Skip to content

[Fix] Remove divisibility requirement between num_kv_heads and tp_size in bailing_moe #26876

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
youkaichao merged 2 commits into from
Oct 15, 2025
Merged

[Fix] Remove divisibility requirement between num_kv_heads and tp_size in bailing_moe #26876

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
aa9433c
fix(num_kv_heads): correct num_kv_heads calculation
ant-yy Oct 15, 2025
867667e
Merge branch 'main' into ling_v2_fix
ant-yy Oct 15, 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
3 changes: 1 addition & 2 deletions vllm/model_executor/models/bailing_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,13 +86,12 @@ def __init__(
tp_size = get_tensor_model_parallel_world_size()

assert self.total_num_heads % tp_size == 0
assert self.total_kv_heads % tp_size == 0
assert self.total_num_heads >= self.total_kv_heads

self.num_heads = self.total_num_heads // tp_size
self.head_dim = config.head_dim or (self.hidden_size // self.total_num_heads)
self.q_size_per_rank = self.head_dim * self.num_heads
self.num_kv_heads = self.total_kv_heads // tp_size
self.num_kv_heads = max(1, self.total_kv_heads // tp_size)
Comment thread
ant-yy marked this conversation as resolved.
self.kv_size_per_rank = self.num_kv_heads * self.head_dim
self.scale = self.head_dim**-0.5
self.use_qk_norm = getattr(config, "use_qk_norm", False)
Expand Down