#how to convert pth to link so i can use it as a module
1 messages · Page 1 of 1 (latest)
you need to upload the model to Hugging Face, check how there https://docs.aihub.wtf/essentials/voice-models/#uploading-to-hugging-face
Last update: Mar 8, 2024
hello what license should i use
nvm
i need to read
You have been rate-limited; you can retry this action later. If you're a new user, your limits will raise progressively over time. Get in touch with us at website@huggingface.co if you need access now.
?
Rate limited? Could u show a screenshot? Did you use your hugging face acc before or
again
i need to read
need to verify my accout
@vague ospreysorry for tagging you
how can i paste my link in this colab
so i can have my voice in path and index
U have to go to download and paste the link to download your model
in your hugging face repository of the model, next to the model.zip file, there should be a lil download icon, you need to right click it and do copy link and paste that one instead, cus thats the direct download link
ahh okeyy
still
i dont know what wrong?
@vague osprey sorry again for tag por favor
mmm could u send me the link lemme check
Ayo? @supple tapir level 3 !!! 
@vague osprey
theres a macos folder in the model .zip for some reason, u should delete that
are you using a mac
also theres for sure a way to auto remove that folder when zipping something in mac, mm this is what i found https://apple.stackexchange.com/questions/288568/how-to-stop-the-hidden-macosx-folder-from-being-created-when-compressing-files, you could try to do some searches for the next times tho
the model sounds bad?
did you use the tensorboard while making the model ?
i used
Ultimate vocal remover
i recorded my voice in memo
then i removed background noises
and i upload it
Your model probably overtrained
Also how long was ur dataset
Basically, epochs are the ubit of measurement of the cycle of training the ai, if u put too less it undertrains, if u put too much it overtrains, making it worse in both cases, thats why we use the tensorboard which we find the point where its not overtraining and be sure that it wont be ruined
I train him 500 times
I will see that
U mean u putted 500 epochs right
U may wanna check this https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/
Last update: Feb 10, 2024
yw