That Sounds Right: Auditory Self-Supervision for Dynamic Robot Manipulation
For robot rollouts and data collection: Refer sound-supervised-bot
- Clone the repo
git clone git@github.com:abitha-thankaraj/audio-robot-learning.git
cd audio-robot-learning/
- Install the required python packages:
conda env create -f setup/environment.yaml
- Install the
aurl
library
pip install -e .
- Download the dataset here. Update path to the data_dir in your dataset configs.
To pretrain the self supervised encoder for AuRL:
cd aurl
python pretrain.py dataset=<task_name>/mel ssl=<aurl_variant>
To finetune aurl:
cd aurl
python train.py encoder_dir=<ssl_encoder_save_dir> dataset=<task_name>/mel