diff --git a/mindspeed_llm/mindspore/mindspore_adaptor.py b/mindspeed_llm/mindspore/mindspore_adaptor.py
index cd3992a2b5dc57d05bf72382a0ae204708ce49bc..d0d2c5015b3b455fd09ee7265c9e6de125237a85 100644
--- a/mindspeed_llm/mindspore/mindspore_adaptor.py
+++ b/mindspeed_llm/mindspore/mindspore_adaptor.py
@@ -159,6 +159,12 @@ class MindSporeAdaptation(MegatronAdaptationABC):
                                     apply_seq_aux_loss)
         MindSporeAdaptation.register('megatron.core.transformer.moe.router.TopKRouter.gating', topk_router_gating_func)
 
+        from mindspeed.mindspore.core.transformer.transformer import core_mlp_forward_wrapper
+        MindSporeAdaptation.register('megatron.core.transformer.mlp.MLP.forward',
+            core_mlp_forward_wrapper)
+
+
+
         if args.moe_fb_overlap:
             from mindspeed_llm.mindspore.tasks.models.transformer.multi_head_latent_attention import mla_forward
             MindSporeAdaptation.register('mindspeed_llm.tasks.models.transformer.multi_head_latent_attention.MultiHeadLatentAttention.forward',
@@ -267,10 +273,6 @@ class MindSporeAdaptation(MegatronAdaptationABC):
             MindSporeAdaptation.register('mindspeed.core.transformer.moe.comm_utils.async_all_to_all',
                         async_all_to_all)
             
-            from mindspeed.mindspore.core.transformer.transformer import core_mlp_forward_wrapper
-            MindSporeAdaptation.register('megatron.core.transformer.mlp.MLP.forward',
-                core_mlp_forward_wrapper)
-
             from mindspeed.mindspore.core.pipeline_parallel.fb_overlap.modules.token_dispatcher import alltoall_token_perm1, overlap_stream
             MindSporeAdaptation.register('mindspeed.core.pipeline_parallel.fb_overlap.modules.token_dispatcher.alltoall_token_perm1',
                 alltoall_token_perm1)           
@@ -331,7 +333,10 @@ class MindSporeAdaptation(MegatronAdaptationABC):
 
         if args.gemm_gradient_accumulation_fusion:
             from mindspeed.mindspore.ops.npu_groupmatmul_add import npu_groupmatmul_add_fp32
-            MindSporeAdaptation.register('mindspeed.ops.npu_groupmatmul_add', npu_groupmatmul_add_fp32)
+            MindSporeAdaptation.register('mindspeed.ops.npu_groupmatmul_add.npu_groupmatmul_add_fp32', npu_groupmatmul_add_fp32)
+        from mindspeed.mindspore.ops.npu_matmul_add import npu_matmul_add_fp32
+        MindSporeAdaptation.register('fused_weight_gradient_mlp_cuda.wgrad_gemm_accum_fp32', npu_matmul_add_fp32)
+        MindSporeAdaptation.register('mindspeed.ops.npu_matmul_add.npu_matmul_add_fp32', npu_matmul_add_fp32)
 
         if args.use_moba_attn:
             from mindspeed_llm.mindspore.core.transformer.dot_product_attention import flash_attention_forward