Thanks for sending an issue! Please fill in the following template to help quickly solve your problem.

Describe the current behavior / 问题描述 (Mandatory / 必填)

当前日志打印的shape是转换后的nz格式，但是shape拦截是根据原始规格实现的，打印上存在歧义。

Environment / 环境信息 (Mandatory / 必填)

Ascend 950

Steps to reproduce the issue / 重现步骤 (Mandatory / 必填)

构造权重shape为[1000001, 1, 1, 1]的用例，触发shape规格拦截打印。

Describe the expected behavior / 预期结果 (Mandatory / 必填)

期望报incorrect shape时打印原始shape。

Related log / screenshot / 日志 / 截图 (Mandatory / 必填)

[INFO] GE(1572820,atc.bin):2026-07-24-09:58:20.870.973 [error_manager.cc:1071]1572820 GetRawErrorMessages:current work_stream_id:157283572820
[PID: 1572820] 2026-07-24-09:58:20.575.457 Invalid_Argument_Tensor_Shape(EZ0009): Parameter filter of Conv2DV2 has incorrect shape [[1000001,1,1,1]]. Reason: Shape[0] of this parameter must be within the range [1, 1000000].
        Solution: Check whether the shape of input/output tensor is correct.
TraceBack (most recent call last):
        Op Conv2DV2 tiling failed[FUNC:TbeOptilingPyInterfaceNew][FILE:op_tiling_py.cc][LINE:1955][FUNC:TeInnerErrMessageReport][FILE:common_utils.cc][LINE:374]
        TbeOpTilingPyInterfaceOld failed.[FUNC:DoOpTilingForCompile][FILE:op_tiling_py.cc][LINE:2047][FUNC:TeInnerErrMessageReport][FILE:common_utils.cc][LINE:374]
        Compile operator failed, cause: Template constraint, detailed information: Tiling func of op_type Conv2DV2 failed, failure details:
Compile_info: {'device_id': None, '_cube_vector_core_type': 'AiCore'}
Inputs: {'shape': (1, 1, 1, 1), 'ori_shape': (1, 1, 1, 1), 'format': 'NCHW', 'sub_format': 0, 'ori_format': 'NCHW', 'dtype': 'float16', 'addr_type': 0, 'total_shape': [1, 1, 1, 1], 'slice_offset': (), 'L1_addr_offset': 0, 'L1_fusion_type': -1, 'L1_workspace_size': -1, 'valid_shape': (), 'split_index': 0, 'atomic_type': '', 'input_c_values': 1, 'range': [[1, 1], [1, 1], [1, 1], [1, 1]], 'param_name': 'x_in__'}
{'shape': (1, 62501, 16, 16), 'ori_shape': (1000001, 1, 1, 1), 'format': 'FRACTAL_Z_C04', 'sub_format': 0, 'ori_format': 'NCHW', 'dtype': 'float16', 'addr_type': 0, 'total_shape': [1, 62501, 16, 16], 'slice_offset': (), 'L1_addr_offset': 0, 'L1_fusion_type': -1, 'L1_workspace_size': -1, 'valid_shape': (), 'split_index': 0, 'atomic_type': '', 'input_c_values': 1, 'range': [[1, 1], [62501, 62501], [16, 16], [16, 16]], 'param_name': 'filter_in__'}
{'shape': (1000001,), 'ori_shape': (1000001,), 'format': 'ND', 'sub_format': 0, 'ori_format': 'ND', 'dtype': 'float16', 'addr_type': 0, 'total_shape': [1000001], 'slice_offset': (), 'L1_addr_offset': 0, 'L1_fusion_type': -1, 'L1_workspace_size': -1, 'valid_shape': (), 'split_index': 0, 'atomic_type': '', 'input_c_values': -1, 'range': [[1000001, 1000001]], 'param_name': 'bias_in__'}
None
Outputs: {'shape': (1, 1000001, 1, 1), 'ori_shape': (1, 1000001, 1, 1), 'format': 'NCHW', 'sub_format': 0, 'ori_format': 'NCHW', 'dtype': 'float16', 'addr_type': 0, 'total_shape': [1, 1000001, 1, 1], 'slice_offset': (), 'L1_addr_offset': 0, 'L1_fusion_type': -1, 'L1_workspace_size': -1, 'valid_shape': (), 'split_index': 0, 'atomic_type': '', 'input_c_values': 1000001, 'range': [[1, 1], [1000001, 1000001], [1, 1], [1, 1]], 'param_name': 'y_out_'}
[OP_TILING] Attrs: [{'name': 'strides', 'dtype': 'list_int', 'value': (1, 1, 2, 2)}, {'name': 'pads', 'dtype': 'list_int', 'value': (0, 0, 0, 0)}, {'name': 'dilations', 'dtype': 'list_int', 'value': (1, 1, 1, 1)}, {'name': 'groups', 'dtype': 'int', 'value': 1}, {'name': 'data_format', 'dtype': 'str', 'value': 'NCHW'}, {'name': 'offset_x', 'dtype': 'int', 'value': 0}, {'name': 'pad_mode', 'dtype': 'str', 'value': 'SPECIFIC'}, {'name': 'enable_hf32', 'dtype': 'bool', 'value': False}, {'name': 'ascendc_op_para_size', 'dtype': 'int', 'value': 2097152}, {'name': 'fixed_shift_value', 'dtype': 'int64', 'value': 42}].

Special notes for this issue/备注 (Optional / 选填)