TensorRT Conversion Issue "TypeError: pybind11::init(): factory function returned nullptr" about tensorrt-for-yolo-series

[Linaom1214]

Hi,

I ran the ONNX to TensorRT conversion using the following command:

$ python3 export.py -o yolov7-tiny.onnx -e yolov7-tiny-nms.trt -p fp16

However, I encountered the following issue:

Namespace(calib_batch_size=8, calib_cache='./calibration.cache', calib_input=None, calib_num_images=5000, conf_thres=0.4, end2end=False, engine='yolov7-tiny-nms.trt', iou_thres=0.5, max_det=100, onnx='/home/vahid/aid_ws/GitHub-Vbsm/YOLO/PyTorch-ONNX-TensorRT/yolov7-WingKinYiu/yolov7-tiny.onnx', precision='fp16', v8=False, verbose=False, workspace=1) [12/26/2023-13:37:00] [TRT] [W] Unable to determine GPU memory usage [12/26/2023-13:37:00] [TRT] [W] Unable to determine GPU memory usage [12/26/2023-13:37:00] [TRT] [I] [MemUsageChange] Init CUDA: CPU +0, GPU +0, now: CPU 23, GPU 0 (MiB) [12/26/2023-13:37:00] [TRT] [W] CUDA initialization failure with error: 35. Please check your CUDA installation: http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html Traceback (most recent call last): File "export.py", line 308, in main(args) File "export.py", line 266, in main builder = EngineBuilder(args.verbose, args.workspace) File "export.py", line 107, in init self.builder = trt.Builder(self.trt_logger) TypeError: pybind11::init(): factory function returned nullptr

Here are the Python environment dependencies:

$ python3 Python 3.6.9 (default, Mar 10 2023, 16:46:00) [GCC 8.4.0] on linux Type "help", "copyright", "credits" or "license" for more information.

import torch
print(torch.version)
1.10.1+cu102
import torchvision
print(torchvision.version)
0.11.2+cu102
import onnx
print(onnx.version)
1.8.1
import pycuda
import tensorrt
print(tensorrt.version)
8.6.1
exit()

I would appreciate any advice or solutions you can provide to resolve this issue. Thank you.

you can try this branch, https://github.com/Linaom1214/TensorRT-For-YOLO-Series/tree/cuda-python, this error may occur by pycuda, or change the pycuda version.

[VbsmRobotic]

Thank you for your advice; I've addressed the issue.