Learning to Walk in Minutes Using Massively Parallel Deep RL

Автор: Robotic Systems Lab: Legged Robotics at ETH Zürich

Загружено: 5 окт. 2021 г.

Просмотров: 71 579 просмотров

Описание:

We present a training set-up that achieves fast policy generation for real-world robotic tasks by using massive parallelism on a single workstation GPU. The parallel approach allows training policies for flat terrain in under 4 minutes, and in 20 minutes for uneven terrain.

Paper accepted to CoRL 2021.
Code and paper: https://leggedrobotics.github.io/legg...

Learning to Walk in Minutes Using Massively Parallel Deep RL

Доступные форматы для скачивания:

Скачать видео mp4

Информация по загрузке:

Скачать аудио mp3

Похожие видео

DeepGait: Planning and Control of Quadrupedal Gaits using Deep Reinforcement Learning (Presentation)

DeepGait: Planning and Control of Quadrupedal Gaits using Deep Reinforcement Learning (Presentation)

Градиентный спуск, как обучаются нейросети | Глава 2, Глубинное обучение

Градиентный спуск, как обучаются нейросети | Глава 2, Глубинное обучение

Making Real-World Reinforcement Learning Practical

Making Real-World Reinforcement Learning Practical

Perceptive Locomotion through Nonlinear Model Predictive Control

Perceptive Locomotion through Nonlinear Model Predictive Control

Cybersecurity Architecture: Five Principles to Follow (and One to Avoid)

Cybersecurity Architecture: Five Principles to Follow (and One to Avoid)

Reinforcement Learning Series: Overview of Methods

Reinforcement Learning Series: Overview of Methods

I Built an Internal Cycloidal Robotic Actuator

I Built an Internal Cycloidal Robotic Actuator

Reinforcement Learning - My Algorithm vs State of the Art

Reinforcement Learning - My Algorithm vs State of the Art

Design and Control of a Bipedal Robotic Character

Design and Control of a Bipedal Robotic Character

Визуализация внимания, сердце трансформера | Глава 6, Глубокое обучение

Визуализация внимания, сердце трансформера | Глава 6, Глубокое обучение