Jianyi Wang
Jianyi Wang
Hi. Thanks for your interest of our work! We are very busy recently. We may consider this later. Contributions are also really appreciated :)
> @IceClear hi Jianyi - did you see this I put together: https://github.com/WSJUSA/Comfyui-StableSR > > It has bugs because the way Comfy builds the Unet model is different (beyond my...
> came here to say this too. I absolutely love using this extension in Automatic1111 and often have to throw images into it instead of using comfyui workflows. sure wish...
You may have a try on the new 768 model to see if the result gets better.
Hi. Yes. you can run it w/o CFW.
If you would like to try the result w/o CFW, you can set --dec_w to 0. But the CFW module is not deleted and still leads to memory cost. You...
Hi. This is because you are using a tile size larger than your image size. See [here](https://github.com/IceClear/StableSR/blob/main/scripts/sr_val_ddim_text_T_negativeprompt_canvas_tile.py#L288C6-L288C20).
Hi. Sry for the late reply since I am quite busy recently. It seems like a size checking problem since the default tile size of sr_val_ddpm_text_T_vqganfin_oldcanvas_tile is 1280 and for...
Currently maybe not. I am busy with other projects so may not be available for it.
It is hard to say. For training, usually the longer, the better. After all, the official LDM seems to be trained for about 2.6M iterations with 256 batch size. The...