Define the Prediction Output

[chadlagore]

Here is a good idea, though we probably won't transcribe. So the body will be a binary blob (thats OK!). This work is to be done here.

[
    {
          "speaker": 0,
          "start_time": 148172245,
          "end_time": 148172251,
          "body": "0xafb103..."
    },
    {
          "speaker": 1,
          "start_time": 148172251,
          "end_time": 148172253,
          "body": "0xa99f9012..."
    }
]