Missing data

#6
by bonham79 - opened

Heya,

I'm loading the french split and it looks like there's some data missing:

DatasetDict({
    train: Dataset({
        features: ['audio', 'key', 'country', 'language', 'video_id', 'transcript_id', 'start_seconds', 'end_seconds', 'duration_seconds', 'asr_transcript', 'human_transcript', 'cer', 'wer', 'original_transcript_start_idx', 'original_transcript_end_idx'],
        num_rows: 19296
    })

Even at 20s a sample, this is only about ~100hrs.

DISCO org

Thanks for the report! I replied to your issue on GitHub as well. I'm currently re-uploading the missing files. I'll ping you here once the French dataset is fully synced.

Also, thanks! It looks like English was affected as well. German I think is within the reported range.

DISCO org

The French training split should be fixed by now. I'll look into the English split tomorrow afternoon and double check the other languages.

Love it, lemme check and I'll confirm on the user side.

Alrighty, files seem to be there for French. Thanksss

Sign up or log in to comment