lucidrains/q-transformer
Implementation of Q-Transformer, Google DeepMind's offline reinforcement learning method for robotics using transformer-based autoregressive Q-functions.
Not currently ranked — collecting fresh signals.
star history
This repository provides a PyTorch implementation of Q-Transformer, a method for scalable offline reinforcement learning. It reproduces the autoregressive Q-learning formulation for handling continuous action spaces in robotics applications. The implementation includes a vision transformer backbone for processing high-dimensional observations like images, along with tools for environment interaction, replay memory management, and training agents using Q-learning on discrete action bins.
Frequently asked
- What is lucidrains/q-transformer?
- Implementation of Q-Transformer, Google DeepMind's offline reinforcement learning method for robotics using transformer-based autoregressive Q-functions.
- Is q-transformer open source?
- Yes — lucidrains/q-transformer is open source, released under the MIT license.
- What language is q-transformer written in?
- lucidrains/q-transformer is primarily written in Python.
- How popular is q-transformer?
- lucidrains/q-transformer has 407 stars on GitHub.
- Where can I find q-transformer?
- lucidrains/q-transformer is on GitHub at https://github.com/lucidrains/q-transformer.