A 3D computer vision development toolkit based on PaddlePaddle. It supports point-cloud object detection, segmentation, and monocular 3D object detection models.
Apache License 2.0
573
stars
141
forks
source link
ImportError: cannot import name 'ball_query' from 'pointnet2_ops', when running IASSD inference #475
------------------------------------------------
W0903 11:19:05.792060 61632 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 8.6, Driver API Version: 12.2, Runtime API Version: 11.7
W0903 11:19:05.792079 61632 gpu_resources.cc:91] device: 0, cuDNN Version: 8.9.
2024-09-03 11:19:08,971 - WARNING - No custom op pointnet2_ops found, try JIT build
Compiling user custom op, it will cost a few seconds.....
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
W0903 11:19:25.526822 61632 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:19:25.526851 61632 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:19:25.526854 61632 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:19:25.526856 61632 custom_operator.cc:723] Operator (hard_voxelize) has been registered.
W0903 11:19:25.526859 61632 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:19:25.526861 61632 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
2024-09-03 11:19:25,534 - INFO - pointnet2_ops builded success!
/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/paddle/nn/layer/norm.py:712: UserWarning: When training, we now always track global mean and variance.
"When training, we now always track global mean and variance."
2024-09-03 11:19:26,769 - WARNING - No custom op roiaware_pool3d found, try JIT build
Compiling user custom op, it will cost a few seconds.....
W0903 11:19:27.559796 61632 custom_operator.cc:723] Operator (hard_voxelize) has been registered.
W0903 11:19:27.559818 61632 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:19:27.559819 61632 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:19:27.559821 61632 custom_operator.cc:723] Operator (ball_query_stack) has been registered.
W0903 11:19:27.559823 61632 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:19:27.559825 61632 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:19:27.559827 61632 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:19:27.559829 61632 custom_operator.cc:723] Operator (gather_operation) has been registered.
W0903 11:19:27.559831 61632 custom_operator.cc:723] Operator (ball_query_batch) has been registered.
W0903 11:19:27.559834 61632 custom_operator.cc:723] Operator (grouping_operation_batch) has been registered.
W0903 11:19:27.559835 61632 custom_operator.cc:723] Operator (voxel_query_wrapper) has been registered.
W0903 11:19:27.559837 61632 custom_operator.cc:723] Operator (grouping_operation_stack) has been registered.
W0903 11:19:27.559839 61632 custom_operator.cc:723] Operator (farthest_point_sample) has been registered.
2024-09-03 11:19:27,565 - INFO - roiaware_pool3d builded success!
I was then able to export the model with python tools/export.py --config configs/iassd/iassd_kitti.yaml --model outputs/iassd_kitti/epoch_1/model.pdparams --save_dir outputs/iassd_kitti/export, however, there is a UserWarning:
/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/paddle/tensor/creation.py:1806: UserWarning: paddle.assign doesn't support float64 input now due to current platform protobuf data limitation, we convert it to float32
"paddle.assign doesn't support float64 input now due "
/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/paddle/fluid/layers/tensor.py:668: UserWarning: paddle.assign doesn't support float64 input now due to current platform protobuf data limitation, we convert it to float32
"paddle.assign doesn't support float64 input now due "
/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/paddle/fluid/layers/control_flow.py:125: UserWarning: the input shapes of select_input should have the same rank, but get (-1, 7), (1,)
f"the input shapes of select_input should have the same rank, but get {first_shape}, {second_shape}"
2024-09-03 11:21:13,362 - INFO - Exported model is saved in outputs/iassd_kitti/export/iassd
Next, if I try to run inference, python3.7 deploy/iassd/python/infer.py --model_file outputs/iassd_kitti/export/iassd.pdmodel --params_file outputs/iassd_kitti/export/iassd.pdiparams --lidar_file datasets/KITTI/testing/velodyne/000000.bin --gpu_id 0 --run_mode trt_fp32 i get ImportError:
W0903 11:23:02.417436 64919 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:23:02.417456 64919 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:23:02.417460 64919 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:23:02.417511 64919 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:23:02.417515 64919 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:461: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
interp=Image.BILINEAR):
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1502: DeprecationWarning: NEAREST is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.NEAREST or Dither.NONE instead.
'nearest': Image.NEAREST,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1503: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
'bilinear': Image.BILINEAR,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1504: DeprecationWarning: BICUBIC is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BICUBIC instead.
'bicubic': Image.BICUBIC,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1505: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
'box': Image.BOX,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1506: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
'lanczos': Image.LANCZOS,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1507: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
'hamming': Image.HAMMING
ortools not installed, install it by "pip install ortools==9.1.9490" if you run BEVLaneDet model
2024-09-03 11:23:02,580 - WARNING - No custom op pointnet2_ops found, try JIT build
Compiling user custom op, it will cost a few seconds.....
W0903 11:23:03.350497 64919 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:23:03.350534 64919 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:23:03.350538 64919 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:23:03.350539 64919 custom_operator.cc:723] Operator (hard_voxelize) has been registered.
W0903 11:23:03.350541 64919 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:23:03.350543 64919 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
2024-09-03 11:23:03,357 - INFO - pointnet2_ops builded success!
Traceback (most recent call last):
File "deploy/iassd/python/infer.py", line 23, in <module>
from paddle3d.ops.pointnet2_ops import (ball_query, farthest_point_sample,
File "/home/gabriel/Paddle3D/paddle3d/ops/__init__.py", line 189, in __getattr__
attr, self.modulename, module.__file__))
ImportError: cannot import name 'ball_query' from 'pointnet2_ops' (/home/gabriel/.cache/paddle_extensions/pointnet2_ops/pointnet2_ops_140464807302336.py)
I thought maybe the operation was not built properly, so i then try to build it manually by cd into paddle3d/ops and running python setup.py install. However, it seemed to fail at building centerpoint operations.
/usr/local/cuda-11.6/bin/nvcc -I/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/paddle/include -I/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/paddle/include/third_party -I/usr/local/cuda-11.6/include -I/home/gabriel/miniconda3/envs/paddle_37/include/python3.7m -c /home/gabriel/Paddle3D/paddle3d/ops/centerpoint_postprocess/postprocess.cu -o /home/gabriel/Paddle3D/paddle3d/ops/build/centerpoint_postprocess/lib.linux-x86_64-cpython-37/postprocess.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
/home/gabriel/Paddle3D/paddle3d/ops/centerpoint_postprocess/postprocess.cu(147): error: too few arguments in function call
1 error detected in the compilation of "/home/gabriel/Paddle3D/paddle3d/ops/centerpoint_postprocess/postprocess.cu".
error: command '/usr/local/cuda-11.6/bin/nvcc' failed with exit code 1
Since I dont care about centerpoint, I commented out the operations in paddle3d/ops/__init__.py:
After I did this, I was able to run setup.py successfully. However, I still cannot run infer.py
(paddle_37) gabriel@gabriel-Z790-AORUS-ELITE-AX:~/Paddle3D$ python3.7 deploy/iassd/python/infer.py --model_file outputs/iassd_kitti/export/iassd.pdmodel --params_file outputs/iassd_kitti/export/iassd.pdiparams --lidar_file datasets/KITTI/testing/velodyne/000000.bin --gpu_id 0 --run_mode trt_fp32
W0903 11:32:00.621784 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.621804 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.621806 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.621807 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.621809 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.626067 71289 custom_operator.cc:723] Operator (assign_score_withk) has been registered.
W0903 11:32:00.626078 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.626088 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.626089 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.626091 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.626092 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.631099 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.631114 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.631116 71289 custom_operator.cc:723] Operator (ball_query_stack) has been registered.
W0903 11:32:00.631117 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.631119 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.631119 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.631120 71289 custom_operator.cc:723] Operator (assign_score_withk) has been registered.
W0903 11:32:00.631122 71289 custom_operator.cc:723] Operator (gather_operation) has been registered.
W0903 11:32:00.631124 71289 custom_operator.cc:723] Operator (ball_query_batch) has been registered.
W0903 11:32:00.631124 71289 custom_operator.cc:723] Operator (grouping_operation_batch) has been registered.
W0903 11:32:00.631125 71289 custom_operator.cc:723] Operator (voxel_query_wrapper) has been registered.
W0903 11:32:00.631126 71289 custom_operator.cc:723] Operator (grouping_operation_stack) has been registered.
W0903 11:32:00.631129 71289 custom_operator.cc:723] Operator (farthest_point_sample) has been registered.
W0903 11:32:00.636233 71289 custom_operator.cc:723] Operator (ms_deform_attn) has been registered.
W0903 11:32:00.636245 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.636246 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.636250 71289 custom_operator.cc:723] Operator (ball_query_stack) has been registered.
W0903 11:32:00.636250 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.636252 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.636255 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.636255 71289 custom_operator.cc:723] Operator (assign_score_withk) has been registered.
W0903 11:32:00.636257 71289 custom_operator.cc:723] Operator (gather_operation) has been registered.
W0903 11:32:00.636258 71289 custom_operator.cc:723] Operator (ball_query_batch) has been registered.
W0903 11:32:00.636260 71289 custom_operator.cc:723] Operator (grouping_operation_batch) has been registered.
W0903 11:32:00.636261 71289 custom_operator.cc:723] Operator (voxel_query_wrapper) has been registered.
W0903 11:32:00.636263 71289 custom_operator.cc:723] Operator (grouping_operation_stack) has been registered.
W0903 11:32:00.636265 71289 custom_operator.cc:723] Operator (farthest_point_sample) has been registered.
W0903 11:32:00.639941 71289 custom_operator.cc:723] Operator (bev_pool_v2) has been registered.
W0903 11:32:00.639951 71289 custom_operator.cc:723] Operator (ms_deform_attn) has been registered.
W0903 11:32:00.639976 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.639977 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.639978 71289 custom_operator.cc:723] Operator (ball_query_stack) has been registered.
W0903 11:32:00.639981 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.639982 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.639984 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.639986 71289 custom_operator.cc:723] Operator (assign_score_withk) has been registered.
W0903 11:32:00.639987 71289 custom_operator.cc:723] Operator (gather_operation) has been registered.
W0903 11:32:00.639989 71289 custom_operator.cc:723] Operator (ball_query_batch) has been registered.
W0903 11:32:00.639990 71289 custom_operator.cc:723] Operator (grouping_operation_batch) has been registered.
W0903 11:32:00.639992 71289 custom_operator.cc:723] Operator (voxel_query_wrapper) has been registered.
W0903 11:32:00.639993 71289 custom_operator.cc:723] Operator (grouping_operation_stack) has been registered.
W0903 11:32:00.639995 71289 custom_operator.cc:723] Operator (farthest_point_sample) has been registered.
W0903 11:32:00.645074 71289 custom_operator.cc:723] Operator (bev_pool_v2) has been registered.
W0903 11:32:00.645089 71289 custom_operator.cc:723] Operator (ms_deform_attn) has been registered.
W0903 11:32:00.645092 71289 custom_operator.cc:723] Operator (bev_pool_v2_bkwd) has been registered.
W0903 11:32:00.645093 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.645095 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.645097 71289 custom_operator.cc:723] Operator (ball_query_stack) has been registered.
W0903 11:32:00.645098 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.645100 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.645102 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.645103 71289 custom_operator.cc:723] Operator (assign_score_withk) has been registered.
W0903 11:32:00.645105 71289 custom_operator.cc:723] Operator (gather_operation) has been registered.
W0903 11:32:00.645107 71289 custom_operator.cc:723] Operator (ball_query_batch) has been registered.
W0903 11:32:00.645108 71289 custom_operator.cc:723] Operator (grouping_operation_batch) has been registered.
W0903 11:32:00.645109 71289 custom_operator.cc:723] Operator (voxel_query_wrapper) has been registered.
W0903 11:32:00.645112 71289 custom_operator.cc:723] Operator (grouping_operation_stack) has been registered.
W0903 11:32:00.645112 71289 custom_operator.cc:723] Operator (farthest_point_sample) has been registered.
W0903 11:32:00.652513 71289 custom_operator.cc:723] Operator (hard_voxelize) has been registered.
W0903 11:32:00.652525 71289 custom_operator.cc:723] Operator (bev_pool_v2) has been registered.
W0903 11:32:00.652527 71289 custom_operator.cc:723] Operator (ms_deform_attn) has been registered.
W0903 11:32:00.652529 71289 custom_operator.cc:723] Operator (bev_pool_v2_bkwd) has been registered.
W0903 11:32:00.652530 71289 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:32:00.652532 71289 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:32:00.652534 71289 custom_operator.cc:723] Operator (ball_query_stack) has been registered.
W0903 11:32:00.652535 71289 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
W0903 11:32:00.652536 71289 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:32:00.652539 71289 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:32:00.652540 71289 custom_operator.cc:723] Operator (assign_score_withk) has been registered.
W0903 11:32:00.652541 71289 custom_operator.cc:723] Operator (gather_operation) has been registered.
W0903 11:32:00.652544 71289 custom_operator.cc:723] Operator (ball_query_batch) has been registered.
W0903 11:32:00.652544 71289 custom_operator.cc:723] Operator (grouping_operation_batch) has been registered.
W0903 11:32:00.652546 71289 custom_operator.cc:723] Operator (voxel_query_wrapper) has been registered.
W0903 11:32:00.652547 71289 custom_operator.cc:723] Operator (grouping_operation_stack) has been registered.
W0903 11:32:00.652549 71289 custom_operator.cc:723] Operator (farthest_point_sample) has been registered.
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:461: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
interp=Image.BILINEAR):
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1502: DeprecationWarning: NEAREST is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.NEAREST or Dither.NONE instead.
'nearest': Image.NEAREST,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1503: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
'bilinear': Image.BILINEAR,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1504: DeprecationWarning: BICUBIC is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BICUBIC instead.
'bicubic': Image.BICUBIC,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1505: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
'box': Image.BOX,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1506: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
'lanczos': Image.LANCZOS,
/home/gabriel/Paddle3D/paddle3d/transforms/transform.py:1507: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
'hamming': Image.HAMMING
ortools not installed, install it by "pip install ortools==9.1.9490" if you run BEVLaneDet model
Traceback (most recent call last):
File "deploy/iassd/python/infer.py", line 23, in <module>
from paddle3d.ops.pointnet2_ops import (ball_query, farthest_point_sample,
ImportError: cannot import name 'ball_query' from 'pointnet2_ops' (/home/gabriel/miniconda3/envs/paddle_37/lib/python3.7/site-packages/pointnet2_ops-0.0.0-py3.7-linux-x86_64.egg/pointnet2_ops.py)
Even if i omit --run_mode trt_fp32, it still fails with the same error message. Checking the docs, I noticed that it specified that it needs cuda==11.2, so i created another environment with conda install paddlepaddle-gpu==2.4.1 cudatoolkit=11.2 -c https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/Paddle/ -c conda-forge.
(paddle-37-cuda-11_2) gabriel@gabriel-Z790-AORUS-ELITE-AX:~/Paddle3D$ python3.7 deploy/iassd/python/infer.py --model_file outputs/iassd_kitti/export/iassd.pdmodel --params_file outputs/iassd_kitti/export/iassd.pdiparams --lidar_file datasets/KITTI/testing/velodyne/000000.bin --gpu_id 0
ortools not installed, install it by "pip install ortools==9.1.9490" if you run BEVLaneDet model
2024-09-03 11:36:13,607 - WARNING - No custom op iou3d_nms found, try JIT build
Compiling user custom op, it will cost a few seconds.....
2024-09-03 11:36:14,353 - INFO - iou3d_nms builded success!
2024-09-03 11:36:14,353 - WARNING - No custom op pointnet2_ops found, try JIT build
Compiling user custom op, it will cost a few seconds.....
2024-09-03 11:36:14,942 - INFO - Re-Compiling pointnet2_ops.so, because specified cflags have been changed. New signature 36348e91a4cc0bf4ae6ca42e5dea053f has been saved into /home/gabriel/.cache/paddle_extensions/pointnet2_ops/version.txt.
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
W0903 11:36:30.812824 74126 custom_operator.cc:723] Operator (nms_normal_gpu) has been registered.
W0903 11:36:30.812853 74126 custom_operator.cc:723] Operator (nms_gpu) has been registered.
W0903 11:36:30.812855 74126 custom_operator.cc:723] Operator (boxes_overlap_bev_gpu) has been registered.
W0903 11:36:30.812857 74126 custom_operator.cc:723] Operator (boxes_iou_bev_gpu) has been registered.
W0903 11:36:30.812858 74126 custom_operator.cc:723] Operator (boxes_iou_bev_cpu) has been registered.
2024-09-03 11:36:30,819 - INFO - pointnet2_ops builded success!
Traceback (most recent call last):
File "deploy/iassd/python/infer.py", line 23, in <module>
from paddle3d.ops.pointnet2_ops import (ball_query, farthest_point_sample,
File "/home/gabriel/Paddle3D/paddle3d/ops/__init__.py", line 189, in __getattr__
attr, self.modulename, module.__file__))
ImportError: cannot import name 'ball_query' from 'pointnet2_ops' (/home/gabriel/.cache/paddle_extensions/pointnet2_ops/pointnet2_ops_140388231238848.py)
System:
I am trying to train and run IASSD using tensorrt, I followed the installation instructions for python 3.7, my environment:
and was able to train the model:
I was then able to export the model with
python tools/export.py --config configs/iassd/iassd_kitti.yaml --model outputs/iassd_kitti/epoch_1/model.pdparams --save_dir outputs/iassd_kitti/export
, however, there is aUserWarning
:Next, if I try to run inference,
python3.7 deploy/iassd/python/infer.py --model_file outputs/iassd_kitti/export/iassd.pdmodel --params_file outputs/iassd_kitti/export/iassd.pdiparams --lidar_file datasets/KITTI/testing/velodyne/000000.bin --gpu_id 0 --run_mode trt_fp32
i getImportError
:I thought maybe the operation was not built properly, so i then try to build it manually by
cd
intopaddle3d/ops
and runningpython setup.py install
. However, it seemed to fail at building centerpoint operations.Since I dont care about centerpoint, I commented out the operations in
paddle3d/ops/__init__.py
:After I did this, I was able to run
setup.py
successfully. However, I still cannot runinfer.py
Even if i omit
--run_mode trt_fp32
, it still fails with the same error message. Checking the docs, I noticed that it specified that it needscuda==11.2
, so i created another environment withconda install paddlepaddle-gpu==2.4.1 cudatoolkit=11.2 -c https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/Paddle/ -c conda-forge
.If I run
infer.py
, same error:What can be causing this issue?