{"id":20812242,"url":"https://github.com/curegit/precure-stylegan","last_synced_at":"2025-05-11T20:30:29.731Z","repository":{"id":54701047,"uuid":"206583461","full_name":"curegit/precure-stylegan","owner":"curegit","description":"Yet another StyleGAN 1.0 implementation with Chainer to synthesize specific Precure (Cure Beauty) images","archived":true,"fork":false,"pushed_at":"2023-03-19T14:44:17.000Z","size":96019,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":2,"default_branch":"master","last_synced_at":"2025-05-10T04:37:29.273Z","etag":null,"topics":["anime","chainer","deep-learning","generative-adversarial-network","image-synthesis","precure","stylegan"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/curegit.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2019-09-05T14:26:22.000Z","updated_at":"2023-07-20T13:15:33.000Z","dependencies_parsed_at":"2024-11-17T21:01:38.179Z","dependency_job_id":null,"html_url":"https://github.com/curegit/precure-stylegan","commit_stats":null,"previous_names":[],"tags_count":7,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/curegit%2Fprecure-stylegan","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/curegit%2Fprecure-stylegan/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/curegit%2Fprecure-stylegan/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/curegit%2Fprecure-stylegan/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/curegit","download_url":"https://codeload.github.com/curegit/precure-stylegan/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253631809,"owners_count":21939357,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["anime","chainer","deep-learning","generative-adversarial-network","image-synthesis","precure","stylegan"],"created_at":"2024-11-17T20:51:28.969Z","updated_at":"2025-05-11T20:30:24.699Z","avatar_url":"https://github.com/curegit.png","language":"Python","readme":"# Precure StyleGAN\n\nYet another StyleGAN 1.0 implementation with Chainer\n\nWe tried out to generate facial images of a specific Precure (Japanese Anime) character.\n\nThis project is **finished** and [will be continued here for better quality with StyleGAN 2.0 ADA](https://github.com/curegit/precure-stylegan-ada).\n\n## Overview\n\n![Overall Result](examples/overall.png)\n\nStyleGAN is a generative adversarial network introduced by NVIDIA researchers.\nLike PGGAN, its output image resolutions grow progressively during training.\nThis implementation supports generation ranging from 4x4 px (stage 1) to 1024x1024 px (stage 9) images.\n\nMost of the implementation follows the original paper, but we have added some enhancements.\nFor example, we implemented an alternative least-squares objective introduced in LSGAN.\nWe trained the models with facial images of Cure Beauty (Smile Pretty Cure!, 2012) and other common datasets.\n\n## Requirements\n\n- Python \u003e= 3.6\n- Chainer \u003e= 7.0\n- Pillow \u003e= 9.1\n- Numpy \u003c 1.24\n- H5py\n- Matplotlib\n\n### Optional\n\n- Cupy\n- OpenCV-Python\n- Pydot (Graphviz)\n\n## Script Synopses\n\n- `train.py` trains StyleGAN models.\n  Use the `-h` option for more details.\n- `generate.py` generates images from a trained model.\n  Use the `-h` option for more details.\n- `mix.py` mixes styles from latent files.\n  Use the `-h` option for more details.\n- `animate.py` makes an animation of the analogy from a trained model.\n  Use the `-h` option for more details.\n- `visualize.py` draws an example of a computation graph for debugging (Pydot and Graphviz are required).\n  It takes no command-line arguments.\n- `check.py` analyzes the Chainer environment.\n  It takes no command-line arguments.\n\n## Results\n\n### Cure Beauty (Curated, FID = 80.43)\n\n![Cure Beauty](examples/beauty.png)\n\n### MNIST (Uncurated, FID = 1.14)\n\nTry yourself: `python3 generate.py -g models/mnist.hdf5 -x 4 -c 256 16 -z 256 -n 100 -d mnist-images`\n\n![MNIST](examples/mnist.png)\n\n### CIFAR-10 (Uncurated, ψ = 0.7, FID = 18.61)\n\nTry yourself: `python3 generate.py -g models/cifar10.hdf5 -x 4 -c 512 64 -t 0.7 -n 100 -d cifar10-images`\n\n![CIFAR-10](examples/cifar-10.png)\n\n### Anime Face (Uncurated, ψ = 0.8, FID = 13.11)\n\nTry yourself: `python3 generate.py -g models/anime.hdf5 -x 5 -c 512 64 -t 0.8 -n 100 -d anime-images`\n\n![Anime Face](examples/anime.png)\n\n## Bibliography\n\n### Papers\n\n- [Progressive Growing of GANs for Improved Quality, Stability, and Variation](https://arxiv.org/abs/1710.10196)\n- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://arxiv.org/abs/1812.04948)\n- [Least Squares Generative Adversarial Networks](https://arxiv.org/abs/1611.04076)\n\n### Implementation used as reference\n\n- [Chainer implementation of Style-based Generator](https://github.com/pfnet-research/chainer-stylegan)\n- [Chainer-StyleBasedGAN](https://github.com/RUTILEA/Chainer-StyleBasedGAN)\n\n### Datasets for testing\n\n- [THE MNIST DATABASE of handwritten digits](http://yann.lecun.com/exdb/mnist/)\n- [CIFAR-10 and CIFAR-100 datasets](https://www.cs.toronto.edu/~kriz/cifar.html)\n- [Anime-Face-Dataset](https://github.com/Mckinsey666/Anime-Face-Dataset)\n\n## License\n\n[CC BY-NC 4.0](LICENSE)\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcuregit%2Fprecure-stylegan","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fcuregit%2Fprecure-stylegan","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcuregit%2Fprecure-stylegan/lists"}