ny issues

Repositories
Issues
Comments

Results 5 issues of

ny

how to support different size in the latent space when training

hello, thanks for the great work. there is a question when i train the control net. when inference the denoise steps, the ddim sample size is not just the64*64, how...

the Training details of ella

ELLA 后出来的embedding 和diffusion model的 alignment 表现是怎么表现的，可以讲一下吗？谢谢～

the training details for hunyuan dit control net

Thanks for your great work. Can you provide the training loss and training_steps in control net training?

Could you provide the learning rate and training time costs for each stage? and how much datasets, i just found it‘s learning slowly

maybe due to the skip net?

questions about tcd in calculate loss

Thanks for your awesome work. Here is some questions. 1.whether the second part of tn should infer with the same 1-rd ode function from the tn+k predicted of teacher model,...