Placeholder chapter page. Sections are produced by the book-skills pipeline.
Sections
- 58.1 From perception to action: the VLA idea
- 58.2 RT-1/RT-2 and web-scale knowledge transfer
- 58.3 Open generalist policies (Octo, OpenVLA)
- 58.4 Flow-matching and open-world VLAs (π0/π0.5)
- 58.5 Humanoid and general robot foundation models (GR00T, Gemini Robotics)
- 58.6 Sim-to-real and Open X-Embodiment data
- 58.7 Evaluation, safety, and generalization
Lab 58
fine-tune or evaluate an open VLA policy on a manipulation benchmark subset (LIBERO/CALVIN).