[Bug] 在Ascend架构下使用deploy.py转换Ascend的om模型报错

[mutieying]

Checklist

• I have searched related issues but cannot get the expected help.
• 2. I have read the FAQ documentation but cannot get the expected help.
• 3. The bug has not been fixed in the latest version.

Describe the bug

显卡是华为的Atlas 300I Pro，在上面安装了mmdetection和mmdeploy0.12.0。下载了openmmlab的faster-rcnn模型http://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth，使用命令python tools/deploy.py configs/mmdet/detection/detection_ascend_static-800x1344.py ./mmdet_demo/faster_rcnn_r50_fpn_1x_coco.py ./mmdet_demo/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth demo/resources/det.jpg --work-dir mmdeploy_models/mmdet/faster_rcnn/cann --device cpu --dump-info转换模型时报下面的错

我找到报错的文件tensor_getitem.py，加上下面的两句。就可以正常转换成onnx和Ascend的om模型，推理不报错，但是没结果。

请问这是什么原因呢？
另外，我转换mmcls的模型是没问题的，可以正确推理。
整个mmdeploy部署的安装部署过程我是参考这个链接做的。https://bbs.huaweicloud.com/blogs/385709

Reproduction

python tools/deploy.py configs/mmdet/detection/detection_ascend_static-800x1344.py ./mmdet_demo/faster_rcnn_r50_fpn_1x_coco.py ./mmdet_demo/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth demo/resources/det.jpg --work-dir mmdeploy_models/mmdet/faster_rcnn/cann --device cpu --dump-info

Environment

2023-02-15 15:00:19,820 - mmdeploy - INFO - 

2023-02-15 15:00:19,820 - mmdeploy - INFO - **********Environmental information**********
2023-02-15 15:00:20,017 - mmdeploy - INFO - sys.platform: linux
2023-02-15 15:00:20,018 - mmdeploy - INFO - Python: 3.8.0 (default, Dec  9 2021, 17:53:27) [GCC 8.4.0]
2023-02-15 15:00:20,018 - mmdeploy - INFO - CUDA available: False
2023-02-15 15:00:20,018 - mmdeploy - INFO - GCC: x86_64-linux-gnu-gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
2023-02-15 15:00:20,018 - mmdeploy - INFO - PyTorch: 1.8.1+cpu
2023-02-15 15:00:20,018 - mmdeploy - INFO - PyTorch compiling details: PyTorch built with:
  - GCC 7.3
  - C++ Version: 201402
  - Intel(R) Math Kernel Library Version 2020.0.0 Product Build 20191122 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v1.7.0 (Git Hash 7aed236906b1f7a05c0917e5257a1af05e9ff683)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - NNPACK is enabled
  - CPU capability usage: AVX2
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.8.1, USE_CUDA=0, USE_CUDNN=OFF, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=OFF, USE_NNPACK=ON, USE_OPENMP=ON, 

2023-02-15 15:00:20,018 - mmdeploy - INFO - TorchVision: 0.9.1+cu102
2023-02-15 15:00:20,018 - mmdeploy - INFO - OpenCV: 4.7.0
2023-02-15 15:00:20,018 - mmdeploy - INFO - MMCV: 1.7.1
2023-02-15 15:00:20,018 - mmdeploy - INFO - MMCV Compiler: GCC 7.5
2023-02-15 15:00:20,018 - mmdeploy - INFO - MMCV CUDA Compiler: not available
2023-02-15 15:00:20,018 - mmdeploy - INFO - MMDeploy: 0.12.0+31b099a
2023-02-15 15:00:20,018 - mmdeploy - INFO - 

2023-02-15 15:00:20,018 - mmdeploy - INFO - **********Backend information**********
2023-02-15 15:00:20,028 - mmdeploy - INFO - tensorrt:	None
2023-02-15 15:00:20,071 - mmdeploy - INFO - ONNXRuntime:	1.13.1
2023-02-15 15:00:20,071 - mmdeploy - INFO - ONNXRuntime-gpu:	None
2023-02-15 15:00:20,071 - mmdeploy - INFO - ONNXRuntime custom ops:	NotAvailable
2023-02-15 15:00:20,072 - mmdeploy - INFO - pplnn:	None
2023-02-15 15:00:20,074 - mmdeploy - INFO - ncnn:	None
2023-02-15 15:00:20,075 - mmdeploy - INFO - snpe:	None
2023-02-15 15:00:20,076 - mmdeploy - INFO - openvino:	None
2023-02-15 15:00:20,078 - mmdeploy - INFO - torchscript:	1.8.1+cpu
2023-02-15 15:00:20,078 - mmdeploy - INFO - torchscript custom ops:	NotAvailable
2023-02-15 15:00:20,125 - mmdeploy - INFO - rknn-toolkit:	None
2023-02-15 15:00:20,125 - mmdeploy - INFO - rknn2-toolkit:	None
2023-02-15 15:00:20,160 - mmdeploy - INFO - ascend:	None
2023-02-15 15:00:20,161 - mmdeploy - INFO - coreml:	None
2023-02-15 15:00:20,162 - mmdeploy - INFO - tvm:	None
2023-02-15 15:00:20,162 - mmdeploy - INFO - 

2023-02-15 15:00:20,162 - mmdeploy - INFO - **********Codebase information**********
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmdet:	2.28.1
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmseg:	None
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmcls:	0.25.0
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmocr:	None
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmedit:	None
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmdet3d:	None
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmpose:	None
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmrotate:	None
2023-02-15 15:00:20,165 - mmdeploy - INFO - mmaction:	None

Error traceback

No response

[mm-assistant]

We recommend using English or English & Chinese for issues so that we could have broader discussion.

[grimoire]

#1842 Please have a try.