Haohe Liu / 刘濠赫
Haohe Liu / 刘濠赫
You can use this Matlab toolbox: https://www.mathworks.com/matlabcentral/fileexchange/40128-filter-bank-design.
The training code will be available in this repo https://github.com/haoheliu/AudioLDM-training-finetuning. Will work on it when I got time.
Oh yes I forgot this dependency. You can do the following ``` sudo apt-get install espeak ```
The checkpoint for Text-to-Speech will be released later.
@BestVicky Please checkout this repo https://github.com/haoheliu/AudioLDM-training-finetuning. Thanks
AudioLDM 2 is also capable of style transfer. But some extra coding is needed to achieve this. You can refer to the style transfer code in AudioLDM v1 for more...
The project is in active building. I'll add that in later.
The inference code is ready now. Please checkout the main branch
Yes I think you can use your higher sampling rate audio (>16kHz) to finetune our 16kHz model. That sounds good to me.
It's super easy to use your own dataset. I'll update the readme with more tutorials later.