12/27/2023 0 Comments Photo face cleaner onlineComponent locations of FFHQ: FFHQ_eye_mouth_landmarks_512.pth.Pre-trained StyleGAN2 model: StyleGAN2_512_Cmul1_FFHQ_B12G4_scratch_800k.pth.Put them in the experiments/pretrained_models folder. (You can try a simple version ( options/train_gfpgan_v1_simple.yml) that does not require face component landmarks.)ĭownload pre-trained models and other data. You may need to perform some pre-processing, such as beauty makeup.More high quality faces can improve the restoration quality.You could improve it according to your own needs. We provide the training codes for GFPGAN (used in our paper). You can find more models (such as the discriminators) here:, OR □ Training ![]() ✓better results on very low-quality inputs You may need to select different models based on your purpose and inputs. Note that V1.3 is not always better than V1.2. Trained with more data with pre-processing. No colorization no CUDA extensions are required. □ Model Zoo Versionīased on V1.2 more natural restoration results better results on very low-quality / high-quality inputs. If you want to use the original model in our paper, please see PaperModel.md for installation and inference. Options: auto | jpg | png, auto means using the same extension as inputs. Default: 400 -suffix Suffix of the restored faces -only_center_face Only restore the center face -aligned Input are aligned faces -ext Image extension. Default: realesrgan -bg_tile Tile size for background sampler, 0 for no tile during testing. Default: 2 -bg_upsampler background upsampler. Default: 1.3 -s upscale The final upsampling scale of the image. Default: results -v version GFPGAN model version. Default: inputs/whole_imgs -o output Output folder. h show this help -i input Input image or folder. Usage: python inference_gfpgan.py -i inputs/whole_imgs -o results -v 1.3 -s 2. If you want to use the original model in our paper, please see PaperModel.md for installation. We now provide a clean version of GFPGAN, which does not require customized CUDA extensions. Python >= 3.7 (Recommend to use Anaconda or Miniconda).Xintao Wang, Yu Li, Honglun Zhang, Ying ShanĪpplied Research Center (ARC), Tencent PCG □ GFP-GAN: Towards Real-World Blind Face Restoration with Generative Facial Prior ▶️ HandyView: A PyQt5-based image viewer that is handy for view and comparison ▶️ facexlib: A collection that provides useful face-relation functions ▶️ BasicSR: An open-source image and video restoration toolbox ▶️ Real-ESRGAN: A practical algorithm for general image restoration If GFPGAN is helpful in your photos/projects, please help to ⭐ this repo or recommend it to your friends. ✅ We provide an updated model without colorizing faces.✅ We provide a clean version of GFPGAN, which does not require CUDA extensions.✅ Support enhancing non-face regions (background) with Real-ESRGAN.✅ Integrated to Huggingface Spaces with Gradio. ![]() ✅ Add V1.3 model, which produces more natural restoration results, and better results on very low-quality / high-quality inputs. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |