CodeFormer
CodeFormer copied to clipboard
sczhou
[NeurIPS 2022] Towards Robust Blind Face Restoration with Codebook Lookup Transformer
### KYC Using Your GitHub Account Please make sure your GitHub account meets the following requirements (all are required): x The issue title must be: Request GitHub KYc - Your...
Hi, I tried to use the repo but it turned out that there were some circular dependencies (with python=3.9). Therefore, I commented out a few import and found that it...
I have come across this image. This is the picture of a molester. I need to unmask the person. I tried restoring with all sorts of fidelity values and inpainting....
我的显卡是rtx4090，运行命令后，他会持续5分钟左右才开始处理，然后处理完成的效果和没处理的一样，不清楚为什么不起作用
In the code, β times `zq-z.detach()`, https://github.com/sczhou/CodeFormer/blob/e878192ee253cfcc8f19e29d3307c181501f53ae/basicsr/archs/vqgan_arch.py#L55 but i think it should times the commitment loss `zq.detach()-z` to control the learning of Encoder, referred to the CoderFormer and VQVAE papers....
 (codeformer) H:\MoveDisk\WorkSpace\CodeFormer>python inference_codeformer.py --bg_upsampler realesrgan --face_upsample -w 1.0 --input_path H:\MoveDisk\WorkSpace\CodeFormer\inputs\test\test.mp4 Traceback (most recent call last): File "inference_codeformer.py", line 97, in image = vidreader.get_frame() File "H:\MoveDisk\WorkSpace\CodeFormer\basicsr\utils\video_util.py", line 81, in get_frame...
What is this error? I used it a month ago without any problems. 
When I trained the model and saved the checkpoint, I noticed that only the checkpoints for net_g and net_d were saved, but not for the Codebook (latent_gt_code1024). Is this normal?...
