Merge pull request #86 from xhedit/main

typo fix: KMisrtal -> KMistral

ktransformers/operators/experts.py

@@ -727,7 +727,7 @@ class KDeepseekV2MoE(BaseInjectedModule, DeepseekV2MoE):
         )
         return final_out
 
-class KMisrtalSparseMoEBlock(BaseInjectedModule, MixtralSparseMoeBlock):
+class KMistralSparseMoEBlock(BaseInjectedModule, MixtralSparseMoeBlock):
     
     def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         """ """

ktransformers/optimize/optimize_rules/Mixtral.yaml

@@ -19,7 +19,7 @@
     name: "^model\\.layers\\..*\\.block_sparse_moe$"
     class: ktransformers.models.modeling_mixtral.MixtralSparseMoeBlock
   replace: 
-    class: ktransformers.operators.experts.KMisrtalSparseMoEBlock
+    class: ktransformers.operators.experts.KMistralSparseMoEBlock
 - match:
     name: "^model\\.layers\\..*\\.block_sparse_moe\\.experts$"
   replace: