search

v-iashin / SpecVQGAN

Source code for "Taming Visually Guided Sound Generation" (Oral at the BMVC 2021)
https://v-iashin.github.io/SpecVQGAN
MIT License
347 stars 40 forks source link

Obout train Mel-GAN #14

Open yangdongchao opened 2 years ago

yangdongchao commented 2 years ago

Hi, I want to ask whether your code about training Mel-GAN vocoder is support multiple GPUs? In you paper, you use one single GPU training about 14 days. So I want to ask you whether we can use multiple GPUs to decrease the training time.

v-iashin commented 2 years ago

Hi, You are right, it supports only single GPU training. I have not tried to use multiple GPUs to train MelGAN. The good thing is that it will start producing good results pretty soon so you could start building on top of that checkpoint early on.

yangdongchao commented 2 years ago

Hi, You are right, it supports only single GPU training. I have not tried to use multiple GPUs to train MelGAN. The good thing is that it will start producing good results pretty soon so you could start building on top of that checkpoint early on.

Thank you very much.