serpdotai
/

sparsetral-16x7B-v1

Text Generation

Inference Endpoints

Model card Files Files and versions Community

francislabounty commited on Jan 30

Commit

c4e9636

•

1 Parent(s): 90a088c

Update configuration_sparsetral.py

Files changed (1) hide show

configuration_sparsetral.py +1 -1

configuration_sparsetral.py CHANGED Viewed

@@ -125,7 +125,7 @@ class SparsetralConfig(PretrainedConfig):
         moe_scaling=1.0,
         num_experts=16,
         topk=4,
-        output_router_logits=True,
         adapter_dim=64,
         **kwargs,
     ):

         moe_scaling=1.0,
         num_experts=16,
         topk=4,
+        output_router_logits=False,
         adapter_dim=64,
         **kwargs,
     ):