Skip to content

[Attention] Fix FlashMLA metadata builder arguments for q_len > 1 #27368

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
LucasWilkinson merged 2 commits into from
Oct 23, 2025
Merged

[Attention] Fix FlashMLA metadata builder arguments for q_len > 1 #27368

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9e44596
fix num_q_tokens_per_head_k
MatthewBonanni Oct 22, 2025
4fc3912
add note
MatthewBonanni Oct 22, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion vllm/v1/attention/backends/mla/flashmla.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -120,9 +120,13 @@ def _build_decode(
num_decode_tokens: int,
dcp_tot_seq_lens_device: torch.Tensor | None,
) -> FlashMLADecodeMetadata:
query_lens_cpu = query_start_loc_cpu[1:] - query_start_loc_cpu[:-1]
# we use the max but all should be the same due to uniform length requirement
max_query_len = query_lens_cpu.max().item()
num_q_tokens_per_head_k = max_query_len * self.num_q_heads // 1
tile_scheduler_metadata, num_splits = get_mla_metadata(
seq_lens_device,
self.num_q_heads,
num_q_tokens_per_head_k,
1, # MQA for the decode path
is_fp8_kvcache=self.is_fp8_kvcache,
)
Expand Down