Skip to content

[Draft] npu_moe_gating_top_k support renorm=1 #4

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

Draft
ChenxiQ wants to merge 1 commit into
base: br_moe_init_routing_custom
Choose a base branch
from
Draft

[Draft] npu_moe_gating_top_k support renorm=1 #4

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
5 changes: 1 addition & 4 deletions vllm_ascend/ops/fused_moe/experts_selector.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -220,15 +220,12 @@ def _select_experts_with_fusion_ops(
k_group=topk_group,
group_count=num_expert_group,
group_select_mode=1, # 0: the maximum in the group; 1: topk2.sum(fix)
renorm=0, # 0: softmax->topk(fix); 1: topk->softmax
renorm=1, # 0: softmax->topk(fix); 1: topk->softmax
norm_type=norm_type, # 0: softmax; 1: sigmoid
# out_flag=False, # todo new api; should the third output be output
# y2_flag=False, # old api; should the third output be output
routed_scaling_factor=1,
eps=float(1e-20))
if scoring_func == "softmax":
topk_weights = _renormalize_topk_weights(topk_weights, renormalize)

return topk_weights, topk_ids


Expand Down