Jaeger416
Jaeger416
Hi, it seems that you haven't released the codes related to initialization from SD's 2D VAE. Would you like to or do you plan to release that part?
Could you please let me know which version of Stable Diffusion you used for initializing the VAE ?
During train causalvae, It seems that the "logvar" parameter in discriminator loss can't receive any gradient, since opt1.zero_grad() clear the gradient from gan loss and logvar is not used in...
请教下资源出处
请问下线上资源是从哪里获取的呢,大佬能提供下链接不,最近想做一些类似但是AI base的项目
 Hi, I noticed you set bbox_crop to false in stage_2.yaml and keep quote it in the training procedure  but the tgt_box is necessary in video_dataset.__getitem__ and could only...
When I look in to each parameter's gradient, I found all attention blocks except for the last one in guidance encoder get nan gradient. If not use accelerate just torch...