[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-71009":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":16,"stars7d":16,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":16,"compositeScore":18,"rankGlobal":10,"rankLanguage":10,"license":19,"archived":20,"fork":21,"defaultBranch":22,"hasWiki":20,"hasPages":21,"topics":23,"createdAt":10,"pushedAt":10,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":16,"starSnapshotCount":16,"syncStatus":17,"lastSyncTime":27,"discoverSource":28},71009,"monolith","bytedance\u002Fmonolith","bytedance","A Lightweight Recommendation System","",null,"Python",9307,717,1,15,0,2,64.77,"Other",true,false,"master",[],"2026-06-12 04:00:58"," Monolith\n\n## What is it?\n\n[Monolith](https:\u002F\u002Farxiv.org\u002Fabs\u002F2209.07663) is a deep learning framework for large scale recommendation modeling. It introduces two important features which are crucial for advanced recommendation system: \n* collisionless embedding tables guarantees unique represeantion for different id features\n* real time training captures the latest hotspots and help users to discover new intersts rapidly\n\nMonolith is built on the top of TensorFlow and supports batch\u002Freal-time training and serving.\n\n\n## Discussion Group\n\n### Join us at Discord\n\nhttps:\u002F\u002Fdiscord.gg\u002FQYTDeKxGMX\n\n## Quick start\n\n### Build from source\n\nCurrently, we only support compilation on the Linux.\n\nFirst, download bazel 3.1.0\n```bash\nwget https:\u002F\u002Fgithub.com\u002Fbazelbuild\u002Fbazel\u002Freleases\u002Fdownload\u002F3.1.0\u002Fbazel-3.1.0-installer-linux-x86_64.sh && \\\n  chmod +x bazel-3.1.0-installer-linux-x86_64.sh && \\\n  .\u002Fbazel-3.1.0-installer-linux-x86_64.sh && \\\n  rm bazel-3.1.0-installer-linux-x86_64.sh\n```\n\nThen, prepare a python environment\n```bash\npip install -U --user pip numpy wheel packaging requests opt_einsum\npip install -U --user keras_preprocessing --no-deps\n```\n\nFinally, you can build any target in the monolith.\nFor example,\n```bash\nbazel run \u002F\u002Fmonolith\u002Fnative_training:demo --output_filter=IGNORE_LOGS\n```\n\n### Demo and tutorials\n\nThere are a tutorial in [markdown\u002Fdemo](markdown\u002Fdemo) on how to run distributed async training, and few guides on how to use the `MonolithModel` API [here](markdown).  ","Monolith 是一个轻量级的推荐系统框架，专为大规模推荐模型设计。它基于 TensorFlow 构建，支持批量和实时训练及服务，并引入了无碰撞嵌入表和实时训练两大关键特性，前者确保不同 ID 特征的独特表示，后者能够捕捉最新热点并快速帮助用户发现新兴趣。该项目适用于需要处理海量数据且对个性化推荐有高要求的应用场景，如电商、社交媒体等。其开源性质使得开发者可以轻松地根据自身需求进行定制与扩展。","2026-06-11 03:35:27","high_star"]