search

DLR-RM / AugmentedAutoencoder

Official Code: Implicit 3D Orientation Learning for 6D Object Detection from RGB Images
MIT License
336 stars 97 forks source link

Porting to TensorRT #86

Closed davideCremona closed 3 years ago

davideCremona commented 3 years ago

Hi, is it possible to port a trained Augmented Auto-Encoder to TensorRT for inference on embedded devices (e.g. Jetson TX2 that you used in the demo)

Thanks, Davide.

MartinSmeyer commented 3 years ago

I did not use tensorrt in my demo, however it should be possible. The AAE is also integrated in the mostly closed source Nvidia Isaac sdk for Jetson platforms where it is presumably optimized using tensorrt:

https://docs.nvidia.com/isaac/archive/2020.1/packages/object_pose_estimation/doc/ae_3d_object_pose_estimation.html

davideCremona commented 3 years ago

This is great! Thank you for the link, I'll surely use it.