Upload modeling_deepseek.py

modeling_deepseek.py

@@ -530,8 +530,8 @@ class DeepseekV3MoE(nn.Module):
             y = self.moe_infer(hidden_states, topk_idx, topk_weight).view(*orig_shape)
             # if self.config.n_shared_experts is not None:
             #     y = y + self.shared_experts(identity)
-        return y
-
+            return y
+        return 0
     @torch.no_grad()
     def moe_infer(self, x, topk_ids, topk_weight):
         cnts = topk_ids.new_zeros((topk_ids.shape[0], len(self.experts)))