xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-23 18:35:53 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix linting and address depcrecation warnings

Browse Source 
Signed-off-by: Sachin Singh <sachinkumarsingh092@gmail.com>
This commit is contained in:
Sachin Singh 2025-12-23 16:49:36 +05:30
parent ac80c90235
commit aa416c04b6
1 changed files with 5 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/v1/attention/backends/mla/common.py
View File

@ -776,7 +776,11 @@ class MLACommonMetadataBuilder(AttentionMetadataBuilder[M]):
query_start_loc = common_attn_metadata.query_start_loc
query_start_loc_cpu = common_attn_metadata.query_start_loc_cpu
seq_lens = common_attn_metadata.seq_lens
seq_lens_cpu = seq_lens.cpu()
dcp_local_seq_lens = common_attn_metadata.dcp_local_seq_lens
dcp_local_seq_lens_cpu = (
dcp_local_seq_lens.cpu() if dcp_local_seq_lens is not None else None
)
num_decodes, num_prefills, num_decode_tokens, num_prefill_tokens = (
split_decodes_and_prefills(
@ -997,6 +1001,7 @@ class MLACommonMetadataBuilder(AttentionMetadataBuilder[M]):
self.dcp_local_block_size,
)
if dcp_local_seq_lens is None:
assert dcp_local_seq_lens_cpu is not None
dcp_local_seq_lens = dcp_local_seq_lens_cpu.to(
seq_lens.device, non_blocking=True
)