Open Source Robotics

精選高品質的開源機器人專案、工具與軟體,推動機器人社群向前發展。

HuggingFace Robotics Course

  • 提供者: Hugging Face
  • 講師: Hugging Face Robotics Team
  • 程度: beginner
  • 時長: Self-paced (30-45 min per unit)
  • 類別: Machine Learning
  • 語言: English
  • 價格: 免費
  • 標籤: robotics, machine learning, LeRobot, imitation learning, reinforcement learning, diffusion models, pytorch, manipulation

一門全面的免費課程,以 LeRobot 銜接機器人學的理論與實務。學習古典機器人學的基礎,以及以學習為基礎的現代方法,包括模仿學習、強化學習,以及 PI0 與 SmolVLA 等通用機器人策略。

為何值得修讀

This is one of the most accessible entry points into robotics AI. Created by Hugging Face — the team behind LeRobot — it uniquely bridges classical robotics theory with modern learning-based approaches. Unlike most courses that focus on either theory or practice, this one gives you both, with hands-on exercises using the actual LeRobot framework used in cutting-edge research.

你將學到

  • Classical robotics foundations: kinematics, dynamics, and control theory
  • Imitation learning techniques for teaching robots from demonstrations
  • Reinforcement learning applied to real robotic systems
  • How diffusion models are used for robot policy learning
  • Generalist robot policies (PI0, SmolVLA) and how they work
  • Practical skills with the LeRobot framework and PyTorch

課程大綱

The course is structured as self-paced units (30-45 minutes each). It starts with classical robotics foundations, then progressively introduces modern learning-based methods — from imitation learning to reinforcement learning to diffusion models. Each unit combines theory with practical LeRobot exercises. The course culminates with generalist robot policies that can handle multiple tasks.

核心主題

  • Introduction to robotics and the LeRobot ecosystem
  • Classical robotics: kinematics, dynamics, and control
  • Imitation learning and behavior cloning
  • Reinforcement learning for robotics
  • Diffusion policies for robot manipulation
  • Generalist robot policies: PI0 and SmolVLA
  • Deploying learned policies on real robots

先備知識

  • Basic Python programming experience
  • Familiarity with machine learning concepts (helpful but not required)
  • No robotics hardware needed — all exercises run in simulation

適用對象

  • Beginners curious about robotics and AI
  • Machine learning practitioners looking to apply their skills to robotics
  • Students wanting a free, high-quality introduction to robot learning
  • Hobbyists building with LeRobot-compatible hardware (SO-100, LeKiwi)

連結

  • 網站: https://huggingface.co/learn/robotics-course/unit0/1
  • GitHub: https://github.com/huggingface/robotics-course

本網站需要 JavaScript 才能正常運作。請啟用 JavaScript 以瀏覽我們的開源機器人專案。