{
  "title": "Training labels and runtime conditions",
  "caption": "The author made this π0.7 conditioning diagram from Physical Intelligence’s source figure. Training uses labeled instructions, subgoal images and episode metadata. Inference can use task/subtask context, observation memory, desired metadata and world-model subgoals in a supported mode. The prompt interface receives these conditions before continuous action generation. Official performance demos are not shown, and visual subgoals are not necessary for all actions. The playback is an example, and animation time is not measured runtime.",
  "alt": "π0.7 comparison: annotations train weights. Fixed weights and runtime context generate action, with an optional world-model visual goal.",
  "sources": [
    "https://www.pi.website/blog/pi07"
  ],
  "limitations": [
    "The playback is an example. Playback time is not measured runtime.",
    "The two inference scenarios do not use training annotations as runtime inputs.",
    "The training scenario does not activate action generation.",
    "The supported condition modalities are not the same for all operating modes."
  ]
}