[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-82745":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":15,"subscribersCount":15,"size":15,"stars1d":16,"stars7d":17,"stars30d":18,"stars90d":15,"forks30d":15,"starsTrendScore":19,"compositeScore":20,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":23,"topics":24,"createdAt":10,"pushedAt":10,"updatedAt":28,"readmeContent":29,"aiSummary":30,"trendingCount":15,"starSnapshotCount":15,"syncStatus":31,"lastSyncTime":32,"discoverSource":33},82745,"learn-world-model","datawhalechina\u002Flearn-world-model","datawhalechina","Learn everything about world model","https:\u002F\u002Fdatawhalechina.github.io\u002Flearn-world-model\u002F",null,"TypeScript",224,16,1,0,6,63,115,25,82.69,false,"main",true,[25,26,27],"self-teaching","tutorial","world-model","2026-06-12 04:01:38","\u003Cdiv align=\"center\">\n  \u003Cimg src=\".\u002Fdocs\u002Fpublic\u002Fpreface.png\" width=\"100%\" alt=\"Learn World Models Banner\">\n  \u003Cbr>\n\n[English](.\u002FREADME.md) · [中文](.\u002FREADME-CN.md)\n\n# Learn World Models（⚠️ Alpha Preview）\n\n[![Read Online](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FRead-Online-blue?style=for-the-badge&logo=github)](https:\u002F\u002Fdatawhalechina.github.io\u002Flearn-world-model)\n[![GitHub Stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fdatawhalechina\u002Flearn-world-model?style=for-the-badge&logo=github)](https:\u002F\u002Fgithub.com\u002Fdatawhalechina\u002Flearn-world-model\u002Fstargazers)\n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-yellow?style=for-the-badge)](https:\u002F\u002Fgithub.com\u002Fdatawhalechina\u002Flearn-world-model\u002Fblob\u002Fmain\u002FLICENSE)\n\n> **Learn world models by building them: from the intuition behind latent dynamics to a working simulation, planning, and evaluation system.**\n\n\u003C\u002Fdiv>\n\n> [!CAUTION]\n> ⚠️ **Alpha Preview**: This is an early build. Content is still being completed and revised: sections, examples, and wording may continue to change. Feedback via Issues is welcome.\n\n---\n\n## ✨ Preview\n\n### 🏠 Course Home\n> Structured learning path with lecture and project cards.\n\n![Course home](.\u002Fdocs\u002Fpublic\u002Fscreenshots\u002Freadme\u002Fen-home.png)\n\n### 📖 Lecture Pages\n> Concept-first explanations with mermaid diagrams and background callouts for deep-learning readers.\n\n![Lecture page](.\u002Fdocs\u002Fpublic\u002Fscreenshots\u002Freadme\u002Fen-lecture-01.png)\n\n### 🗂️ Architecture Deep Dive\n> Seven architecture families, three planning mechanisms, side-by-side comparison tables.\n\n![Architecture lecture](.\u002Fdocs\u002Fpublic\u002Fscreenshots\u002Freadme\u002Fen-lecture-03.png)\n\n---\n\n## What this course covers\n\nFive lectures and five projects that take you from the intuition behind world models to a working three-model evaluation dashboard.\n\n| # | Type | Title | Core Topics |\n|---|------|-------|-------------|\n| L01 | Lecture | Internal Simulation & Historical Context | Craik's mental models, predictive coding, four eras of world model evolution |\n| L02 | Lecture | Observation Encoding & Latent Dynamics | VAE, CNN encoder, ELBO, GRU → MDN-RNN → RSSM |\n| L03 | Lecture | Architecture Patterns, Learning Paradigms & Planning | Seven architecture families, CEM-MPC, latent Actor-Critic, TD-MPC |\n| L04 | Lecture | Evaluation by World Model | FID, reward correlation, consistency loss, PSNR, horizon drift |\n| L05 | Lecture | Frontier Debates | Language vs physical grounding, Bitter Lesson, AGI as a research target |\n| P01 | Project | Train a VAE Encoder | Compress 64×64 pixels to latent z; reconstruction loss curve |\n| P02 | Project | Build a Latent Dynamics Model | GRU → RSSM; 1-step vs 5-step prediction error |\n| P03 | Project | Full Dreamer Pipeline | Encode → RSSM → latent Actor-Critic → act |\n| P04 | Project | Implement TD-MPC Planning | CEM-MPC + latent consistency loss; compare vs Dreamer |\n| P05 | Project | STORM + Three-Model Evaluation Dashboard | Swap GRU for Transformer; side-by-side Dreamer\u002FTD-MPC\u002FSTORM dashboard |\n\n---\n\n## Curriculum flow\n\n```mermaid\nflowchart TD\n    L01[\"L01 History and Intuition\"] --> L02A\n    L02A[\"L02 Part A: VAE Encoder\"] --> P01[\"P01 Build and visualize VAE\"]\n    L02A --> L02B[\"L02 Part B: GRU to RSSM\"]\n    L02B --> P02[\"P02 Train dynamics, measure drift\"]\n    L02B --> L03A[\"L03 Part A: Architecture Patterns\"]\n    L03A --> L03B[\"L03 Part B: Planning mechanisms\"]\n    L03B --> P03[\"P03 Full Dreamer pipeline\"]\n    P03 --> P04[\"P04 TD-MPC, compare vs Dreamer\"]\n    P04 --> L04[\"L04 Evaluation vocabulary\"]\n    L04 --> P05[\"P05 STORM + Three-Model Dashboard\"]\n    P05 --> L05[\"L05 Frontier Debates\"]\n```\n\nSuggested path: **L01 → L02 → P01 → P02 → L03 → P03 → P04 → L04 → P05 → L05**\n\nYou do not need to finish all theory before starting a project. Build, then come back with questions.\n\n---\n\n## Quick start\n\n```sh\nnpm install\nnpm run docs:dev        # dev server with hot reload\nnpm run docs:build      # production build\nnpm run docs:preview    # preview built site\n```\n\nTo refresh the README screenshots after a build:\n\n```sh\nnpm run docs:build\nnpm run screenshots:readme\n```\n\n---\n\n## Repo structure\n\n```\nlearn-world-model\u002F\n├── docs\u002F                                  # VitePress documentation site\n│   ├── .vitepress\u002Fconfig.mts             # nav and sidebar (EN + ZH)\n│   ├── en\u002Flectures\u002F                       # 5 English lecture pages\n│   ├── zh\u002Flectures\u002F                       # 5 Chinese lecture pages\n│   ├── en\u002Fprojects\u002F                       # 5 English project pages\n│   └── zh\u002Fprojects\u002F                       # 5 Chinese project pages\n├── external\u002Fworld-model-tutorial\u002F         # PyTorch source referenced by projects\n│   └── references.md                      # four-era history and architecture survey\n├── scripts\u002F                               # build utilities (screenshots, PDF)\n└── package.json\n```\n\n---\n\n## Community\n\nScan the QR code to join the WeChat discussion group (微信交流群):\n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\".\u002Fdocs\u002Fpublic\u002Fwechat-group-qr-code.jpg\" width=\"300\" alt=\"WeChat Group QR Code\">\n\u003C\u002Fdiv>\n\n---\n\n## Contributors (Tutorial)\n\n| Name | Role | Affiliation | GitHub |\n| ---- | ---- | ----------- | ------ |\n| Zhimin Zhao | Project Lead | Queen's University | [@zhimin-z](https:\u002F\u002Fgithub.com\u002Fzhimin-z) |\n| Qi Wang | Project Lead | Shanghai Jiao Tong University | [@qiwang067](https:\u002F\u002Fgithub.com\u002Fqiwang067) |\n","Learn World Models 是一个旨在通过构建世界模型来学习其原理和应用的项目。该项目使用 TypeScript 编写，提供了一条结构化的学习路径，包括五个讲座和五个实践项目，覆盖了从直觉理解到实现完整的模拟、规划与评估系统的全过程。内容涉及内部模拟与历史背景、观察编码与潜在动态、架构模式及规划机制等主题，并通过实际操作如训练 VAE 编码器、构建潜在动力学模型等加深理解。适合对机器学习特别是强化学习领域感兴趣的研究者或开发者自学使用。",2,"2026-06-11 04:09:05","CREATED_QUERY"]