From the source
Hugging Face announces the integration of the Decision Transformer, an offline reinforcement learning method, into the transformers library and the Hugging Face Hub, along with nine pre-trained model checkpoints for continuous control tasks.
