[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-72584":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":14,"stars30d":18,"stars90d":16,"forks30d":16,"starsTrendScore":15,"compositeScore":19,"rankGlobal":10,"rankLanguage":10,"license":20,"archived":21,"fork":22,"defaultBranch":23,"hasWiki":22,"hasPages":22,"topics":24,"createdAt":10,"pushedAt":10,"updatedAt":25,"readmeContent":26,"aiSummary":27,"trendingCount":16,"starSnapshotCount":16,"syncStatus":17,"lastSyncTime":28,"discoverSource":29},72584,"chameleon","facebookresearch\u002Fchameleon","facebookresearch","Repository for Meta Chameleon, a mixed-modal early-fusion foundation model from FAIR.","https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.09818",null,"Python",2099,117,3,6,0,2,4,28.22,"Other",true,false,"main",[],"2026-06-12 02:03:05","# Meta Chameleon\n\n**[Paper](\u002F\u002Farxiv.org\u002Fabs\u002F2405.09818) | [Blog](\u002F\u002Fai.meta.com\u002Fblog\u002Fmeta-fair-research-new-releases\u002F) | [Model Checkpoint Download](\u002F\u002Fai.meta.com\u002Fresources\u002Fmodels-and-libraries\u002Fchameleon-downloads) | [HuggingFace](\u002F\u002Fhuggingface.co\u002Fcollections\u002Ffacebook\u002Fchameleon-668da9663f80d483b4c61f58)**\n\nThis repository contains artifacts for the Meta Chameleon model from FAIR, Meta AI. In this repository is:\n- [Standalone Inference Code](.\u002Fchameleon\u002Finference) — a fast GPU-based inference implementation for running [model checkpoints](\u002F\u002Fai.meta.com\u002Fresources\u002Fmodels-and-libraries\u002Fchameleon-downloads)\n- [Input-Output Viewing](.\u002Fchameleon\u002Fviewer) — a harness for richly viewing multimodal model inputs and outputs with a browser-based tool\n- [Evaluation Prompts](.\u002Fdata) — mixed-modal and text-only prompts for human evaluation\n\n### System Requirements\n\nRunning constituent components for inference and the input-output viewer currently requires a CUDA-capable GPU. If you'd like to run inference on other hardware, other inference implementations, including [HuggingFace](\u002F\u002Fhuggingface.co\u002Ffacebook\u002Fchameleon), are platform agnostic.\n\n## Getting Started\n\nFirst, pip install this repository:\n```sh\npip install -U git+https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fchameleon.git\n```\n\nAlternatively, if you want access to the full visualizer, you'll need to clone this repository (instead of installing), then pip install from the repository root:\n```sh\ngit clone https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fchameleon.git\ncd chameleon\npip install -e .\n```\n\nModel checkpoints and configs must be downloaded before running inference or the viewer. After [requesting model access](\u002F\u002Fai.meta.com\u002Fresources\u002Fmodels-and-libraries\u002Fchameleon-downloads\u002F), run the following script, adding pre-signed download URL you were emailed when prompted:\n```sh\npython -m chameleon.download_data [pre-signed URL]\n```\n(you can also paste the command given in the email containing the download link)\n\n### Running the Viewer\n\nThe [viewer](.\u002Fchameleon\u002Fviewer) visualizes multi-modal model input and output. It is most easily run with [`docker-compose`](\u002F\u002Fdocs.docker.com\u002Fcompose\u002Finstall\u002F). You'll need to clone the repository, not just a pip install.\n\nThe following runs both the service and viewer interface. \n> **By default, this runs the 7B parameter model. You can change the `model_path` variable in [`.\u002Fconfig\u002Fmodel_viewer.yaml`](.\u002Fconfig\u002Fmodel_viewer.yaml)** to select another model and alter other configuration:\n```sh\ndocker-compose up --build\n```\n\nYou can open the viewer at http:\u002F\u002Flocalhost:7654\u002F\n\n### Running the MiniViewer\n\nThe [miniviewer](.\u002Fchameleon\u002Fminiviewer) is a light weight debug visualizer, that can be run with:\n```sh\npython -m chameleon.miniviewer\n```\nThis runs the 7B parameter model. To run the 30B model, use the following command:\n```sh\npython -m chameleon.miniviewer --model-size 30b\n```\n\nYou can open the miniviewer at `http:\u002F\u002Flocalhost:5000\u002F`.\n\n### License\n\nUse of this repository and related resources are governed by the [Chameleon Research License](\u002F\u002Fai.meta.com\u002Fresources\u002Fmodels-and-libraries\u002Fchameleon-license) and the [LICENSE](.\u002FLICENSE) file.\n\n#### Citation\n\nTo cite the paper, model, or software, please use the below:\n```\n@article{Chameleon_Team_Chameleon_Mixed-Modal_Early-Fusion_2024,\n  author = {Chameleon Team},\n  doi = {10.48550\u002FarXiv.2405.09818},\n  journal = {arXiv preprint arXiv:2405.09818},\n  title = {Chameleon: Mixed-Modal Early-Fusion Foundation Models},\n  url = {https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fchameleon},\n  year = {2024}\n}\n```\n","Meta Chameleon 是由FAIR（Facebook AI Research）开发的一个混合模态早期融合基础模型。该项目的核心功能包括基于GPU的快速推理代码、浏览器端的输入输出可视化工具以及用于人类评估的混合模态和纯文本提示数据集。它特别适用于需要处理多模态数据（如图像与文本结合）的应用场景，例如多媒体内容分析、跨模态检索等。技术上，Chameleon 要求CUDA兼容的GPU来运行其组件，但通过HuggingFace平台也支持其他硬件环境下的部署。此外，项目提供了详细的文档指导用户如何安装依赖、下载模型权重文件以及启动可视化界面，便于研究人员和技术爱好者快速上手使用。","2026-06-11 03:42:42","high_star"]