Cuda Error in copyToDevice: 222 (the provided PTX was compiled with an unsupported toolchain.)

[thudongbn]

Env

• Gtx 1650
• Cuda version 11.3
• TensorRT version 7.2.3.4
• docker image nvcr.io/nvidia/tensorrt:21.05-py3
• Driver Version: 460.91.03

About this repo

• master
• ctdet_coco_dla_2x.pth

Your problem

• python3 centernet.py -m ctdet_coco_dla_2x.pth -s
• I compiled an engine with your code but at runtime I see this error:

[TensorRT] ERROR: /dong/TensorRT/plugin/dcnv2Plugin/dcnv2Plugin.cpp (180) - Cuda Error in copyToDevice: 222 (the provided PTX was compiled with an unsupported toolchain.) Traceback (most recent call last): File "centernet.py", line 336, in main(args) File "centernet.py", line 311, in main with Centernet_dla34(weights).engine as engine: File "centernet.py", line 39, in init self.engine = self.build_engine() File "centernet.py", line 293, in build_engine self.populate_network() File "centernet.py", line 271, in populate_network dla_up = self.add_dla_up(y, first_level, 'module.dla_up') File "centernet.py", line 241, in add_dla_up input_tensors = self.add_ida_up( File "centernet.py", line 219, in add_ida_up proj = self.add_deform_conv( File "centernet.py", line 206, in add_deform_conv DCN = dcnCreator.create_plugin( RuntimeError: std::exception

[wang-xinyu]

@SsisyphusTao Please assist.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.