#I need suggestion
1 messages · Page 1 of 1 (latest)
take a look at: https://discord.com/channels/1159260121998827560/1235952130855010365
TITAN may help produce clean results but you can try others, though your dataset quality is still the significant factor
so, you're saying that TITAN is the best among all the other models?
go give it a try before rashing your judgment.
Itaila is also quite good despite being only 32k sample rate, and although it was trained on Italian dataset should also work fine on English and other languages.
Which language and sample is your model?
look at #1233407331405004954
Alright, I'll try
Japanese
Maybe Snowie 3.1 could help as its made for Japanese (& Russian)
English pretrains also work fine on it
I'll give it a try once I resolve my issue. Btw, is this the correct link to fill the "d" part: https://huggingface.co/MUSTAR/SnowieV3.1-40k/resolve/main/D_SnowieV3.1_40k.pth?download=true
Ayo? @ruby oxide level 4 !!! 