FaceChain - A framework for generating portraits and personal images launched by Alibaba.
FaceChain is an open-source AI framework for generating portraits and digital avatars (similar to a free and open-source version of Miaoduck Camera) launched by Alibaba DAMO Academy. Users only need to provide a minimum of one photo to generate a unique...
FaceChain is an open-source AI framework for generating portraits and digital avatars, developed by Alibaba DAMO Academy (similar to a free and open-source version of Miaoduck Camera). Users only need to provide a minimum of one photo to generate a unique digital avatar. This AI framework utilizes the text-to-image function of the Stable Diffusion model, combined with portrait stylization LoRA model training and face-related perception and understanding models, to train the input image and then infer and output a personalized portrait image.
FaceChain Features
- Image customization trainingUsers only need to provide at least one head and shoulder photo to use for LoRA stylization training and generate a digital avatar with a personalized style.
- Generate personal photos in various stylesFaceChain can generate personal photos in various styles, including Hanfu style, work photos, Barbie doll style, school uniform style, Christmas style, gentleman style, comic style, etc., to meet the diverse personalized needs of users.
- Supports SD WebUI plugin callsFaceChain supports invocation via the SD WebUI plugin, allowing interaction with AI models through the SD interface, and convenient generation and editing of personal avatars.
- Support attitude controlUsers can control the pose of the generated digital avatar, making it possible to create dynamic or action-oriented personal avatars.
- Custom prompt wordsUsers can change the clothing, accessories, etc. of their digital avatar by entering specific prompts, achieving more personalized customization.
FaceChain's official website entrance
- FaceChain's GitHub repository:https://github.com/modelscope/facechain
- FaceChain's research paper:https://arxiv.org/abs/2308.14256
- FaceChain's ModelScope link:https://www.modelscope.cn/brand/view/FaceChain
- FaceChain's Hugging Face address:https://huggingface.co/spaces/modelscope/FaceChain
How to use FaceChain
FaceChain supports model training and inference capabilities within the Grado interface, allows experienced developers to use Python scripts for training and inference, and also supports installation of plugins within the SD WebUI. This article demonstrates the experience and use of FaceChain through the online Grado version running on the ModelScope community.
- accessFaceChain's ModelScope demo pageUnder the "Character Image Training" tab, select 1-10 personal face/headshot photos (avoiding multiple faces or faces obscured in the images) and upload them.
- After uploading, click "Start Training" to begin customized image training. Each image will take approximately 1.5 minutes to complete.
- Once training is complete, switch to the "Portrait" tab on the right, adjust and set the relevant parameters to generate your style photo.
- Alternatively, you can use it directly.FaceChain Agent VersionNo complicated operations are required; you can obtain various portrait photos simply through dialogue. This method is highly recommended.