InstantMimic: A High Performance System for Learning Physics-based Skills in Seconds

2026-09-09

AI Summary

本文提出了InstantMimic系统，通过将整个训练流程GPU化来解决物理技能学习中的计算效率问题，实现快速训练。

Abstract

Physics-based character control is a long-standing challenge in computer graphics and robotics, requiring policies that satisfy complex dynamics while producing realistic motion. Recent Deep RL approaches, particularly imitation learning methods such as DeepMimic, have had broad impact beyond animation, influencing robotics by enabling agile and expressive behaviors. While these approaches achieve impressive results, they remain computationally inefficient to train in practice. Despite GPU-accelerated simulation, we find that end-to-end pipelines often underutilize hardware due to overheads outside the physics solver, caused by fragmented GPU kernels and CPU memory access in the critical path. We present InstantMimic, a system that addresses these inefficiencies by making the entire training loop GPU-native. Built on a GPU-native physics backend, our unified pipeline integrates simulation, environment computation, policy inference, and policy updates within a single execution flow. As a result, InstantMimic reduces training time for diverse physics-based skills to a few seconds and makes LLM-agent-driven hyperparameter search practical.

Problem

• Physics-based character control
• Deep RL
• Imitation learning
• GPU-accelerated simulation
• Hardware utilization

Innovation

• GPU-native
• physics-based skills
• training efficiency
• unified pipeline