Search results

报错RuntimeError: The size of tensor a (4608) must match the size of tensor b (16896) at non-singleton dimension 2

Status: Open.

#15 In Huage001/CLEAR;

· hanhanhanhanhan929 opened on Oct 6, 2025

蒸馏过程是否是必要的

Status: Open.

#14 In Huage001/CLEAR;

· kxding opened on Apr 13, 2025

Error when training the proposed method

Status: Open.

#13 In Huage001/CLEAR;

· syf-fgnb opened on Feb 16, 2025

把原有的注意力机制统统换成q不变，kv圆形区域内？新的网络和老师网络蒸馏方法训练？

Status: Open.

#12 In Huage001/CLEAR;

· henbucuoshanghai opened on Feb 12, 2025

Speedup should be much more than 6.3x on 8K resolution?

Status: Open.

#11 In Huage001/CLEAR;

· sjtuzq opened on Jan 29, 2025

Can this method be trained on images with non-1:1 resolution.

Status: Open.

#10 In Huage001/CLEAR;

· super782 opened on Jan 21, 2025

How to reproduce the 6 times speed up reported in the paper?

Status: Open.

#8 In Huage001/CLEAR;

· jzhang38 opened on Jan 18, 2025

Inference error

Status: Open.

#6 In Huage001/CLEAR;

· phyllispeng123 opened on Dec 26, 2024

Synthetic Dataset

Status: Open.

#5 In Huage001/CLEAR;

· A-BigBao opened on Dec 25, 2024

LTX-Video support?

Status: Open.

#4 In Huage001/CLEAR;

· yoavhacohen opened on Dec 24, 2024

Inference time saving for 1024 * 1024

Status: Open.

#3 In Huage001/CLEAR;

· phyllispeng123 opened on Dec 24, 2024

Have you considered implementing it in ComfyUI?

Status: Open.

#2 In Huage001/CLEAR;

· sungod189 opened on Dec 24, 2024