Skip to content

[Bugfix] Fix DeepseekV32 AssertionError: num_kv_heads == 1 #33090

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
NickLucche merged 1 commit into from
Jan 27, 2026
Merged

[Bugfix] Fix DeepseekV32 AssertionError: num_kv_heads == 1 #33090

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion vllm/distributed/kv_transfer/kv_connector/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -322,7 +322,7 @@ def __post_init__(self):
# Figure out whether the first dimension of the cache is K/V
# or num_blocks. This is used to register the memory regions correctly.
kv_cache_shape = self.attn_backend.get_kv_cache_shape(
num_blocks=1, block_size=16, num_kv_heads=4, head_size=1
num_blocks=1, block_size=16, num_kv_heads=1, head_size=1
)
# Non-MLA backends caches have 5 dims [2, num_blocks, H,N,D],
# we just mock num_blocks to 1 for the dimension check below.
Expand Down
Loading