UniversalTechno

I did using the default script on the repository https://github.com/lucidrains/spear-tts-pytorch and I got a folder output_folder with models from 0 TO 9.pt so I replaced /path/to/trained/model.pt by /output_folder/model.pt but still...

you mean i should use directly [trainer.py](https://github.com/lucidrains/spear-tts-pytorch/blob/main/spear_tts_pytorch/trainer.py) is there any pre trained model so i just can test quickly or just any instruction for how to train many thanks

I think the best method its to to look first if there are a features or to modifies the existing great translation model like plachtaa mentionned MetaAI's SeamlessM4T. or others...

yes for sure its not about library , its common issue in linux

> it will re-use original as base code and we will build on it to make it End-2-End. Original model just predicts mel and then uses hifigan to convert mel-wav;...

Thanks I will do it right now On Tue, Oct 24, 2023, 10:30 p0p ***@***.***> wrote: > Add me on discord p0p4k > > — > Reply to this email...

Sure I can take in charge please can you share the data and méthodes requested for training . On Thu, Nov 2, 2023, 08:18 p0p ***@***.***> wrote: > Hi, do...