Feature Extraction
NeMo

Fine-tuning

#3
by yukiarimo - opened

Hello! Can you please give me a step-by-step fine-tuning guide of this codec (this exact checkpoint)? Thanks!

There is a ft guide and code on NVIDIA`s GitHub ))

It’s for old models

Is there any update on this issue?
The NanoCodec has become the main bottleneck for TTS inference speed.
I want to train a NanoCodec specifically for telephony systems, where 8 kHz audio is required.
Using a 22 kHz codec is too slow and unnecessary for phone-quality speech.

Sign up or log in to comment