alumae / kaldi-offline-transcriber

Offline transcription system for Estonian using Kaldi
Other
226 stars 55 forks source link

json2trs dies on no speakers #25

Open hillar opened 3 years ago

hillar commented 3 years ago

https://github.com/alumae/kaldi-offline-transcriber/blob/caadc224137c0c86e71e6bb8f65ab2b8d718f09b/local/json2trs.py#L40

Traceback (most recent call last):
  File "./local/json2trs.py", line 64, in <module>
    print_sections(trans["sections"], trans["speakers"])
  File "./local/json2trs.py", line 40, in print_sections
    print('<Turn speaker="%s" startTime="%0.3f" endTime="%0.3f">' % (speakers[turn["speaker"]]["transcriber_id"], turn["start"], turn["end"]))
KeyError: 'S0'

tdnn_7d_online_pruned_rescored_main_rnnlm_unk.punctuated.normalized.json

{
    "speakers": {},
    "sections": [
        {
            "start": 0,
            "type": "speech",
            "end": 110.4,
            "turns": [
                {
                    "transcript":