[DEBUG] OP(292804,python3):2026-03-04-10:53:12.407.792 [contiguous.cpp:659][NNOP][ViewCopy][292804] OpName:[aclnnAdaptiveMaxPool3d_0_ViewCopy] ViewCopy data addr: 0x120000018000
[ERROR] OP(292804,python3):2026-03-04-10:53:12.407.801 [contiguous.cpp:630][NNOP][CheckViewCopyParams][292804] errno[161002] OpName:[aclnnAdaptiveMaxPool3d_0_ViewCopy] Input tensor's dtype[DT_INT32] should be same with output's dtype[DT_INT64].
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)
当输入索引数据类型为int64时，通过aclnnAdaptiveMaxPool3d调用算子，出现类型不匹配报错
Environment / 环境信息 (Mandatory / 必填)
950
Steps to reproduce the issue / 重现步骤 (Mandatory / 必填)
在950环境下执行下述用例：
输入 (x)：
shape [2, 15, 134, 102, 142]
dtype float32
属性 (output_size):
[1, 1, 1]
输出：
y :
shape [2, 15, 1, 1 1]
dtype float32
indices:
shape [2. 15, 1, 1, 1]
dtype int64
Describe the expected behavior / 预期结果 (Mandatory / 必填)
当索引值类型为int64时aclnnAdaptiveMaxPool3d能正常执行
Related log / screenshot / 日志 / 截图 (Mandatory / 必填)
[DEBUG] OP(292804,python3):2026-03-04-10:53:12.407.792 [contiguous.cpp:659][NNOP][ViewCopy][292804] OpName:[aclnnAdaptiveMaxPool3d_0_ViewCopy] ViewCopy data addr: 0x120000018000
[ERROR] OP(292804,python3):2026-03-04-10:53:12.407.801 [contiguous.cpp:630][NNOP][CheckViewCopyParams][292804] errno[161002] OpName:[aclnnAdaptiveMaxPool3d_0_ViewCopy] Input tensor's dtype[DT_INT32] should be same with output's dtype[DT_INT64].
Special notes for this issue/备注 (Optional / 选填)