{"id":50493420,"url":"https://github.com/ZeroLu/awesome-happy-horse","last_synced_at":"2026-06-18T22:00:59.118Z","repository":{"id":350155163,"uuid":"1205492751","full_name":"ZeroLu/awesome-happy-horse","owner":"ZeroLu","description":null,"archived":false,"fork":false,"pushed_at":"2026-04-09T06:16:59.000Z","size":73757,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2026-04-09T06:22:00.262Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":null,"has_issues":false,"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/ZeroLu.png","metadata":{"files":{"readme":"README-zh.md","changelog":null,"contributing":"CONTRIBUTING.md","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,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-04-09T02:31:13.000Z","updated_at":"2026-04-09T06:17:04.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/ZeroLu/awesome-happy-horse","commit_stats":null,"previous_names":["zerolu/awesome-happy-horse"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/ZeroLu/awesome-happy-horse","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZeroLu%2Fawesome-happy-horse","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZeroLu%2Fawesome-happy-horse/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZeroLu%2Fawesome-happy-horse/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZeroLu%2Fawesome-happy-horse/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ZeroLu","download_url":"https://codeload.github.com/ZeroLu/awesome-happy-horse/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZeroLu%2Fawesome-happy-horse/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34508867,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-18T02:00:06.871Z","response_time":128,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":[],"created_at":"2026-06-02T05:00:33.188Z","updated_at":"2026-06-18T22:00:59.112Z","avatar_url":"https://github.com/ZeroLu.png","language":null,"funding_links":[],"categories":["🤖 AI \u0026 Machine Learning"],"sub_categories":[],"readme":"# Awesome Happy Horse 1.0\n\n[![Awesome](https://awesome.re/badge.svg)](https://github.com/sindresorhus/awesome)\n[![License: CC BY 4.0](https://img.shields.io/badge/License-CC_BY_4.0-lightgrey.svg)](https://creativecommons.org/licenses/by/4.0/)\n\n\u003e 精选的 Happy Horse 1.0 提示词、视频生成技术和资源合集，助你充分发挥 AI 视频模型的潜力。\n\n\u003cdiv align=\"center\"\u003e\n\n| [English](README.md) | [简体中文](README-zh.md) |\n|----------------------|--------------------------|\n\n\u003c/div\u003e\n\n---\n\n**最后更新：** 2026 年 4 月\n\nHappy Horse 1.0 是一个**150 亿参数统一 Transformer 模型**，可从文本或图像提示生成视频和同步音频，支持电影级 1080p 画质和 7 种语言的唇形同步。完全开源，支持商业用途。\n\n**官方网站：** [happyhorseone.com](https://happyhorseone.com)\n\n## 🐴 为什么选择 Happy Horse 1.0？\n\n### 🏆 当前排行榜排名\n\n**Artificial Analysis T2V 排行榜（无音频）- 2026 年 4 月初**\n\n| 排名 | 模型 | Elo | API 可用 | 发布时间 |\n|------|------|-----|----------|----------|\n| **#1** | **HappyHorse-1.0** | **1333** | ❌ 否 | 2026 年 4 月 |\n| #2 | Seedance 2.0 720p | 1273 | ❌ 无公共 API | 2026 年 3 月 |\n| #3 | SkyReels V4 | 1245 | ✅ 是 ($7.20/min) | 2026 年 3 月 |\n| #4 | Kling 3.0 1080p Pro | 1241 | ✅ 是 ($13.44/min) | 2026 年 2 月 |\n| #5 | PixVerse V6 | 1240 | ✅ 是 ($5.40/min) | 2026 年 3 月 |\n\n![Artificial Analysis T2V 排行榜 - 2026 年 4 月](https://d8r3hf7cxo9nn.cloudfront.net/cms/2026-04-08/20260408071633_pi0gts3g.png)\n\n*Artificial Analysis T2V 排行榜 - 2026 年 4 月初*\n\n### 技术对比\n\n| 特性 | Happy Horse 1.0 | Seedance 2.0 | LTX-2.3 |\n|------|-----------------|--------------|---------|\n| 模型规模 | 15B (统一架构) | ~4.5B | 22B |\n| 架构类型 | 单流 Transformer | 扩散 Transformer | 双流架构 |\n| 文生视频 Elo | **1355** | 1273 | 1290 |\n| 图生视频 Elo | **1404** | 1357 | 1345 |\n| 去噪步数 | **8 步** (DMD-2) | 25-50 步 | 12-20 步 |\n| 1080p 渲染时间 | **~38 秒** (H100) | ~55 秒 | ~45 秒 |\n| 唇形同步语言 | **7 种语言** (原生) | 需外部工具 | 有限支持 |\n| 开源 | ✅ **是** | ❌ 否 | ✅ 是 |\n\n**胜率：** 80.0% vs OVI 1.1 · 60.9% vs LTX 2.3\n\n---\n\n## 📑 目录\n\n- [电影风格](#1-电影风格)\n- [广告与品牌](#2-广告与品牌)\n- [社交媒体与病毒内容](#3-社交媒体与病毒内容)\n- [UGC 风格](#4-ugc 风格)\n- [动漫与动画](#5-动漫与动画)\n- [短剧与网剧](#6-短剧与网剧)\n- [视觉特效与实验风格](#7-视觉特效与实验风格)\n- [资源链接](#8-资源链接)\n- [贡献指南](#9-贡献指南)\n\n---\n\n## 🎬 核心能力\n\nHappy Horse 1.0 的统一多模态架构专为联合视频和音频生成而设计：\n\n- **🎯 统一 Transformer：** 40 层自注意力网络，含模态特定层\n- **🎵 联合视频 + 音频：** 单次通过生成同步对话、环境音和音效\n- **⚡ 8 步 DMD-2 蒸馏：** 无需分类器自由引导的快速生成\n- **🌍 多语言唇形同步：** 英语、普通话、粤语、日语、韩语、德语、法语\n- **📺 1080p 输出：** 5-8 秒电影级画质片段\n- **🔓 开源可自托管：** 完整权重，支持商业用途\n\n---\n\n## 📑 目录\n\n- [资源链接](#资源链接)\n- [安装与部署](#安装与部署)\n- [提示词工程指南](#提示词工程指南)\n- [贡献指南](#贡献指南)\n\n---\n\n## 资源链接\n\n### 官方资源\n\n- [🏠 Happy Horse 官网](https://happyhorseone.com) - 官方网站，含演示和文档\n- [📦 GitHub 仓库](https://github.com/happy-horse/happyhorse-1) - 源代码和模型权重\n- [📄 技术报告](https://happyhorseone.com/report) - 完整架构和基准测试详情\n- [🎮 在线演示](https://happyhorseone.com/demo) - 免费试用 Happy Horse 1.0\n\n### 社区与资讯\n\n- [🔥 Brent Lynch - 来源揭秘](https://x.com/BrentLynch/status/2041903421851365838) - Happy Horse 来自阿里通义实验室（前 Kling 团队负责人）\n- [📊 GMI Cloud - 对比测试](https://x.com/gmi_cloud/status/2041952066873221288) - Happy Horse vs Veo, Kling, SkyReel, PixVerse\n- [⚠️ 虚假内容警告](https://x.com/BrentLynch/status/2041997193373237561) - 警惕 Seedance 2.0 冒充 Happy Horse\n- [FaceSwap-AI 宣布](https://x.com/chennnhhe/status/2041914942509740035) - \"开源 AI 视频的 Sora 时刻\"\n- [HappyhorseAI 官方](https://x.com/Happyhorseteam/status/2041909880496517550) - 官方发布公告\n- [Textideo 深度分析](https://textideo.com/blog/happy-horse-1-0-redefining-open-source-sota-ai-video-generation) - 技术深度解析\n- [AI Research 周报](https://x.com/airesearch_ai/status/2041638371538432256) - 每周 AI 新闻精选\n\n### 对比与基准测试\n\n- [Artificial Analysis Video Arena](https://artificialanalysis.ai) - Happy Horse 排名第一的排行榜\n- [模型对比表](https://happyhorses.io#compare) - 与 Seedance 2.0、LTX-2.3、OVI 1.1 的对比\n\n---\n\n## 安装与部署\n\n### 系统要求\n\n- **GPU：** NVIDIA H100 或 A100（推荐 ≥48GB 显存）\n- **存储：** 约 30GB 用于模型权重\n- **内存：** 推荐 64GB RAM\n\n### 快速开始\n\n```bash\n# 克隆仓库\ngit clone https://github.com/happy-horse/happyhorse-1.git\ncd happyhorse-1\n\n# 安装依赖\npip install -r requirements.txt\n\n# 下载模型权重\nbash download_weights.sh\n\n# 生成你的第一个视频\npython demo_generate.py --prompt \"a robot dancing on the moon\" --duration 5\n```\n\n### Python API\n\n```python\nfrom happyhorse import HappyHorseModel\n\nmodel = HappyHorseModel.from_pretrained(\"happy-horse/happyhorse-1.0\")\n\nvideo, audio = model.generate(\n    prompt=\"一位老人站在山峰上俯瞰山谷\",\n    duration_seconds=5,\n    fps=24,\n    language=\"zh\",\n)\n\nvideo.save(\"output.mp4\")\naudio.save(\"output.wav\")\n```\n\n### 性能基准\n\n| 分辨率 | 时间 (H100) | 适用场景 |\n|--------|-------------|----------|\n| 256p 预览 | ~2.0 秒 | 快速迭代 |\n| 540p + 超分 | ~8.0 秒 | 社交媒体 |\n| 1080p 高清 | ~38.4 秒 | 生产质量 |\n\n---\n\n## 提示词工程指南\n\n### 最佳实践\n\n1. **具体描述：** 包含光线、摄像机角度和动作的详细信息\n2. **音频提示：** 提及对话、环境音或音乐风格\n3. **语言选择：** 为对话场景指定唇形同步语言\n4. **时长控制：** 提示词与 5-8 秒输出长度对齐\n\n### 示例提示词结构\n\n```\n[风格/类型]：电影级，戏剧性灯光\n[主体]：主要角色描述和动作\n[场景]：环境和背景细节\n[摄像机]：镜头类型、移动、角度\n[音频]：对话语言、环境音、音乐风格\n[时长]：5-8 秒\n```\n\n### 社区示例提示词\n\n来自 @BrentLynch 的 X 收藏夹：\n\n```\n手电筒光束探索洞穴系统，照亮湿润的石灰岩地层。\n光线捕捉到结晶的方解石沉积物，闪烁着光芒。\n光束穿过浅水区时，在水底形成明亮的焦散图案。\n钟乳石随着手电筒移动投下长长的、摆动的阴影。\n\n音频：滴水声在洞穴中回响，脚步声踩在湿岩石上，封闭空间中的呼吸声。\n```\n\n**为什么这个提示词有效：**\n- ✅ 详细的环境描述（洞穴系统、湿石灰岩）\n- ✅ 动态光影效果（手电筒光束、晶体闪烁、阴影）\n- ✅ 水面效果（浅水、焦散图案）\n- ✅ 运动元素（光束移动、阴影摆动）\n- ✅ 音频提示（滴水声、脚步声、呼吸声）\n- ✅ 空间感（封闭空间声学）\n\n### 多语言支持\n\nHappy Horse 1.0 原生支持 7 种语言的唇形同步：\n- 🇺🇸 英语\n- 🇨🇳 普通话（含方言）\n- 🇭🇰 粤语\n- 🇯🇵 日语\n- 🇰🇷 韩语\n- 🇩🇪 德语\n- 🇫🇷 法语\n\n---\n\n## 贡献指南\n\n欢迎贡献！如果你有优秀的 Happy Horse 1.0 提示词或资源，请提交 Pull Request。\n\n### 如何贡献\n\n1. Fork 本仓库\n2. 创建新分支 (`git checkout -b feature/add-new-prompt`)\n3. 在相应部分添加你的提示词/资源\n4. 提交 PR\n\n**请确保包含以下内容：**\n- 完整提示词文本（如适用）\n- 来源链接（原创者的 X/Twitter 帖子）\n- 风格/效果简述\n- 示例视频或图片（如有）\n\n### 贡献模板\n\n```markdown\n### [你的标题]\n\n*简述。*\n\n**提示词：**\n```\n[你的提示词]\n```\n\n**来源：** [创作者名称](https://x.com/@username) - [帖子链接](https://x.com/...)\n\n**设置：**\n- 时长：5 秒\n- 语言：中文\n- 分辨率：1080p\n```\n\n---\n\n## 许可证\n\n[![CC BY 4.0](https://img.shields.io/badge/License-CC_BY_4.0-lightgrey.svg)](https://creativecommons.org/licenses/by/4.0/)\n\n本作品采用 [知识共享署名 4.0 国际许可协议](https://creativecommons.org/licenses/by/4.0/) 进行许可。\n\nHappy Horse 1.0 模型权重在允许商业用途的许可下发布。\n\n---\n\n\u003cdiv align=\"center\"\u003e\n\n**觉得有帮助？** ⭐ Star 支持社区！\n\n🐴 由 Happy Horse 社区 ❤️ 构建\n\n\u003c/div\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FZeroLu%2Fawesome-happy-horse","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FZeroLu%2Fawesome-happy-horse","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FZeroLu%2Fawesome-happy-horse/lists"}