ZDisket

Multispeaker

13

First I wanna say that I love this repo, it performs excellently with many of my small (down to 80 seconds) and noisy datasets, having good transfer learning. Multispeaker support...

I managed to export jaywalnut310's VITS with TorchScript - [x] Write VITS class with speed control - [x] Integrate single-speaker VITS into the program - [ ] Add support for...

This PR will add DE-VITS (codename VITS iCMOTT "6 Elements"). It will replace VITSTM in-place. - [x] Create BERT tokenizer - [x] Create BERT model inference - [x] Create DE-VITS...

Support v-prediction objective, fix some SDXL models outputting clouds

2

If you try to use models trained with the v-prediction objective with the current repo, you'll get nonsense output. I detailed in [this issue](https://github.com/axodox/axodox-machinelearning/issues/25#issue-2152960991) This PR adds v-prediction to both...

Some SD finetunes (and the official Stable Diffusion 2) use v-prediction because they [observe that it's better](https://medium.com/@zljdanceholic/three-stable-diffusion-training-losses-x0-epsilon-and-v-prediction-126de920eb73) Currently, v-prediction models will export with Olive and run fine with the ONNX...