Tatsuya Matsushima's Projects
Code for "TD-MPC2: Scalable, Robust World Models for Continuous Control"
Differentiable ODE solvers with full GPU support and O(1)-memory backpropagation.
Code for the paper "Offline Reinforcement Learning as One Big Sequence Modeling Problem"
RoboCup challenge implementations
Codebase for "Uni[MASK]: Unified Inference in Sequential Decision Problems"
Python parser for URDFs
Official Algorithm Implementation of Paper "VIMA: General Robot Manipulation with Multimodal Prompts"
Official repository for "VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training"
A Python framework for high performance GPU simulation and graphics
Deep Reinforcement Learning framework