: Weights for the CaRL planner, which is noted as a top open-source RL planner for urban driving scenarios.
: A 2026 model that uses RL to generate adversarial code samples to test the vulnerabilities of programming language models. (2025) paper?
: CaRL: Learning Scalable Planning Policies with Simple Rewards
The zip file typically contains the codebase and pre-trained weights for the following:
: Proceedings of the Conference on Robot Learning (CoRL), 2025