Zhen Li
Zhen Li
How to replicate this error?
Download corresponding `safetensors` file to `./civitai_models`
Hi all, you could follow this [link](https://github.com/TencentARC/PhotoMaker/discussions/114) for saving GPU memory.
Here are links of models we used in PhotoMaker-Style: 1. [stableUnstableDiffusers_v11](https://civitai.com/models/84040?modelVersionId=276923) 2. [xl_more_art-full](https://civitai.com/models/124347?modelVersionId=152309)
We're not familiar with Windows setup, maybe you could leave an issue to https://github.com/bmaltais/PhotoMaker/tree/v1.0.1
Hello, you could refer to this issue #7.
For scale factor x4, the results of Set5 and Set14 are 32.39/0.897(PSNR/SSIM) and 28.77/0.786, respectively.
You can change the value of `split_batch` in json file for training on one GPU.
I'm sorry that our code can't support CPU currently. We will solve this problem in the next code update.
Instead of applying an early stopping strategy, choosing the biggest evaluated PSNR value during 1000 training epochs is our preferred way. The reason is that the PSNR value still has...