TRI-ML / realtime_panoptic

Official PyTorch implementation of CVPR 2020 Oral: Real-Time Panoptic Segmentation from Dense Detections
MIT License
115 stars 25 forks source link

Torchscript c++ #14

Open Joywalker opened 3 years ago

Joywalker commented 3 years ago

I'm trying to trace the model in order to do inference in c++ but I've encountered problems when running on cuda. When doing inference with a CPU model it works fine, but when exporting a cuda model, i get the following error.

1c39dbf8-3d33-49e0-a6c5-c871204b75db

Was anyone able to do inference with cuda in c++? Any response will be highly appreciated.