Transcript is from an independent captioning model — not RABBiT's input.
loading…
Replay a prediction example
Precomputed predictions, synchronized with the original audio.
0:00 / 0:24
Microphone inference (experimental)
Optional captions download a separate speech-recognition model. They are display-only.
Load the model first; then start your microphone. Weights are cached when browser storage allows. Audio is processed on this device. Allow about 10 seconds of audio context plus CPU processing time before the first prediction.
Predicted activation
deact.act.
Drag to rotate; scroll to zoom. With the brain focused, use arrow keys to rotate and +/− to zoom. Two halves show the left and right hemispheres (pial fsaverage6).