Collection of Hinglish dataset - Githubissues

ChakshuGautam / whisper-hinglish

1 stars 0 forks source link

Collection of Hinglish dataset #3

Open rayaanoidPrime opened 6 months ago

rayaanoidPrime commented 6 months ago

Compile sources of Hinglish audio (100 hours) that are:

Collect Hinglish audio sources:
- Podcasts | Conversational (English heavy)
- [ ] 70% of the dataset or around 70 hours
- Product review | Monologues | YouTube videos (Hindi heavy)
- [ ] 30% of the dataset or 30 hours
- Note: Sources with transcripts to be preferred