Search papers, labs, and topics across Lattice.
1
0
3
Turn your robot's clumsy pre-trained behaviors into expert-level skills with DICE-RL, a surprisingly stable and efficient RL fine-tuning method.
