{
  "title": "Predict a subtask and then generate a chunk",
  "caption": "The author made this π0.5 inference diagram from Physical Intelligence’s source figure. A language-level subtask explicitly conditions continuous action generation with the current observation and original instruction. The two context cards show reuse, not two sensors. Token-level details and official illustrative images are not shown. The playback is an example, and animation time is not measured runtime.",
  "alt": "π0.5 flow: predict a language subtask, then condition a continuous action chunk on it, the current observation and original instruction.",
  "sources": [
    "https://www.pi.website/blog/pi05"
  ],
  "limitations": [
    "The playback is an example. Playback time is not measured runtime.",
    "No training or weight update occurs in this shown inference path."
  ]
}