Shangchen Zhou

Fixed it! Thanks.

We have no plan to release this colorization model since it only works for the cropped face and could be hard to harmonize with the whole image color for practical...

It will be a new project

The input created image has a large resolution of 1536x1024 which is beyond our model with an output face of 512. Since the face restoration models are originally designed for...

Hi all @andyderuyter @MarcusAdams @caacoe, I add the face upsampling '--face_upsample' option for high-resolution AI-created faces. Please have a try! e.g., `python inference_codeformer.py --w 0.7 --test_path inputs/user_upload --bg_upsampler realesrgan --face_upsample`

The result of using `--face_upsample` 

> @sczhou I'm not seeing a difference. I tried --face_upsample with weight 1.0 and weight 7.0, with both --upscale 1 and --upscale 2, but I can't discern a difference between...

Yes, FID is considered a non-reference metric for image quality assessment since **NO paired ground truth (the corresponding HQ version) is needed for this metric**. The reference dataset just stands...

Thanks @chaofengc for your exhaustive explanation 👍

@estelen `FACE_UPSAMPLE` has been already added. You may need to check the new version.