{"name":"LAM: Large Avatar Model for One-shot Animatable Gaussian Head","description":"LAM is a cutting-edge project that enables the creation of ultra-realistic, animatable 3D avatars from just a single image in seconds. Leveraging advanced Gaussian Head technology, it offers super-fast cross-platform rendering and a low-latency SDK for real-time interactive chatting. This innovative model is set to be presented at SIGGRAPH 2025.","github":"https://github.com/aigc3d/LAM","url":"https://osrepos.com/repo/aigc3d-lam","source":"osrepos.com","sourceDescription":"This repository profile is provided by osrepos.com, an open source repository discovery platform.","repositoryProfile":"https://osrepos.com/repo/aigc3d-lam","generatedFor":"open source discovery and AI-assisted research","markdown":"https://osrepos.com/repo/aigc3d-lam.md","json":"https://osrepos.com/repo/aigc3d-lam.json","topics":["Python","3D Avatar","Gaussian Splatting","AI","Computer Graphics","Real-time","Animatable Head","SIGGRAPH"],"keywords":["Python","3D Avatar","Gaussian Splatting","AI","Computer Graphics","Real-time","Animatable Head","SIGGRAPH"],"stars":null,"summary":"LAM is a cutting-edge project that enables the creation of ultra-realistic, animatable 3D avatars from just a single image in seconds. Leveraging advanced Gaussian Head technology, it offers super-fast cross-platform rendering and a low-latency SDK for real-time interactive chatting. This innovative model is set to be presented at SIGGRAPH 2025.","content":"## Introduction\n\nLAM, or Large Avatar Model, is an innovative project designed to revolutionize the creation of 3D digital humans. It allows users to generate ultra-realistic, animatable 3D avatars from a single image in mere seconds. This technology is particularly notable for its super-fast cross-platform animating and rendering capabilities, making it suitable for various devices. Furthermore, LAM provides a low-latency SDK, enabling the development of real-time interactive chatting avatars. The project is slated for presentation at SIGGRAPH 2025, highlighting its significance in the field of computer graphics and AI.\n\n## Installation\n\nGetting started with LAM is straightforward, with options for both Windows and Linux users.\n\n### Online Demos\n\nFor a quick preview, you can explore the online demos:\n\n*   **Avatar Generation from One Image:**\n    *   [HuggingFace Space](https://huggingface.co/spaces/3DAIGC/LAM)\n    *   [ModelScope Space](https://www.modelscope.cn/studios/Damo_XR_Lab/LAM_Large_Avatar_Model)\n*   **Interactive Chatting:**\n    *   [HuggingFace Space](https://huggingface.co/spaces/HumanAIGC-Engineering-Team/open-avatar-chat)\n    *   [ModelScope Space](https://www.modelscope.cn/studios/HumanAIGC-Engineering/open-avatar-chat)\n\n### Environment Setup\n\n**Windows One-Click Installer:**\n\nA convenient one-click installation package for Windows (Cuda 12.8) is available. You can find a [video guide](https://www.bilibili.com/video/BV13QGizqEey) and the [download link](https://virutalbuy-public.oss-cn-hangzhou.aliyuncs.com/share/aigc3d/data/LAM/Installation/LAM-windows-one-click-install.zip).\n\n**Linux:**\n\nbash\ngit clone https://github.com/aigc3d/LAM.git\ncd LAM\n# Install with Cuda 12.1\nsh ./scripts/install/install_cu121.sh\n# Or Install with Cuda 11.8\nsh ./scripts/install/install_cu118.sh\n\n\n**Windows (Manual):**\n\nFor manual installation on Windows, please refer to the [Windows Install Guide](https://github.com/aigc3d/LAM/blob/master/scripts/install/WINDOWS_INSTALL.md).\n\n### Model Weights\n\nModel weights can be downloaded from HuggingFace or ModelScope.\n\n**HuggingFace Download:**\n\nbash\n# Download Assets\nhuggingface-cli download 3DAIGC/LAM-assets --local-dir ./tmp\ntar -xf ./tmp/LAM_assets.tar && rm ./tmp/LAM_assets.tar\ntar -xf ./tmp/thirdparty_models.tar && rm -r ./tmp/\n# Download Model Weights\nhuggingface-cli download 3DAIGC/LAM-20K --local-dir ./model_zoo/lam_models/releases/lam/lam-20k/step_045500/\n\n\n**ModelScope Download:**\n\nbash\npip3 install modelscope\n# Download Assets\nmodelscope download --model \"Damo_XR_Lab/LAM-assets\" --local_dir \"./tmp/\"\ntar -xf ./tmp/LAM_assets.tar && rm ./tmp/LAM_assets.tar\ntar -xf ./tmp/thirdparty_models.tar && rm -r ./tmp/\n# Download Model Weights\nmodelscope download \"Damo_XR_Lab/LAM-20K\" --local_dir \"./model_zoo/lam_models/releases/lam/lam-20k/step_045500/\"\n\n\n## Examples\n\nAfter setting up the environment, you can run LAM locally.\n\n### Gradio Run\n\nTo run the Gradio demo:\n\nbash\npython app_lam.py\n\n\nIf you wish to export ZIP files for real-time conversations on OpenAvatarChat, refer to the [Avatar Export Guide](https://github.com/aigc3d/LAM/blob/master/tools/AVATAR_EXPORT_GUIDE.md) and run with the Blender path:\n\nbash\npython app_lam.py --blender_path /path/blender\n\n\n### Inference\n\nFor inference, use the provided script:\n\nbash\nsh ./scripts/inference.sh ${CONFIG} ${MODEL_NAME} ${IMAGE_PATH_OR_FOLDER} ${MOTION_SEQ}\n\n\n## Why Use LAM?\n\nLAM offers compelling advantages for anyone interested in 3D avatar creation and interactive AI applications:\n\n*   **Ultra-realistic 3D Avatar Creation from One Image in Seconds:** Quickly generate high-fidelity avatars with minimal input.\n*   **Super-fast Cross-platform Animating and Rendering on Any Devices:** Experience smooth performance across various hardware, from powerful GPUs to mobile phones.\n*   **Low-latency SDK for Realtime Interactive Chatting Avatar:** Build responsive and engaging conversational AI experiences with integrated 3D digital humans.\n\n## Links\n\n*   **GitHub Repository:** [https://github.com/aigc3d/LAM](https://github.com/aigc3d/LAM)\n*   **Project Website:** [https://aigc3d.github.io/projects/LAM/](https://aigc3d.github.io/projects/LAM/)\n*   **arXiv Paper:** [https://arxiv.org/pdf/2502.17796](https://arxiv.org/pdf/2502.17796)\n*   **HuggingFace Space (Avatar Generation):** [https://huggingface.co/spaces/3DAIGC/LAM](https://huggingface.co/spaces/3DAIGC/LAM)\n*   **ModelScope Space (Avatar Generation):** [https://www.modelscope.cn/studios/Damo_XR_Lab/LAM_Large_Avatar_Model](https://www.modelscope.cn/studios/Damo_XR_Lab/LAM_Large_Avatar_Model)\n*   **HuggingFace Space (Interactive Chatting):** [https://huggingface.co/spaces/HumanAIGC-Engineering-Team/open-avatar-chat](https://huggingface.co/spaces/HumanAIGC-Engineering-Team/open-avatar-chat)\n*   **ModelScope Space (Interactive Chatting):** [https://www.modelscope.cn/studios/HumanAIGC-Engineering/open-avatar-chat](https://www.modelscope.cn/studios/HumanAIGC-Engineering/open-avatar-chat)\n*   **OpenAvatarChat:** [https://github.com/HumanAIGC-Engineering/OpenAvatarChat](https://github.com/HumanAIGC-Engineering/OpenAvatarChat)","metrics":{"detailViews":18,"githubClicks":12},"dates":{"published":null,"modified":"2025-12-07T08:00:59.000Z"}}