Skip to content

[None][fix] Fix moe backend setting #15861

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

Open
shuyixiong wants to merge 1 commit into
base: main
Choose a base branch
from
+0 −5
Open

[None][fix] Fix moe backend setting #15861

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: 0 additions & 5 deletions tensorrt_llm/_torch/modules/fused_moe/create_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -213,11 +213,6 @@ def resolve_moe_cls(
moe_cls = get_moe_cls(model_config, override_quant_config, layer_idx)

effective_quant_config = override_quant_config or model_config.quant_config
has_quant = (effective_quant_config is not None
and effective_quant_config.layer_quant_mode.has_any_quant(
exclude_kv_cache=True))
if (moe_cls == TRTLLMGenFusedMoE and not has_quant):
moe_cls = CutlassFusedMoE

# Routed-expert LoRA is supported only on CutlassFusedMoE with unquantized
# base weights. Fail loudly here rather than at runtime if the user-selected
Expand Down
Loading