Will you release the training code of VAE.

[serend1p1ty]

Feature request / 功能建议

There is not much description of the training details of VAE in the paper, including the weights of various losses, the implementation of discriminators, learning rate scheduler, and so on.

Will you release the training code of VAE?

Motivation / 动机

Reproduce your VAE.

Your contribution / 您的贡献

None

[zRzRzRzRzRzRzR]

We currently don't have any related plans, but we are working on fine-tuning code based on Diffusers models. We can consider adding fine-tuning for VAE as well.

[serend1p1ty]

Thanks for your reply.

[serend1p1ty]

I have trained a stronger VAE than CogVideoX using Open-Sora-Plan's code.