What is this model?
I tried using it, and it seems to separate the inst (instrumental) into the low frequencies and vocals into the high frequencies. What would be some good use cases for this?
this model is for next release, I change to public this repo to upload, so Dont mind to this repo
For next releases, I would like to use your Mel Roformer cinematic speech separation, for separate two speakers or more.
btw, I tried to use test3.ckpt but it gives me an error on torch size.
Please, fix this model error, I tried to use your model and nothing, it gives me always the same error on torch size and everything! Or what model type is your new test model?
Have you got any updated config for your new model?
Gobzaluigi, I will upload new model soon about a day, And I must say this model is so big. config will be publicly soon, so please wait a bit.
Ok, thanks for your information.
Gobzaluigi, What is Mel Roformer cinematic speech separation? If I could understand that, I may be able to make for that model.
Gobzaluigi, What is Mel Roformer cinematic speech separation? If I could understand that, I may be able to make for that model.
I'll explain you. Mel Roformer cinematic speech separation will be a Mel Band Roformer model that isolates main speech from multiple speeches, and it's the speech separation model with multilingual speech (including spanish, both castillian and latin american)