search

open-mmlab / mmdeploy

OpenMMLab Model Deployment Framework
https://mmdeploy.readthedocs.io/en/latest/
Apache License 2.0
2.73k stars 627 forks source link

[ONNXRuntimeError] : 1 : FAIL : Load model from ./../end2end.onnx failed:Fatal error: TRTBatchedBEVNMS is not a registered function/op #1782

Open kunnalparihar opened 1 year ago

kunnalparihar commented 1 year ago

Checklist

Describe the bug

Hi, I am using mmdeploy to convert point pillar model to onnx and tensorRT. I am able to generate .onnx and .engine file using tools/deploy.py but when I test using tools/test.py, it shows above error.

I download the model file from here https://github.com/open-mmlab/mmdetection3d/tree/master/configs/pointpillars

Reproduction

I used docker file to create the required environment. Dockerfile.txt Inside docker container, I installed mmdetection3d repo. using this below command, pip install openmim mim install mmdet mim install mmsegmentation git clone https://github.com/open-mmlab/mmdetection3d.git cd mmdetection3d pip install -e .

pip install numpy==1.23.5 export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/root/workspace/onnxruntime-linux-x64-1.8.1/lib/

Commands I used to deploy and test the model are python tools/deploy.py configs/mmdet3d/voxel-detection/voxel-detection_tensorrt_dynamic-nus-64x4.py ../mmdetection3d/configs/pointpillars/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d.py /usr/src/app/dataset/mmdeploy_dataset/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d_20210826_225857-f19d00a3.pth /usr/src/app/ws/n015-2018-07-24-11-22-45+0800__LIDAR_TOP__1532402927647951.pcd.bin --work-dir ./../ --device cuda:0

python tools/test.py configs/mmdet3d/voxel-detection/voxel-detection_onnxruntime_dynamic.py ./../mmdetection3d/configs/pointpillars/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d.py --model ./../end2end.onnx --metrics bbox --device cuda:0

Environment

python tools/check_env.py
2023-02-20 01:13:20,748 - mmdeploy - INFO - 

2023-02-20 01:13:20,748 - mmdeploy - INFO - **********Environmental information**********
2023-02-20 01:13:20,919 - mmdeploy - INFO - sys.platform: linux
2023-02-20 01:13:20,919 - mmdeploy - INFO - Python: 3.8.16 (default, Jan 17 2023, 23:13:24) [GCC 11.2.0]
2023-02-20 01:13:20,920 - mmdeploy - INFO - CUDA available: True
2023-02-20 01:13:20,920 - mmdeploy - INFO - GPU 0,1,2,3,4,5,6,7: NVIDIA GeForce GTX 1080 Ti
2023-02-20 01:13:20,920 - mmdeploy - INFO - CUDA_HOME: /usr/local/cuda
2023-02-20 01:13:20,920 - mmdeploy - INFO - NVCC: Cuda compilation tools, release 11.7, V11.7.99
2023-02-20 01:13:20,920 - mmdeploy - INFO - GCC: gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
2023-02-20 01:13:20,920 - mmdeploy - INFO - PyTorch: 1.8.0
2023-02-20 01:13:20,920 - mmdeploy - INFO - PyTorch compiling details: PyTorch built with:
  - GCC 7.3
  - C++ Version: 201402
  - Intel(R) Math Kernel Library Version 2020.0.2 Product Build 20200624 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
  - CUDA Runtime 10.2
  - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_37,code=compute_37
  - CuDNN 7.6.5
  - Magma 2.5.2
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=10.2, CUDNN_VERSION=7.6.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -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=1.8.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, 

2023-02-20 01:13:20,920 - mmdeploy - INFO - TorchVision: 0.9.0
2023-02-20 01:13:20,920 - mmdeploy - INFO - OpenCV: 4.7.0
2023-02-20 01:13:20,920 - mmdeploy - INFO - MMCV: 1.6.2
2023-02-20 01:13:20,920 - mmdeploy - INFO - MMCV Compiler: GCC 7.3
2023-02-20 01:13:20,920 - mmdeploy - INFO - MMCV CUDA Compiler: 10.2
2023-02-20 01:13:20,920 - mmdeploy - INFO - MMDeploy: 0.13.0+b1be9c6
2023-02-20 01:13:20,920 - mmdeploy - INFO - 

2023-02-20 01:13:20,920 - mmdeploy - INFO - **********Backend information**********
2023-02-20 01:13:20,971 - mmdeploy - INFO - tensorrt:   8.4.2.4
2023-02-20 01:13:20,971 - mmdeploy - INFO - tensorrt custom ops:    Available
2023-02-20 01:13:21,019 - mmdeploy - INFO - ONNXRuntime:    None
2023-02-20 01:13:21,019 - mmdeploy - INFO - ONNXRuntime-gpu:    1.8.1
2023-02-20 01:13:21,020 - mmdeploy - INFO - ONNXRuntime custom ops: Available
2023-02-20 01:13:21,021 - mmdeploy - INFO - pplnn:  None
2023-02-20 01:13:21,025 - mmdeploy - INFO - ncnn:   None
2023-02-20 01:13:21,028 - mmdeploy - INFO - snpe:   None
2023-02-20 01:13:21,029 - mmdeploy - INFO - openvino:   None
2023-02-20 01:13:21,030 - mmdeploy - INFO - torchscript:    1.8.0
2023-02-20 01:13:21,031 - mmdeploy - INFO - torchscript custom ops: NotAvailable
2023-02-20 01:13:21,088 - mmdeploy - INFO - rknn-toolkit:   None
2023-02-20 01:13:21,088 - mmdeploy - INFO - rknn2-toolkit:  None
2023-02-20 01:13:21,090 - mmdeploy - INFO - ascend: None
2023-02-20 01:13:21,092 - mmdeploy - INFO - coreml: None
2023-02-20 01:13:21,094 - mmdeploy - INFO - tvm:    None
2023-02-20 01:13:21,094 - mmdeploy - INFO - 

2023-02-20 01:13:21,094 - mmdeploy - INFO - **********Codebase information**********
2023-02-20 01:13:21,096 - mmdeploy - INFO - mmdet:  2.28.1
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmseg:  0.30.0
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmcls:  0.25.0
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmocr:  None
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmedit: None
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmdet3d:    1.0.0rc6
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmpose: None
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmrotate:   None
2023-02-20 01:13:21,097 - mmdeploy - INFO - mmaction:   None

Error traceback

***Output of running deploy.py***
python tools/deploy.py configs/mmdet3d/voxel-detection/voxel-detection_tensorrt_dynamic-nus-64x4.py ../mmdetection3d/configs/pointpillars/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d.py /usr/src/app/dataset/mmdeploy_dataset/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d_20210826_225857-f19d00a3.pth /usr/src/app/ws/n015-2018-07-24-11-22-45+0800__LIDAR_TOP__1532402927647951.pcd.bin --work-dir ./../ --device cuda:0
2023-02-19 14:25:55,902 - mmdeploy - INFO - Start pipeline mmdeploy.apis.pytorch2onnx.torch2onnx in subprocess
/root/workspace/mmdetection3d/mmdet3d/models/dense_heads/anchor3d_head.py:84: UserWarning: dir_offset and dir_limit_offset will be depressed and be incorporated into box coder in the future
  warnings.warn(
load checkpoint from local path: /usr/src/app/dataset/mmdeploy_dataset/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d_20210826_225857-f19d00a3.pth
2023-02-19 14:26:01,446 - mmdeploy - WARNING - DeprecationWarning: get_onnx_config will be deprecated in the future. 
2023-02-19 14:26:01,447 - mmdeploy - INFO - Export PyTorch model to ONNX: ./../end2end.onnx.
2023-02-19 14:26:01,509 - mmdeploy - WARNING - Can not find torch._C._jit_pass_onnx_autograd_function_process, function rewrite will not be applied
2023-02-19 14:26:01,512 - mmdeploy - WARNING - Can not find torch.nn.functional._scaled_dot_product_attention, function rewrite will not be applied
2023-02-19 14:26:01,513 - mmdeploy - WARNING - Can not find torch._C._jit_pass_onnx_deduplicate_initializers, function rewrite will not be applied
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/core/anchor/anchor_3d_generator.py:44: TracerWarning: torch.tensor results are registered as constants in the trace. You can safely ignore this warning if you use this function to create tensors out of constant variables that would be the same every time you call this function. In any other case, this might cause the trace to be incorrect.
  anchor_range = torch.tensor(anchor_range, device=device)
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/core/anchor/anchor_3d_generator.py:65: TracerWarning: torch.tensor results are registered as constants in the trace. You can safely ignore this warning if you use this function to create tensors out of constant variables that would be the same every time you call this function. In any other case, this might cause the trace to be incorrect.
  sizes = torch.tensor(sizes, device=device).reshape(-1, 3) * scale
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/core/anchor/anchor_3d_generator.py:66: TracerWarning: torch.tensor results are registered as constants in the trace. You can safely ignore this warning if you use this function to create tensors out of constant variables that would be the same every time you call this function. In any other case, this might cause the trace to be incorrect.
  rotations = torch.tensor(rotations, device=device)
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/core/anchor/anchor_3d_generator.py:74: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  tile_shape[-2] = int(sizes.shape[0])
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/models/anchor3d_head.py:65: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  assert cls_score.size()[-2:] == bbox_pred.size()[-2:]
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/models/anchor3d_head.py:79: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if nms_pre > 0 and scores.shape[1] > nms_pre:
/root/workspace/mmdeploy/mmdeploy/pytorch/functions/topk.py:58: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if k > size:
/root/workspace/mmdetection3d/mmdet3d/core/bbox/coders/delta_xyzwhlr_bbox_coder.py:72: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if box_ndim > 7:
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/models/anchor3d_head.py:100: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if mlvl_bboxes.shape[0] > 0:
/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/core/post_processing/box3d_nms.py:106: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  num_classes = int(mlvl_scores.shape[-1])
/root/workspace/mmdeploy/mmdeploy/mmcv/ops/nms_rotated.py:197: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  out_boxes = min(num_boxes, after_topk)
2023-02-19 14:26:06,222 - mmdeploy - INFO - Execute onnx optimize passes.
2023-02-19 14:26:06,544 - mmdeploy - INFO - Finish pipeline mmdeploy.apis.pytorch2onnx.torch2onnx
2023-02-19 14:26:09,155 - mmdeploy - INFO - Start pipeline mmdeploy.apis.utils.utils.to_backend in subprocess
2023-02-19 14:26:09,197 - mmdeploy - INFO - Successfully loaded tensorrt plugins from /root/workspace/mmdeploy/mmdeploy/lib/libmmdeploy_tensorrt_ops.so
[02/19/2023-14:26:09] [TRT] [I] [MemUsageChange] Init CUDA: CPU +195, GPU +0, now: CPU 257, GPU 3642 (MiB)
[02/19/2023-14:26:10] [TRT] [I] [MemUsageChange] Init builder kernel library: CPU +6, GPU +2, now: CPU 282, GPU 3644 (MiB)
[02/19/2023-14:26:10] [TRT] [I] ----------------------------------------------------------------
[02/19/2023-14:26:10] [TRT] [I] Input filename:   ./../end2end.onnx
[02/19/2023-14:26:10] [TRT] [I] ONNX IR version:  0.0.6
[02/19/2023-14:26:10] [TRT] [I] Opset version:    11
[02/19/2023-14:26:10] [TRT] [I] Producer name:    pytorch
[02/19/2023-14:26:10] [TRT] [I] Producer version: 1.8
[02/19/2023-14:26:10] [TRT] [I] Domain:           
[02/19/2023-14:26:10] [TRT] [I] Model version:    0
[02/19/2023-14:26:10] [TRT] [I] Doc string:       
[02/19/2023-14:26:10] [TRT] [I] ----------------------------------------------------------------
[02/19/2023-14:26:10] [TRT] [W] parsers/onnx/onnx2trt_utils.cpp:367: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[02/19/2023-14:26:10] [TRT] [W] parsers/onnx/onnx2trt_utils.cpp:395: One or more weights outside the range of INT32 was clamped
[02/19/2023-14:26:11] [TRT] [W] Tensor DataType is determined at build time for tensors not marked as input or output.
[02/19/2023-14:26:12] [TRT] [I] No importer registered for op: TRTBatchedBEVNMS. Attempting to import as plugin.
[02/19/2023-14:26:12] [TRT] [I] Searching for plugin: TRTBatchedBEVNMS, plugin_version: 1, plugin_namespace: 
[02/19/2023-14:26:12] [TRT] [I] Successfully created plugin: TRTBatchedBEVNMS
[02/19/2023-14:26:13] [TRT] [I] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +255, GPU +106, now: CPU 562, GPU 3750 (MiB)
[02/19/2023-14:26:13] [TRT] [I] [MemUsageChange] Init cuDNN: CPU +114, GPU +46, now: CPU 676, GPU 3796 (MiB)
[02/19/2023-14:26:13] [TRT] [I] Local timing cache in use. Profiling results in this builder pass will not be stored.
[02/19/2023-14:26:20] [TRT] [I] Some tactics do not have sufficient workspace memory to run. Increasing workspace size will enable more tactics, please check verbose output for requested sizes.
[02/19/2023-14:26:22] [TRT] [I] Detected 3 inputs and 3 output network tensors.
[02/19/2023-14:26:22] [TRT] [I] Total Host Persistent Memory: 22256
[02/19/2023-14:26:22] [TRT] [I] Total Device Persistent Memory: 1036288
[02/19/2023-14:26:22] [TRT] [I] Total Scratch Memory: 2277160320
[02/19/2023-14:26:22] [TRT] [I] [MemUsageStats] Peak memory usage of TRT CPU/GPU memory allocators: CPU 8 MiB, GPU 2243 MiB
[02/19/2023-14:26:22] [TRT] [I] [BlockAssignment] Algorithm ShiftNTopDown took 6.56618ms to assign 11 blocks to 54 nodes requiring 2367475200 bytes.
[02/19/2023-14:26:22] [TRT] [I] Total Activation Memory: 2367475200
[02/19/2023-14:26:22] [TRT] [I] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +8, now: CPU 933, GPU 3940 (MiB)
[02/19/2023-14:26:22] [TRT] [I] [MemUsageChange] Init cuDNN: CPU +0, GPU +8, now: CPU 933, GPU 3948 (MiB)
[02/19/2023-14:26:22] [TRT] [I] [MemUsageChange] TensorRT-managed allocation in building engine: CPU +2, GPU +31, now: CPU 2, GPU 31 (MiB)
[02/19/2023-14:26:22] [TRT] [W] The getMaxBatchSize() function should not be used with an engine built from a network created with NetworkDefinitionCreationFlag::kEXPLICIT_BATCH flag. This function will always return 1.
[02/19/2023-14:26:22] [TRT] [W] The getMaxBatchSize() function should not be used with an engine built from a network created with NetworkDefinitionCreationFlag::kEXPLICIT_BATCH flag. This function will always return 1.
2023-02-19 14:26:22,947 - mmdeploy - INFO - Finish pipeline mmdeploy.apis.utils.utils.to_backend
2023-02-19 14:26:23,804 - mmdeploy - INFO - visualize tensorrt model start.
2023-02-19:14:26:26 - trimesh - WARNING - No FCL -- collision checking will not work
2023-02-19 14:26:28,153 - mmdeploy - INFO - Successfully loaded tensorrt plugins from /root/workspace/mmdeploy/mmdeploy/lib/libmmdeploy_tensorrt_ops.so
2023-02-19:14:26:28 - mmdeploy - INFO - Successfully loaded tensorrt plugins from /root/workspace/mmdeploy/mmdeploy/lib/libmmdeploy_tensorrt_ops.so
2023-02-19 14:26:28,154 - mmdeploy - INFO - Successfully loaded tensorrt plugins from /root/workspace/mmdeploy/mmdeploy/lib/libmmdeploy_tensorrt_ops.so
2023-02-19:14:26:28 - mmdeploy - INFO - Successfully loaded tensorrt plugins from /root/workspace/mmdeploy/mmdeploy/lib/libmmdeploy_tensorrt_ops.so
2023-02-19 14:26:34,078 - mmdeploy - WARNING - render and display result skipped for headless device, exception no display name and no $DISPLAY environment variable
2023-02-19:14:26:34 - mmdeploy - WARNING - render and display result skipped for headless device, exception no display name and no $DISPLAY environment variable
2023-02-19 14:26:35,494 - mmdeploy - INFO - visualize tensorrt model success.
2023-02-19 14:26:35,495 - mmdeploy - INFO - visualize pytorch model start.
2023-02-19:14:26:38 - trimesh - WARNING - No FCL -- collision checking will not work
/root/workspace/mmdetection3d/mmdet3d/models/dense_heads/anchor3d_head.py:84: UserWarning: dir_offset and dir_limit_offset will be depressed and be incorporated into box coder in the future
  warnings.warn(
load checkpoint from local path: /usr/src/app/dataset/mmdeploy_dataset/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d_20210826_225857-f19d00a3.pth
2023-02-19 14:26:42,149 - mmdeploy - WARNING - render and display result skipped for headless device, exception no display name and no $DISPLAY environment variable
2023-02-19:14:26:42 - mmdeploy - WARNING - render and display result skipped for headless device, exception no display name and no $DISPLAY environment variable
2023-02-19 14:26:42,931 - mmdeploy - INFO - visualize pytorch model success.
2023-02-19 14:26:42,932 - mmdeploy - INFO - All process success.

***Output of running test.py***
python tools/test.py configs/mmdet3d/voxel-detection/voxel-detection_onnxruntime_dynamic.py ./../mmdetection3d/configs/pointpillars/hv_pointpillars_secfpn_sbn-all_4x8_2x_nus-3d.py --model ./../end2end.onnx --metrics bbox --device cuda:0
2023-02-19 14:43:01,001 - mmdeploy - INFO - Successfully loaded onnxruntime custom ops from /root/workspace/mmdeploy/mmdeploy/lib/libmmdeploy_onnxruntime_ops.so
Traceback (most recent call last):
  File "tools/test.py", line 168, in <module>
    main()
  File "tools/test.py", line 123, in main
    model = task_processor.init_backend_model(args.model, uri=args.uri)
  File "/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/deploy/voxel_detection.py", line 38, in init_backend_model
    model = build_voxel_detection_model(
  File "/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/deploy/voxel_detection_model.py", line 199, in build_voxel_detection_model
    backend_detector = __BACKEND_MODEL.build(
  File "/opt/conda/lib/python3.8/site-packages/mmcv/utils/registry.py", line 237, in build
    return self.build_func(*args, **kwargs, registry=self)
  File "/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/deploy/voxel_detection_model.py", line 16, in __build_backend_voxel_model
    return registry.module_dict[cls_name](*args, **kwargs)
  File "/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/deploy/voxel_detection_model.py", line 47, in __init__
    self._init_wrapper(
  File "/root/workspace/mmdeploy/mmdeploy/codebase/mmdet3d/deploy/voxel_detection_model.py", line 61, in _init_wrapper
    self.wrapper = BaseBackendModel._build_wrapper(
  File "/root/workspace/mmdeploy/mmdeploy/codebase/base/backend_model.py", line 63, in _build_wrapper
    return backend_mgr.build_wrapper(backend_files, device, input_names,
  File "/root/workspace/mmdeploy/mmdeploy/backend/onnxruntime/backend_manager.py", line 34, in build_wrapper
    return ORTWrapper(
  File "/root/workspace/mmdeploy/mmdeploy/backend/onnxruntime/wrapper.py", line 56, in __init__
    sess = ort.InferenceSession(
  File "/opt/conda/lib/python3.8/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 283, in __init__
    self._create_inference_session(providers, provider_options, disabled_optimizers)
  File "/opt/conda/lib/python3.8/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 310, in _create_inference_session
    sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from ./../end2end.onnx failed:Fatal error: TRTBatchedBEVNMS is not a registered function/op
RhythmWings commented 4 months ago

Have you solved this problem? If you have solved, could you please tell me the way? Thanks!