DiffTORI: Differentiable Trajectory Optimization for Deep Reinforcement and Imitation Learning | Read Paper on Bytez