mindspore 2.6.0rc1__cp311-cp311-win_amd64.whl → 2.7.0__cp311-cp311-win_amd64.whl

mindspore/ops/_vmap/vmap_array_ops.py

@@ -214,6 +214,31 @@ def get_transpose_vmap_rule(prim, axis_size):
     return vmap_rule
 
 
+@vmap_rules_getters.register("TransposeExtView")
+def get_transpose_ext_vmap_rule(prim, axis_size):
+    """VmapRule for `TransposeExtView` operation."""
+    if isinstance(prim, str):
+        prim = Primitive(prim)
+
+    def vmap_rule(x_bdim, dim1_bdim, dim2_bdim):
+        is_all_none, result = vmap_general_preprocess(prim, x_bdim, dim1_bdim, dim2_bdim)
+        if is_all_none:
+            return result
+
+        x, dim = x_bdim
+        dim1, dim1_dim = dim1_bdim
+        dim2, dim2_dim = dim2_bdim
+        if dim1_dim is not None or dim2_dim is not None:
+            _raise_value_error("The source axis of dim1_dim and dim2_dim in `TransposeExtView` must be None, "
+                               "but got {} and {}.".format(dim1_dim, dim2_dim))
+        batch_dim1 = dim1 if dim1 < dim else dim1 + 1
+        batch_dim2 = dim2 if dim2 < dim else dim2 + 1
+        out = prim(x, batch_dim1, batch_dim2)
+        return out, dim
+
+    return vmap_rule
+
+
 @vmap_rules_getters.register("Tile")
 def get_tile_vmap_rule(prim, axis_size):
     """VmapRule for `P.Tile` operation."""
@@ -1502,14 +1527,13 @@ def get_meshgrid_vmap_rule(prim, axis_size):
 
         if not isinstance(inputs_bdim, (tuple)):
             _raise_value_error("The inputs of P.Meshgrid is not tuple.")
-        args = inputs_bdim
-        if len(args) <= 1:
+        if len(inputs_bdim) <= 1:
             _raise_value_error(
                 "The input number of P.Meshgrid must be greater than 1.")
 
         output_shape = []
         ones_shape = []
-        for each_arg in args:
+        for each_arg in inputs_bdim:
             x, bdim = each_arg
             if bdim is None:
                 _raise_value_error(
@@ -1523,22 +1547,16 @@ def get_meshgrid_vmap_rule(prim, axis_size):
         output_shape.insert(0, axis_size)
         ones_shape.insert(0, axis_size)
 
-        indexing, _ = indexing_bdim
-
-        if indexing == Indexing.xy.value:
+        if indexing_bdim[0] == Indexing.xy.value:
             output_shape[1], output_shape[2] = output_shape[2], output_shape[1]
-        shape = tuple(output_shape)
-
-        input_0, _ = args[0]
-        dtype = F.dtype(input_0)
-        ones_tensor = F.fill(dtype, shape, 1)
+        ones_tensor = F.fill(F.dtype(inputs_bdim[0][0]), tuple(output_shape), 1)
 
         index = 0
         vals_out_tuple = ()
-        for each_arg in args:
+        for each_arg in inputs_bdim:
             x, bdim = each_arg
             x = _bdim_at_front(x, bdim, axis_size)
-            shape_index = (1 - index) if (index <= 1 and indexing == Indexing.xy.value) else index
+            shape_index = (1 - index) if (index <= 1 and indexing_bdim[0] == Indexing.xy.value) else index
             ones_shape[shape_index + 1] = output_shape[shape_index + 1]
             x = P.Reshape()(x, tuple(ones_shape))
             output = P.Mul()(x, ones_tensor)

mindspore/ops/_vmap/vmap_nn_ops.py

@@ -2068,14 +2068,15 @@ def get_sparse_apply_adagrad_vmap_rule(prim, axis_size):
         indices, indices_dim = indices_bdim
         if var_dim is None:
             if any(dim is not None for dim in [accum_dim, grad_dim, indices_dim]):
-                ValueError("The source axis of `var` is None, but the source "
-                           "axis of `accum/grad/indices` is not None. The execution order of "
-                           "operator `{}` cannot be guaranteed.".format(prim_name))
+                _raise_value_error("The source axis of `var` is None, but the source "
+                                   "axis of `accum/grad/indices` is not None. The execution "
+                                   "order of operator `{}` cannot be guaranteed.".format(prim_name))
             var, accum = prim(var, accum, grad, indices, u_monad)
             return (var, None), (accum, None)
         if var_dim != 0 or accum_dim != var_dim:
-            ValueError("For `{}`, the source axis of `var` must be equal to `accum`, and not equal to 0, "
-                       "but got the source axis of `var`: {}, `accum`: {}.".format(prim_name, var_dim, accum_dim))
+            _raise_value_error("For `{}`, the source axis of `var` must be equal to `accum`, "
+                               "and not equal to 0, but got the source axis of `var`: {}, "
+                               "`accum`: {}.".format(prim_name, var_dim, accum_dim))
 
         grad = _bdim_at_front(grad, grad_dim, axis_size)
         indices = _bdim_at_front(indices, indices_dim, axis_size)
@@ -2094,27 +2095,18 @@ def get_sparse_apply_ftrl_vmap_rule(prim, axis_size):
     else:
         batch_rank = 1
 
-    prim_name = prim.name
     batch_prim = _vmap_clone_prim(prim)
     batch_prim.add_prim_attr('batch_rank', batch_rank)
 
     def vmap_rule(var_bdim, accum_bdim, linear_bdim, grad_bdim, indices_bdim, u_monad):
         var, var_dim = var_bdim
-        accum, accum_dim = accum_bdim
-        linear, linear_dim = linear_bdim
+        accum, _ = accum_bdim
+        linear, _ = linear_bdim
         grad, grad_dim = grad_bdim
         indices, indices_dim = indices_bdim
         if var_dim is None:
-            if any(dim is not None for dim in [accum_dim, linear_dim, grad_dim, indices_dim]):
-                ValueError("The source axis of `var` is None, but the source "
-                           "axis of `accum/linear/grad/indices` is not None. The execution order of "
-                           "operator `{}` cannot be guaranteed.".format(prim_name))
             var, accum, linear = prim(var, accum, linear, grad, indices, u_monad)
             return (var, None), (accum, None), (linear, None)
-        if var_dim != 0 or accum_dim != var_dim or linear_dim != var_dim:
-            ValueError("For `{}`, the source axis of `var`, `accum` and `linear` must be equal, and "
-                       "not equal to 0, but got the source axis of `var`: {}, `accum`: {}, "
-                       "`linear`:{}.".format(prim_name, var_dim, accum_dim, linear_dim))
 
         grad = _bdim_at_front(grad, grad_dim, axis_size)
         indices = _bdim_at_front(indices, indices_dim, axis_size)

mindspore/ops/auto_generate/cpp_create_prim_instance_helper.py

@@ -53,9 +53,9 @@ op_args_default_value = {
     "Baddbmm": {"beta": 1, "alpha": 1},
     "BatchMatMul": {"transpose_a": False, "transpose_b": False},
     "BatchNormElemt": {"weight": None, "bias": None, "mean": None, "invstd": None, "eps": 1e-5},
-    "BatchNormExt": {"running_mean": None, "runnning_var": None, "training": False, "momentum": 0.1, "epsilon": 1e-5},
+    "BatchNormExt": {"weight": None, "bias": None, "running_mean": None, "runnning_var": None, "training": False, "momentum": 0.1, "epsilon": 1e-5},
     "BatchNormGatherStatsWithCounts": {"running_mean": None, "running_var": None, "momentum": 1e-1, "eps": 1e-5, "counts": None},
-    "BatchNormGradExt": {"running_mean": None, "running_var": None, "saved_mean": None, "saved_rstd": None, "training": False, "eps": 1e-5, "output_mask": (1, 1, 1)},
+    "BatchNormGradExt": {"weight": None, "running_mean": None, "running_var": None, "saved_mean": None, "saved_rstd": None, "training": False, "eps": 1e-5, "output_mask": (1, 1, 1)},
     "BatchNormGradGrad": {"is_training": False, "epsilon": 1e-5, "data_format": 'NCHW'},
     "BatchNormGrad": {"is_training": False, "epsilon": 1e-5, "data_format": 'NCHW'},
     "BatchNormGradWithActivation": {"is_training": False, "epsilon": 1e-5, "data_format": 'NCHW'},
@@ -76,6 +76,7 @@ op_args_default_value = {
     "CholeskyInverse": {"upper": False},
     "Cholesky": {"upper": False},
     "Chunk": {"dim": 0},
+    "ChunkView": {"dim": 0},
     "ClampScalar": {"min": None, "max": None},
     "ClampTensor": {"min": None, "max": None},
     "Col2ImExt": {"dilation": 1, "padding": 0, "stride": 1},
@@ -95,6 +96,8 @@ op_args_default_value = {
     "ConvTranspose2D": {"bias": None, "stride": 1, "padding": 0, "output_padding": 0, "groups": 1, "dilation": 1},
     "Correlate": {"pad_mode": 'valid'},
     "CountNonZero": {"dim": None},
+    "CrossEntropyLossGrad": {"weight": None, "grad_zloss": None, "lse_for_zloss": None, "reduction": 'mean', "ignore_index": -100, "label_smoothing": 0.0, "lse_square_scale_for_zloss": 0.0},
+    "CrossEntropyLoss": {"weight": None, "reduction": 'mean', "ignore_index": -100, "label_smoothing": 0.0, "lse_square_scale_for_zloss": 0.0, "return_zloss": False},
     "Cross": {"dim": -65530},
     "CumsumExt": {"dtype": None},
     "CumProd": {"exclusive": False, "reverse": False},
@@ -103,6 +106,7 @@ op_args_default_value = {
     "DCT": {"type": 2, "n": None, "axis": -1, "norm": None},
     "Dense": {"bias": None},
     "Diagonal": {"offset": 0, "dim1": 0, "dim2": 1},
+    "DiagonalView": {"offset": 0, "dim1": 0, "dim2": 1},
     "DiagExt": {"diagonal": 0},
     "DivMods": {"rounding_mode": None},
     "DivMod": {"rounding_mode": None},
@@ -140,13 +144,12 @@ op_args_default_value = {
     "FlashAttentionScore": {"real_shift": None, "drop_mask": None, "padding_mask": None, "attn_mask": None, "prefix": None, "actual_seq_qlen": None, "actual_seq_kvlen": None, "keep_prob": 1.0, "scale_value": 1.0, "pre_tokens": 2147483647, "next_tokens": 2147483647, "inner_precise": 0, "input_layout": 'BSH', "sparse_mode": 0},
     "FlattenExt": {"start_dim": 0, "end_dim": -1},
     "FullLike": {"dtype": None},
+    "FusedAddTopKDiv": {"activate_type": 0, "is_norm": True, "scale": 2.5, "mapping_num": None, "mapping_table": None, "enable_expert_mapping": False},
     "Gather": {"batch_dims": 0},
     "GeluExt": {"approximate": 'none'},
     "GeluGradExt": {"approximate": 'none'},
     "GenerateEodMaskV2": {"start": 0, "steps": 1, "error_mode": 'cycle', "flip_mode": 'bitflip', "multiply_factor": 0.0, "bit_pos": 0, "flip_probability": 0.0},
     "GLU": {"axis": -1},
-    "GmmBackward": {"group_list": None},
-    "GmmV2Backward": {"group_list": None, "group_list_type": 0},
     "GridSampler2DGrad": {"interpolation_mode": 'bilinear', "padding_mode": 'zeros', "align_corners": False, "output_mask": (1, 1)},
     "GridSampler2D": {"interpolation_mode": 'bilinear', "padding_mode": 'zeros', "align_corners": False},
     "GridSampler3DGrad": {"interpolation_mode": 'bilinear', "padding_mode": 'zeros', "align_corners": False, "output_mask": (1, 1)},
@@ -175,11 +178,13 @@ op_args_default_value = {
     "IncreFlashAttention": {"attn_mask": None, "actual_seq_lengths": None, "pse_shift": None, "dequant_scale1": None, "quant_scale1": None, "dequant_scale2": None, "quant_scale2": None, "quant_offset2": None, "antiquant_scale": None, "antiquant_offset": None, "block_table": None, "kv_padding_size": None, "num_heads": 1, "input_layout": 'BSH', "scale_value": 1.0, "num_key_value_heads": 0, "block_size": 0, "inner_precise": 1},
     "IndexAddExt": {"alpha": 1},
     "InnerInplaceIndexPut": {"accumulate": False},
+    "InnerMoeTokenUnpermute": {"probs": None, "padded_mode": False, "restore_shape": None},
     "InplaceAddmm": {"beta": 1, "alpha": 1},
     "InplaceAddsExt": {"alpha": 1},
     "InplaceAddExt": {"alpha": 1},
     "InplaceClampScalar": {"min": None, "max": None},
     "InplaceClampTensor": {"min": None, "max": None},
+    "InplaceCopy": {"non_blocking": False},
     "InplaceDivMods": {"rounding_mode": None},
     "InplaceDivMod": {"rounding_mode": None},
     "InplaceElu": {"alpha": 1.0},
@@ -237,15 +242,19 @@ op_args_default_value = {
     "Meshgrid": {"indexing": 'xy'},
     "MinimumGrad": {"grad_x": True, "grad_y": True},
     "MinDim": {"keepdim": False},
+    "Mla": {"attn_mask": None, "deq_scale_qk": None, "deq_scale_pv": None, "q_seq_lens": None, "context_lens": None, "head_num": 32, "scale_value": 0.0, "kv_head_num": 1, "mask_mode": 'MASK_NONE', "is_ring": 0},
+    "MlaPreprocess": {"param_cache_mode": 0},
+    "MoeDistributeCombine": {"tp_send_counts": None, "x_active_mask": None, "activate_scale": None, "weight_scale": None, "group_list": None, "expand_scales": None, "group_ep": None, "group_tp": None, "tp_world_size": 0, "tp_rank_id": 0, "expert_shard_type": 0, "shared_expert_num": 0, "shared_export_rank_num": 0, "global_bs": 0, "out_dtype": 0, "common_quant_mode": 0, "group_list_type": 0},
+    "MoeDistributeDispatch": {"expert_scales": None, "scales": None, "x_active_mask": None, "group_ep": None, "group_tp": None, "tp_world_size": 0, "tp_rank_id": 0, "expert_shard_type": 0, "shared_expert_num": 0, "shared_expert_rank_num": 0, "quant_mode": 0, "global_bs": 0, "expert_token_nums_type": 0},
     "MoeTokenPermuteGrad": {"num_topk": 1, "padded_mode": False},
     "MoeTokenPermute": {"num_out_tokens": None, "padded_mode": False},
     "MoeTokenUnpermuteGrad": {"probs": None, "padded_mode": False, "restore_shape": None},
-    "MoeTokenUnpermute": {"probs": None, "padded_mode": False, "restore_shape": None},
     "MSELossExt": {"reduction": 'mean'},
     "MSELossGradExt": {"reduction": 'mean'},
     "Nansum": {"dim": None, "keepdim": False, "dtype": None},
     "NanToNum": {"nan": None, "posinf": None, "neginf": None},
     "NewEmpty": {"dtype": None, "device": None},
+    "NewFull": {"dtype": None},
     "NewOnes": {"dtype": None},
     "NewZeros": {"dtype": None},
     "NLLLoss2d": {"reduction": 'mean', "ignore_index": -100},
@@ -257,10 +266,9 @@ op_args_default_value = {
     "OneHotExt": {"axis": -1},
     "OneHot": {"axis": -1},
     "PagedAttentionMask": {"antiquant_scale": None, "antiquant_offset": None, "alibi_mask": None, "kv_cache_quant_mode": 'DEFAULT'},
-    "PagedAttention": {"antiquant_scale": None, "antiquant_offset": None, "attn_mask": None, "q_seq_lens": None, "alibi_mask": None, "kv_cache_quant_mode": 'DEFAULT', "mask_mode": 'MASK_DEFAULT', "mla_v_dim": 0},
+    "PagedAttention": {"value_cache": None, "block_tables": None, "context_lens": None, "antiquant_scale": None, "antiquant_offset": None, "attn_mask": None, "q_seq_lens": None, "alibi_mask": None, "kv_cache_quant_mode": 'DEFAULT', "mask_mode": 'MASK_DEFAULT', "mla_v_dim": 0},
     "ProdExt": {"dim": None, "keepdim": False, "dtype": None},
     "PromptFlashAttention": {"attn_mask": None, "actual_seq_lengths": None, "actual_seq_lengths_kv": None, "pse_shift": None, "deq_scale1": None, "quant_scale1": None, "deq_scale2": None, "quant_scale2": None, "quant_offset2": None, "num_heads": 1, "scale_value": 1.0, "pre_tokens": 2147483647, "next_tokens": 0, "input_layout": 'BSH', "num_key_value_heads": 0, "sparse_mode": 0, "inner_precise": 1},
-    "PromptKVCache": {"align_mode": 'LEFT'},
     "Qr": {"full_matrices": False},
     "RandIntLike": {"dtype": None},
     "RandInt": {"dtype": None},
@@ -297,6 +305,7 @@ op_args_default_value = {
     "RFFTFreq": {"d": 1.0, "dtype": None},
     "RFFTN": {"s": None, "dim": None, "norm": None},
     "RFFT": {"n": None, "dim": -1, "norm": None},
+    "RingAttentionUpdate": {"actual_seq_qlen": None, "layout": 'SBH'},
     "RmsNorm": {"epsilon": 1e-6},
     "Roll": {"dims": None},
     "RotaryPositionEmbeddingGrad": {"dx": None, "mode": 0},
@@ -326,7 +335,9 @@ op_args_default_value = {
     "SpeedFusionAttention": {"pse": None, "padding_mask": None, "atten_mask": None, "scale": 1.0, "keep_prob": 1.0, "pre_tokens": 2147483647, "next_tokens": 2147483647, "inner_precise": 0, "prefix": None, "actual_seq_qlen": None, "actual_seq_kvlen": None, "sparse_mode": 0, "gen_mask_parallel": True, "sync": False, "pse_type": 1, "q_start_idx": None, "kv_start_idx": None},
     "Split": {"axis": 0, "output_num": 1},
     "SplitTensor": {"dim": 0},
+    "SplitTensorView": {"dim": 0},
     "SplitWithSize": {"dim": 0},
+    "SplitWithSizeView": {"dim": 0},
     "Squeeze": {"axis": ()},
     "StackExt": {"dim": 0},
     "StdMean": {"dim": None, "correction": 1, "keepdim": False},
@@ -341,6 +352,7 @@ op_args_default_value = {
     "TensorScatterElements": {"axis": 0, "reduce": 'none'},
     "TopKRouter": {"drop_type": 0},
     "TopkExt": {"dim": -1, "largest": True, "sorted": True},
+    "TopPRouter": {"drop_type": 0, "threshold": 0.0, "router_prob": 0.0},
     "TraceV2Grad": {"offset": 0, "axis1": 1, "axis2": 0},
     "TraceV2": {"offset": 0, "axis1": 1, "axis2": 0, "dtype": None},
     "TriangularSolve": {"upper": True, "transpose": False, "unitriangular": False},
@@ -349,7 +361,7 @@ op_args_default_value = {
     "TupleToTensor": {"dtype": None},
     "Unique2": {"sorted": True, "return_inverse": False, "return_counts": False},
     "UniqueConsecutive": {"return_inverse": False, "return_counts": False, "dim": None},
-    "UnstackExt": {"dim": 0},
+    "UnstackExtView": {"dim": 0},
     "UpsampleBicubic2DGrad": {"output_size": None, "scales": None, "align_corners": False},
     "UpsampleBicubic2D": {"output_size": None, "scales": None, "align_corners": False},
     "UpsampleBilinear2DGrad": {"output_size": None, "scales": None, "align_corners": False},
@@ -370,6 +382,7 @@ op_args_default_value = {
     "Zeros": {"dtype": None},
     "AddRmsNormDynamicQuant": {"smooth_scale2": None, "epsilon": 1e-5},
     "AddRmsNormQuantV2": {"epsilon": 1e-5},
+    "DynamicNTK": {"dtype": mstype.float16},
     "DynamicQuantExt": {"smooth_scales": None},
     "FusedInferAttentionScore": {"pse_shift": None, "attn_mask": None, "actual_seq_lengths": None, "actual_seq_lengths_kv": None, "dequant_scale1": None, "quant_scale1": None, "dequant_scale2": None, "quant_scale2": None, "quant_offset2": None, "antiquant_scale": None, "antiquant_offset": None, "block_table": None, "query_padding_size": None, "kv_padding_size": None, "key_antiquant_scale": None, "key_antiquant_offset": None, "value_antiquant_scale": None, "value_antiquant_offset": None, "key_shared_prefix": None, "value_shared_prefix": None, "actual_shared_prefix_len": None, "num_heads": 1, "scale_value": 1.0, "pre_tokens": 2147483647, "next_tokens": 2147483647, "input_layout": 'BSH', "num_key_value_heads": 0, "sparse_mode": 0, "inner_precise": 1, "block_size": 0, "antiquant_mode": 0, "softmax_lse_flag": False, "key_antiquant_mode": 0, "value_antiquant_mode": 0},
     "GroupedMatmul": {"bias": None, "scale": None, "offset": None, "antiquant_scale": None, "antiquant_offset": None, "group_list": None, "split_item": 0, "group_type": -1, "transpose_a": False, "transpose_b": False},
@@ -379,25 +392,35 @@ op_args_default_value = {
     "MatmulAllReduceAddRmsNorm": {"reduce_op": 'sum', "comm_turn": 0, "stream_mode": 1},
     "MoeFinalizeRouting": {"x2": None, "bias": None, "scales": None, "expanded_row_idx": None, "expanded_expert_idx": None},
     "MoeGatingTopKSoftmax": {"finished": None, "k": 1},
+    "MoeInitRoutingQuantV2": {"scale": None, "offset": None},
     "QuantBatchMatmul": {"offset": None, "bias": None, "pertokenScaleOptional": None, "transpose_x1": False, "transpose_x2": False, "dtype": mstype.float16},
+    "QuantMatmul": {"offset": None, "pertoken_scale": None, "bias": None, "output_dtype": None, "x1_dtype": None, "x2_dtype": None, "pertoken_scale_dtype": None, "scale_dtype": None, "group_sizes": None},
     "QuantV2": {"sqrt_mode": False, "rounding_mode": 'ROUND', "dst_type": mstype.int8},
     "RmsNormQuant": {"beta": None, "epsilon": 1e-6},
+    "SwiGLUDynamicQuant": {"smooth_scale": None},
     "TransposeBatchMatmulTranspose": {"transpose_a": False, "transpose_b": False},
     "WeightQuantBatchMatmul": {"antiquant_offset": None, "quant_scale": None, "quant_offset": None, "bias": None, "transpose_x": False, "transpose_weight": False, "antiquant_group_size": 0},
+    "AnyExt": {"keepdim": False},
+    "FuncMaxPool2D": {"stride": None, "padding": 0, "dilation": (1, 1), "ceil_mode": False, "return_indices": False},
+    "GmmBackwardFusion": {"group_list": None, "group_list_type": 0},
+    "GmmBackward": {"group_list": None, "group_list_type": 0},
+    "Gmm": {"bias": None, "group_list": None, "group_type": 0, "group_list_type": 0},
+    "GmmV2BackwardFusion": {"group_list": None, "group_list_type": 0},
+    "GmmV2Backward": {"group_list": None, "group_list_type": 0},
+    "GmmV2": {"bias": None, "group_list": None, "group_type": 0, "group_list_type": 0},
+    "MoeTokenUnpermute": {"probs": None, "padded_mode": False, "restore_shape": None},
     "DeprecatedAddbmm": {"beta": 1, "alpha": 1},
     "DeprecatedAddmm": {"beta": 1, "alpha": 1},
     "DeprecatedAddmv": {"beta": 1, "alpha": 1},
     "DeprecatedReduceAll": {"dim": None, "keepdim": False},
     "DeprecatedAllclose": {"rtol": 1e-05, "atol": 1e-08, "equal_nan": False},
     "DeprecatedReduceAny": {"axis": None, "keep_dims": False},
-    "DeprecatedAny": {"dim": None, "keepdim": False},
     "DeprecatedArgmax": {"axis": -1, "keepdims": False},
     "DeprecatedArgmin": {"axis": None, "keepdims": False},
     "DeprecatedArgsort": {"axis": -1, "descending": False},
     "DeprecatedBaddbmm": {"beta": 1, "alpha": 1},
     "DeprecatedBincount": {"weights": None, "minlength": 0},
     "DeprecatedChunk": {"axis": 0},
-    "DeprecatedInplaceCopy": {"non_blocking": False},
     "DeprecatedCountNonzero": {"axis": (), "keep_dims": False, "dtype": None},
     "DeprecatedCumsum": {"axis": None, "dtype": None},
     "DeprecatedDiv": {"rounding_mode": None},
@@ -437,7 +460,9 @@ op_labels = {
     "AssignSub": {"side_effect_mem": True},
     "BatchNormElemt": {"side_effect_mem": True},
     "BatchNormGatherStatsWithCounts": {"side_effect_mem": True},
-    "DecoderKVCache": {"side_effect_mem": True},
+    "BroadcastToView": {"side_effect_mem": True},
+    "ChunkView": {"side_effect_mem": True},
+    "DiagonalView": {"side_effect_mem": True},
     "DistCommAllReduce": {"side_effect_mem": True},
     "DistCommReduce": {"side_effect_mem": True},
     "DropoutExt": {"side_effect_hidden": True},
@@ -451,12 +476,15 @@ op_labels = {
     "EmbeddingApplySgd": {"_process_node_engine_id": 'PS'},
     "Embedding": {"side_effect_mem": True},
     "EmbeddingTableEvict": {"_process_node_engine_id": 'PS'},
+    "ExpandDimsView": {"side_effect_mem": True},
     "Generator": {"side_effect_mem": True},
     "GroupTopk": {"side_effect_mem": True},
     "InnerInplaceIndexPut": {"side_effect_mem": True},
     "InplaceAddmm": {"side_effect_mem": True},
     "InplaceAddsExt": {"side_effect_mem": True},
     "InplaceAddExt": {"side_effect_mem": True},
+    "InplaceBernoulliScalar": {"side_effect_mem": True},
+    "InplaceBernoulliTensor": {"side_effect_mem": True},
     "InplaceCopy": {"side_effect_mem": True},
     "InplaceDivMods": {"side_effect_mem": True},
     "InplaceDivMod": {"side_effect_mem": True},
@@ -478,28 +506,41 @@ op_labels = {
     "InplaceLog": {"side_effect_mem": True},
     "InplaceMaskedFillScalar": {"side_effect_mem": True},
     "InplaceMaskedFillTensor": {"side_effect_mem": True},
+    "InplaceMatmulAdd": {"side_effect_mem": True},
     "InplaceMuls": {"side_effect_mem": True},
     "InplaceMul": {"side_effect_mem": True},
     "InplaceNormal": {"side_effect_mem": True},
     "InplacePut": {"side_effect_mem": True},
     "InplaceRandom": {"side_effect_mem": True},
     "InplaceReLU": {"side_effect_mem": True},
+    "InplaceRemainderTensorScalar": {"side_effect_mem": True},
+    "InplaceRemainderTensorTensor": {"side_effect_mem": True},
     "InplaceScatterAdd": {"side_effect_mem": True},
     "InplaceScatterSrc": {"side_effect_mem": True},
     "InplaceScatterSrcReduce": {"side_effect_mem": True},
     "InplaceScatterValue": {"side_effect_mem": True},
     "InplaceScatterValueReduce": {"side_effect_mem": True},
+    "InplaceSiLU": {"side_effect_mem": True},
     "InplaceSubExt": {"side_effect_mem": True},
     "InplaceSubScalar": {"side_effect_mem": True},
     "InplaceTanh": {"side_effect_mem": True},
     "InplaceThreshold": {"side_effect_mem": True},
     "InplaceUniform": {"side_effect_mem": True},
+    "KvScaleCache": {"side_effect_mem": True},
     "Log": {"cust_aicpu": 'Log', "base": -1.0, "scale": 1.0, "shift": 0.0},
-    "PromptKVCache": {"side_effect_mem": True},
+    "MlaPreprocess": {"side_effect_mem": True},
+    "NarrowView": {"side_effect_mem": True},
     "ReshapeAndCache": {"side_effect_mem": True},
     "ResizeD": {"mode": 'linear'},
+    "SelectExtView": {"side_effect_mem": True},
     "SilentCheckV2": {"side_effect_mem": True},
     "SilentCheckV3": {"side_effect_mem": True},
+    "SliceExtView": {"side_effect_mem": True},
+    "SplitTensorView": {"side_effect_mem": True},
+    "SplitWithSizeView": {"side_effect_mem": True},
+    "TransposeExtView": {"side_effect_mem": True},
+    "TransposeView": {"side_effect_mem": True},
+    "UnstackExtView": {"side_effect_mem": True},
     "KVCacheScatterUpdate": {"side_effect_mem": True},
-    "DeprecatedInplaceCopy": {"side_effect_mem": True},
+    "InplaceExponential": {"side_effect_mem": True},
 }