A GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.

C++-based high-performance parallel environment execution engine (vectorized env) for general RL environments.