Driver error when deploying #163
Description
When I run following command
python /home/admin/workspace/deploy/MMDeploy/tools/deploy.py \
/home/admin/workspace/deploy/MMDeploy/configs/mmdet/detection/detection_tensorrt_dynamic-320x320-1344x1344.py\
/home/admin/workspace/algorithms/mm_algorithms/mmdetection/configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py \
/home/admin/workspace/algorithms/mm_algorithms/mmdetection/checkpoints/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth \
/home/admin/workspace/algorithms/mm_algorithms/mmdetection/demo/demo.jpg \
--work-dir ./sdk_results \
--device cuda \
--dump-info
I got error messages
2022-02-16:19:50:22,root ERROR [utils.py:41] Driver error:
Traceback (most recent call last):
File "/home/admin/workspace/deploy/MMDeploy/mmdeploy/utils/utils.py", line 36, in target_wrapper
result = target(*args, **kwargs)
File "/home/admin/workspace/deploy/MMDeploy/mmdeploy/backend/tensorrt/onnx2tensorrt.py", line 72, in onnx2tensorrt
device_id=device_id)
File "/home/admin/workspace/deploy/MMDeploy/mmdeploy/backend/tensorrt/utils.py", line 114, in create_trt_engine
engine = builder.build_engine(network, config)
RuntimeError: Driver error:
2022-02-16 19:50:23,980 - mmdeploy - ERROR - onnx2tensorrt of ./sdk_results/end2end.onnx failed.
It looks like a problem related to Nvidia Driver. Here is my envs
NVIDIA: Tesla T4
Driver Version: 470.103.01
CUDA Version: 11.4
tensorRT Version: 7.2.3.4