{
  "title": "FAST encoding and continuous generation",
  "caption": "FAST encodes demonstrated action chunks as compact frequency-structured tokens. Flow matching jointly generates a continuous chunk from noise. Knowledge Insulation can use the token objective to train the backbone. Its continuous expert supplies inference-time actions. The author made this explanatory diagram. The playback is an example, and animation time is not measured runtime.",
  "alt": "Two-lane comparison: FAST encodes demonstrated chunks into tokens. Flow matching generates continuous chunks from noise.",
  "sources": [
    {
      "path": "content/blog/2026-05-25_from-pi0-to-pi07-open-pi-series/post.md",
      "description": "Article mechanism and retained citations"
    },
    "https://www.pi.website/research/fast",
    "https://www.pi.website/research/knowledge_insulation",
    "https://arxiv.org/abs/2410.24164"
  ],
  "limitations": [
    "The playback is an example. Playback time is not measured runtime.",
    "The two lanes show different interfaces. They are not two necessary consecutive runtime stages.",
    "The diagram does not give the detailed token codebook or numerical integration schedule."
  ]
}