Search results

关于分布匹配损失的计算

Status: Open.

#84 In tianweiy/DMD2;

· zhangXJ147 opened on Dec 28, 2025

ValueError: <code>FlatParameter</code> requires uniform dtype but got torch.bfloat16 and torch.float32

Status: Open.

#82 In tianweiy/DMD2;

· J-zin opened on Dec 21, 2025

How to train LORAs on DMD2 SDXL Checkpoints

Status: Open.

#81 In tianweiy/DMD2;

· todochenxi opened on Nov 17, 2025

which model to use in ComfyUI - the mentioned ones fail

Status: Open.

#78 In tianweiy/DMD2;

· yekruz opened on Aug 30, 2025

What is the difference between DMD and VSD?

Status: Open.

#77 In tianweiy/DMD2;

· heishuini opened on Aug 23, 2025

Can I continue finetune a DMD distilled model？

Status: Open.

#76 In tianweiy/DMD2;

· muziyongshixin opened on Aug 14, 2025

训练的时候怎么设置对应的loss ratio呢？

Status: Open.

#73 In tianweiy/DMD2;

· littletomatodonkey opened on Aug 1, 2025

I really like DMD2 - not because of its speed. But because with DMD SDXL improves distortions caused by VAE, details like small objects are much more coherent.

Status: Open.

#72 In tianweiy/DMD2;

· madeirafranciscoluxio opened on Jun 20, 2025

How to choose scheduler after the distillation?

Status: Open.

#71 In tianweiy/DMD2;

· yyfcc17 opened on Jun 19, 2025

Request for a more precise requirements.txt with exact dependency versions

Status: Open.

#70 In tianweiy/DMD2;

· hrzhou-ai opened on Jun 7, 2025

why compute_loss_fake when gan_loss

Status: Open.

#69 In tianweiy/DMD2;

· vv12kant opened on Apr 2, 2025

Question on how DMD loss is calculated in code

Status: Open.

#68 In tianweiy/DMD2;

· ruben-nithyaganesh opened on Mar 18, 2025