ErickRosete/tacorl

TACO-RL: Latent Plans for Task-Agnostic Offline Reinforcement Learning

29
/ 100
Experimental

TACO-RL helps roboticists and automation engineers teach robots many different tasks efficiently by leveraging large amounts of previously collected, unorganized robot demonstration data. It takes raw, unlabeled video and robot movement data as input and produces a versatile robot control policy that can perform various complex actions in the real world or simulations. This is for researchers and engineers developing general-purpose robots or automated systems.

No commits in the last 6 months.

Use this if you need to train a robot to perform a wide variety of tasks using existing, diverse, uncurated offline datasets, rather than collecting new data for each specific task.

Not ideal if you have very little existing offline data or if your robot only needs to perform a single, highly specific, and unchanging task.

robotics automation robot-learning robot-control robot-training
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 6 / 25

How are scores calculated?

Stars

30

Forks

2

Language

Python

License

MIT

Last pushed

Jan 26, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/ErickRosete/tacorl"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.