Note: This upscaler is not mine

Credit to Helaman

Originally uploaded at: https://openmodeldb.info/models/4x-NomosUniDAT-bokeh-jpg

Info on version 2

The model and everything is the same, but I have converted it to .safetensors

Note: I have had issues getting Forge and Automatic1111 to load the .safetensors upscale model, but it works like a charm in ComfyUI.

Info

Trained on DAT with Adan, U-Net SN, huber pixel loss, huber perceptial loss, vanilla gan loss, huber ldl loss and huber focal-frequency loss, on paired nomos_uni (universal dataset containing photographs, anime, text, maps, music sheets, paintings ..) with added jpg compression 40-100 and down_up, bicubic, bilinear, box, nearest and lanczos scales. No blur degradation had been introduced in the training dataset to keep the model from trying to sharpen blurry backgrounds.

The three strengths of this model (design purpose):

• Multipurpose

• Handles bokeh effect

• Handles jpg compression

This model will not:

• Denoise

• Deblur

Where does it go?

To use this (and other DAT upscalers) with Automatic1111 and Forge follow these steps

• Download the file either here or from the source

• Place the file in \webui\models\DAT\

• Restart your webui

Note: If you have issues getting the model to work, change the file name from .pt to .pth