[Bug] 使用瑞芯微rk3588部署mmpose模型报错

[vicnoah]

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

1.执行pose推理报错。
./pose_detection cpu ../mmdeploy_models/mmpose/box ./rail/demo.jpg

报错信息:

[2024-05-11 14:10:18.592] [mmdeploy] [info] [model.cpp:35] [DirectoryModel] Load model: "../mmdeploy_models/mmpose/box"
[2024-05-11 14:10:18.652] [mmdeploy] [info] [rknn_net.cpp:105] input tensors: 
[2024-05-11 14:10:18.652] [mmdeploy] [info] [rknn_net.cpp:107]  - index=0, name=input, type=INT8, n_dims=4, dims=[1, 256, 192, 3], n_elems=147456, size=147456, fmt=NHWC, qnt_type=AFFINE, zp=-15, scale=0.018591292
[2024-05-11 14:10:18.652] [mmdeploy] [info] [rknn_net.cpp:105] output tensors: 
[2024-05-11 14:10:18.652] [mmdeploy] [info] [rknn_net.cpp:107]  - index=0, name=simcc_x, type=INT8, n_dims=4, dims=[1, 8, 384, 1], n_elems=3072, size=3072, fmt=NCHW, qnt_type=AFFINE, zp=-24, scale=0.0025815577
[2024-05-11 14:10:18.652] [mmdeploy] [info] [rknn_net.cpp:107]  - index=1, name=simcc_y, type=INT8, n_dims=4, dims=[1, 8, 512, 1], n_elems=4096, size=4096, fmt=NCHW, qnt_type=AFFINE, zp=0, scale=0.0036405057
[2024-05-11 14:10:18.670] [mmdeploy] [error] [simcc_label.cpp:50] unsupported `simcc_x` tensor, shape: [1, 8, 384, 1], dtype: 0
terminate called after throwing an instance of 'system_error2::status_error<mmdeploy::StatusDomain>'
  what():  not supported (2) @ /root/workspace/mmdeploy/csrc/mmdeploy/codebase/mmpose/simcc_label.cpp:52
./run_pose.sh：行 5: 2150391 已放弃               ./pose_detection cpu ../mmdeploy_models/mmpose/box ./rail/demo.jpg

报错部分的代码信息，我不明白发生错误的原因。

Reproduction

运行的命令
./pose_detection cpu ../mmdeploy_models/mmpose/box ./rail/demo.jpg

模型转换命令:

python ../mmdeploy/tools/deploy.py \
    ../mmdeploy/configs/mmpose/pose-detection_simcc_rknn-int8_static-256x192.py \
    configs/body_2d_keypoint/rtmpose/coco/rtmpose-m_8xb256-420e_coco-256x192_rail.py \
    ../models/mmpose/box/best_rail_publish.pth \
    ../models/demo.jpg \
    --work-dir ../mmdeploy/mmdeploy_models/mmpose/box \
    --device cpu \
    --dump-info

模型目录数据：

ls -al ../mmdeploy_models/mmpose/box/
总用量 65824
drwxrwxr-x 2 linaro linaro     4096  5月  6 16:54 .
drwxrwxr-x 4 linaro linaro     4096  5月  6 16:56 ..
-rw-rw-r-- 1 linaro linaro      350  5月  6 17:13 deploy.json
-rw-rw-r-- 1 linaro linaro     1320  5月  6 17:13 detail.json
-rw-rw-r-- 1 linaro linaro 52967951  5月  6 17:14 end2end.onnx
-rw-rw-r-- 1 linaro linaro 14412237  5月  6 17:14 end2end.rknn
-rw-rw-r-- 1 linaro linaro     3609  5月  6 17:13 pipeline.json

修改的配置mmdeploy/configs/base/backends/rknn.py

backend_config = dict(
    type='rknn',
    common_config=dict(
        target_platform='rk3588',  # 'rk3588'
        optimization_level=1),
    quantization_config=dict(
        do_quantization=True,
        dataset=None,
        pre_compile=False,
        rknn_batch_size=-1))

Environment

05/11 14:17:49 - mmengine - INFO - 

05/11 14:17:49 - mmengine - INFO - **********Environmental information**********
05/11 14:17:49 - mmengine - INFO - sys.platform: linux
05/11 14:17:49 - mmengine - INFO - Python: 3.8.19 (default, Mar 20 2024, 19:58:24) [GCC 11.2.0]
05/11 14:17:49 - mmengine - INFO - CUDA available: False
05/11 14:17:49 - mmengine - INFO - MUSA available: False
05/11 14:17:49 - mmengine - INFO - numpy_random_seed: 2147483648
05/11 14:17:49 - mmengine - INFO - GCC: gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
05/11 14:17:49 - mmengine - INFO - PyTorch: 1.8.1
05/11 14:17:49 - mmengine - INFO - PyTorch compiling details: PyTorch built with:
  - GCC 7.3
  - C++ Version: 201402
  - Intel(R) oneAPI Math Kernel Library Version 2023.1-Product Build 20230303 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, 

05/11 14:17:49 - mmengine - INFO - TorchVision: 0.9.1
05/11 14:17:49 - mmengine - INFO - OpenCV: 4.9.0
05/11 14:17:49 - mmengine - INFO - MMEngine: 0.10.4
05/11 14:17:49 - mmengine - INFO - MMCV: 2.0.1
05/11 14:17:49 - mmengine - INFO - MMCV Compiler: GCC 7.3
05/11 14:17:49 - mmengine - INFO - MMCV CUDA Compiler: not available
05/11 14:17:49 - mmengine - INFO - MMDeploy: 1.3.1+bc75c9d
05/11 14:17:49 - mmengine - INFO - 

05/11 14:17:49 - mmengine - INFO - **********Backend information**********
05/11 14:17:49 - mmengine - INFO - tensorrt:    None
05/11 14:17:49 - mmengine - INFO - ONNXRuntime: 1.8.1
05/11 14:17:49 - mmengine - INFO - ONNXRuntime-gpu:     None
05/11 14:17:49 - mmengine - INFO - ONNXRuntime custom ops:      Available
05/11 14:17:49 - mmengine - INFO - pplnn:       None
05/11 14:17:49 - mmengine - INFO - ncnn:        None
05/11 14:17:49 - mmengine - INFO - snpe:        None
05/11 14:17:49 - mmengine - INFO - openvino:    None
05/11 14:17:49 - mmengine - INFO - torchscript: 1.8.1
05/11 14:17:49 - mmengine - INFO - torchscript custom ops:      NotAvailable
05/11 14:17:50 - mmengine - INFO - rknn-toolkit:        None
05/11 14:17:50 - mmengine - INFO - rknn-toolkit2:       1.5.0+1fa95b5c
05/11 14:17:50 - mmengine - INFO - ascend:      None
05/11 14:17:50 - mmengine - INFO - coreml:      None
05/11 14:17:50 - mmengine - INFO - tvm: None
05/11 14:17:50 - mmengine - INFO - vacc:        None
05/11 14:17:50 - mmengine - INFO - 

05/11 14:17:50 - mmengine - INFO - **********Codebase information**********
05/11 14:17:50 - mmengine - INFO - mmdet:       3.3.0
05/11 14:17:50 - mmengine - INFO - mmseg:       None
05/11 14:17:50 - mmengine - INFO - mmpretrain:  1.1.1
05/11 14:17:50 - mmengine - INFO - mmocr:       None
05/11 14:17:50 - mmengine - INFO - mmagic:      None
05/11 14:17:50 - mmengine - INFO - mmdet3d:     None
05/11 14:17:50 - mmengine - INFO - mmpose:      1.1.0
05/11 14:17:50 - mmengine - INFO - mmrotate:    None
05/11 14:17:50 - mmengine - INFO - mmaction:    None
05/11 14:17:50 - mmengine - INFO - mmrazor:     None
05/11 14:17:50 - mmengine - INFO - mmyolo:      0.6.0

Error traceback

No response