attn_backend = attn_backends[layer_name]
if (
getattr(kv_cache_spec, "head_size_v", kv_cache_spec.head_size)
!= kv_cache_spec.head_size
hasattr(kv_cache_spec, "head_size_v")
and kv_cache_spec.head_size_v != kv_cache_spec.head_size
):
kwargs = {"head_size_v": kv_cache_spec.head_size_v}
stride_kwargs = {"diff_kv": True}

4
vllm/v1/worker/gpu_model_runner.py
View File

@ -5207,8 +5207,8 @@ class GPUModelRunner(
kernel_num_blocks = num_blocks * num_blocks_per_kv_block
if (
getattr(kv_cache_spec, "head_size_v", kv_cache_spec.head_size)
!= kv_cache_spec.head_size
hasattr(kv_cache_spec, "head_size_v")
and kv_cache_spec.head_size_v != kv_cache_spec.head_size
):
kwargs = {"head_size_v": kv_cache_spec.head_size_v}
stride_kwargs = {"diff_kv": True}

8
vllm/v1/worker/kv_connector_model_runner_mixin.py
View File

@ -191,8 +191,8 @@ class KVConnectorModelRunnerMixin:
attn_backend = attn_group.backend
if (
getattr(kv_cache_spec, "head_size_v", kv_cache_spec.head_size)
!= kv_cache_spec.head_size
hasattr(kv_cache_spec, "head_size_v")
and kv_cache_spec.head_size_v != kv_cache_spec.head_size
):
kwargs = {"head_size_v": kv_cache_spec.head_size_v}
stride_kwargs = {"diff_kv": True}
@ -269,8 +269,8 @@ class KVConnectorModelRunnerMixin:
attn_backend = attn_group.backend
if (
getattr(kv_cache_spec, "head_size_v", kv_cache_spec.head_size)
!= kv_cache_spec.head_size
hasattr(kv_cache_spec, "head_size_v")
and kv_cache_spec.head_size_v != kv_cache_spec.head_size
):
kwargs = {"head_size_v": kv_cache_spec.head_size_v}
stride_kwargs = {"diff_kv": True}