Jianyi Wang
Jianyi Wang
Hi～ thanks for your interest. Do you mind sharing the failure case for 768 model？Without prompts, is there still artifacts？Maybe we can further improve it if possible.
I think you may use the wrong model.
The most possible reason is that your cpu memory is not enough.
Sry. Currently I have no ideas about which part leads to high memory cost. So no solution from my side.
Maybe you can check if the model ckpt is loaded twice since in my current setting, the diffusion model and vqgan are initialized from the same large ckpt and I...
> > Maybe you can check if the model ckpt is loaded twice since in my current setting, the diffusion model and vqgan are initialized from the same large ckpt...
You can replace the model architecture accordingly and load SD 1.5 for finetuning.
> thanks~ BTW, I am not quite sure but this repo should contain the model architecture of SD 1.5 since I used to try it using this repo : )....
> It seems that the 'LatentDiffusionSRTextWT' class can not load the SD 1.5 ckpt. Are there any easy methods to solve it? I checked the code. I think I removed...
BTW, SD v2 shows much better performance than SD v1. Since SD is regarded as a fixed prior in StableSR, SD v1 may lead to inferior performance compared with SD...