lawsonxwl

Question of Overfitting!

1

Thanks for solid work! I use 80 scenes in megadepth to train the superglue model but only after 5 epoch,the model converges and the validation loss explodes after epoch 5....

Could you tell me the validation scenes in Megadepth?

1

Also, if I resize the longest dimension of the image to 1600, how can I resize the keypoint?

使用源码config训练CAN-DWAP，在CROHME14测试集上得到的ExpRate比论文中要低1.4%

12

config未作任何修改，Wordrate为0.9005，ExpRate为55.6%，问下大佬这个是啥情况？为什么复现不了呢

Prompt: Landscape photography,highest resolution, 8K, realistic, detailed, dynamic, full detail, creative, awesome, perfect, unique, amazing, beautiful Settings: Steps: 30, Sampler: Euler a, CFG scale: 5, Seed: 12345, Size: 832x1216, Model...

A technical question: In the "sharpness" implementation, why isn't the gaussian degradation applyed on the unconditional branch but conditional branch?

2

According to the SAG paper 5.1(Improving Sample Quality of Diffusion Models Using Self-Attention Guidance), the blur guidance is performed on the unconditional model output, in order to guide the sampling...

Any plans to do some research on masking based removing methods?

3

It's a great work! I wonder if this method can be transferred to mask based removing? As instruction sometimes cannot convey the user's accurate intention! I have tried SD-Inpaint, but...

is the traindata model input correct？

10

### Describe the bug I try to fine tune the official sdxl brushnet with my own data, but get bad result, I wonder if my input is correct, below is...