Jianyi Wang

I guess longer training and more data should help.

> > It is hard to say. For training, usually the longer, the better. After all, LDM is trained for 2.6M iterations with 256 batch size. The performance between different...

> > > > It is hard to say. For training, usually the longer, the better. After all, LDM is trained for 2.6M iterations with 256 batch size. The performance...

I guess the version of the torchvision is not correct?

Hi. It seems that your path to the dataset is not correct. The code did not find related images under the folder so the length of the dataset is 0.

Sry for the late reply. I am now busy with other projects. I am not sure of the reasons. I did not use additional data. We noticed that sometimes the...

> Thank you for your outstanding work！ I am very interested in the work with SD-Turbo. May I ask if its training is the same as SD-2.1? Do you only...

Hi. You are using a zoomed image, i.e., the LR image is already a 4x bicubic upsampled version of the original LR image. We guess that a too-large upsampling scale,...

You may test more images using the original images [here](https://github.com/zsyOAOA/ResShift/tree/journal/testdata/RealSet65). If you are interested in finding the reasons, you can first try upsampling these images with 16x to see if...

I tested on the original image using DDPM 50 steps and it works. I think DDIM 20 steps should work too.  The original image for your reference: