Srinivas Billa

srinivasbilla

AI & ML interests

None yet

Recent Activity

Organizations

None yet

srinivasbilla's activity

view reply

It's not prompted. The source Audio had that emotional context and the model simply copied it.

New activity in srinivasbilla/llasa-8b-tts 6 days ago
New activity in srinivasbilla/llasa-3b-tts 6 days ago
view reply

around 10gb, and around 300 chars is the sweet spot. you can chunk text and do it though

New activity in srinivasbilla/llasa-3b-tts 2 months ago
view reply

I had a look at both, it seems doable. Ill try follow the repeng example. But its a bit confusing how they generate the dataset