HareshKarnan/sterling_corl23

STERLING: Self-Supervised Terrain Representation Learning from Unconstrained Robot Experience

Haresh Karnan*¹, Elvin Yang¹, Daniel Farkash², Garrett Warnell¹, Joydeep Biswas¹, Peter Stone^1,3

¹The University of Texas at Austin, ²Cornell University, ³Sony AI

Project Page | arXiV | Summary Video

---

Training representations using STERLING

To launch the training script for STERLING:

python train.py --data_config_path spot_data/test.yaml

Note : The code release only contains the training script for STERLING. The ROS node to process rosbags from the spot for the offline pre-processing step will be released shortly!.