[MoE Refactor] Move the shared/fused expert output sum into MoERunnerBase (#35949)

Signed-off-by: Bill Nell <bnell@redhat.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
This commit is contained in:
bnellnm
2026-04-20 12:28:46 -04:00
committed by GitHub
co-authored by Robert Shaw
parent 595562651a
commit 726efe177b
52 changed files with 325 additions and 700 deletions
-1
View File
@@ -85,7 +85,6 @@ class DbrxExperts(FusedMoE):
hidden_size=config.d_model,
intermediate_size=config.ffn_config.ffn_hidden_size,
params_dtype=params_dtype,
reduce_results=True,
renormalize=True,
quant_config=quant_config,
tp_size=get_tensor_model_parallel_world_size(),