Fix condition for expert fusion by updating the check for 'enable_deepep_moe' to include 'enable_ep_moe' in DeepseekV2ForCausalLM. sgl-project#6767

Author: xutizhou

python/sglang/srt/models/deepseek_v2.py

@@ -1688,7 +1688,12 @@ def determine_n_share_experts_fusion(
                 and torch.cuda.get_device_capability("cuda") >= (9, 0)
                 and self.config.architectures[0] == architecture
                 and self.config.n_routed_experts == 256
-                and (not global_server_args_dict["enable_deepep_moe"])
+                and (
+                    not (
+                        global_server_args_dict["enable_deepep_moe"]
+                        or global_server_args_dict["enable_ep_moe"]
+                    )
+                )
             ):
                 self.n_share_experts_fusion = self.tp_size
                 global_server_args_dict["n_share_experts_fusion"] = self.tp_size