Skip to content

Commit d1bd389

Browse files
committed
first commit
1 parent 98229d8 commit d1bd389

537 files changed

Lines changed: 119185 additions & 0 deletions

File tree

  • assets
  • desktop
  • phases_cn
    • 00-setup-and-tooling
    • 01-math-foundations
      • 01-linear-algebra-intuition
      • 02-vectors-matrices-operations
      • 03-matrix-transformations
      • 04-calculus-for-ml
      • 05-chain-rule-and-autodiff
      • 06-probability-and-distributions
      • 07-bayes-theorem
      • 08-optimization
      • 09-information-theory
      • 10-dimensionality-reduction
      • 11-singular-value-decomposition
      • 12-tensor-operations
      • 13-numerical-stability
      • 14-norms-and-distances
      • 15-statistics-for-ml
      • 16-sampling-methods
      • 17-linear-systems
      • 18-convex-optimization
      • 19-complex-numbers
      • 20-fourier-transform
      • 21-graph-theory
      • 22-stochastic-processes
    • 02-ml-fundamentals
    • 03-deep-learning-core
    • 04-computer-vision
      • 01-image-fundamentals
      • 02-convolutions-from-scratch
      • 03-cnns-lenet-to-resnet
      • 04-image-classification
      • 05-transfer-learning
      • 06-object-detection-yolo
      • 07-semantic-segmentation-unet
      • 08-instance-segmentation-mask-rcnn
      • 09-image-generation-gans
      • 10-image-generation-diffusion
      • 11-stable-diffusion
      • 12-video-understanding
      • 13-3d-vision-nerf
      • 14-vision-transformers
      • 15-real-time-edge
      • 16-vision-pipeline-capstone
      • 17-self-supervised-vision
      • 18-open-vocab-clip
      • 19-ocr-document-understanding
      • 20-image-retrieval-metric
      • 21-keypoint-pose
      • 22-3d-gaussian-splatting
      • 23-diffusion-transformers-rectified-flow
      • 24-sam3-open-vocab-segmentation
      • 25-vision-language-models
      • 26-monocular-depth
      • 27-multi-object-tracking
      • 28-world-models-video-diffusion
    • 05-nlp-foundations-to-advanced
      • 01-text-processing
      • 02-bag-of-words-tfidf
      • 03-word-embeddings-word2vec
      • 04-glove-fasttext-subword
      • 05-sentiment-analysis
      • 06-named-entity-recognition
      • 07-pos-tagging-parsing
      • 08-cnns-rnns-for-text
      • 09-sequence-to-sequence
      • 10-attention-mechanism
      • 11-machine-translation
      • 12-text-summarization
      • 13-question-answering
      • 15-topic-modeling
      • 16-text-generation-pre-transformer
      • 17-chatbots-rule-to-neural
      • 18-multilingual-nlp
      • 19-subword-tokenization
      • 20-structured-outputs-constrained-decoding
      • 21-nli-textual-entailment
      • 22-embedding-models-deep-dive
      • 23-chunking-strategies-rag
      • 24-coreference-resolution
      • 25-entity-linking
      • 26-relation-extraction-kg
      • 27-llm-evaluation-frameworks
      • 28-long-context-evaluation
      • 29-dialogue-state-tracking
    • 06-speech-and-audio
      • 01-audio-fundamentals
      • 02-spectrograms-mel-features
      • 03-audio-classification
      • 04-speech-recognition-asr
      • 05-whisper-architecture-finetuning
      • 06-speaker-recognition-verification
      • 07-text-to-speech
      • 08-voice-cloning-conversion
      • 09-music-generation
      • 10-audio-language-models
      • 11-real-time-audio-processing
      • 12-voice-assistant-pipeline
      • 13-neural-audio-codecs
      • 14-voice-activity-detection-turn-taking
      • 15-streaming-speech-to-speech-moshi-hibiki
      • 16-anti-spoofing-audio-watermarking
      • 17-audio-evaluation-metrics
    • 07-transformers-deep-dive
      • 01-why-transformers
      • 02-self-attention-from-scratch
      • 03-multi-head-attention
      • 04-positional-encoding
      • 05-full-transformer
      • 06-bert-masked-language-modeling
      • 07-gpt-causal-language-modeling
      • 08-t5-bart-encoder-decoder
      • 09-vision-transformers
      • 10-audio-transformers-whisper
      • 11-mixture-of-experts
      • 12-kv-cache-flash-attention
      • 13-scaling-laws
      • 14-build-a-transformer-capstone
      • 15-attention-variants
      • 16-speculative-decoding
    • 08-generative-ai
      • 01-generative-models-taxonomy-history
      • 02-autoencoders-vae
      • 03-gans-generator-discriminator
      • 04-conditional-gans-pix2pix
      • 05-stylegan
      • 06-diffusion-ddpm-from-scratch
      • 07-latent-diffusion-stable-diffusion
      • 08-controlnet-lora-conditioning
      • 09-inpainting-outpainting-editing
      • 10-video-generation
      • 11-audio-generation
      • 12-3d-generation
      • 13-flow-matching-rectified-flows
      • 14-evaluation-fid-clip-score
      • 19-visual-autoregressive-var
    • 09-reinforcement-learning
    • 10-llms-from-scratch
    • 11-llm-engineering
    • 12-multimodal-ai
      • 01-vision-transformer-patch-tokens
      • 02-clip-contrastive-pretraining
      • 03-blip2-qformer-bridge
      • 04-flamingo-gated-cross-attention
      • 05-llava-visual-instruction-tuning
      • 06-any-resolution-patch-n-pack
      • 07-open-weight-vlm-recipes
      • 08-llava-onevision-single-multi-video
      • 09-qwen-vl-family-dynamic-fps
      • 10-internvl3-native-multimodal
      • 11-chameleon-early-fusion-tokens
      • 12-emu3-next-token-for-generation
      • 13-transfusion-autoregressive-diffusion
      • 14-show-o-discrete-diffusion-unified
      • 15-janus-pro-decoupled-encoders
      • 16-mio-any-to-any-streaming
      • 17-video-language-temporal-grounding
      • 18-long-video-million-token
      • 19-audio-language-whisper-to-af3
      • 20-omni-models-thinker-talker
      • 21-embodied-vlas-openvla-pi0-groot
      • 22-document-diagram-understanding
      • 23-colpali-vision-native-rag
      • 24-multimodal-rag-cross-modal
      • 25-multimodal-agents-computer-use
    • 13-tools-and-protocols
      • 01-the-tool-interface
      • 02-function-calling-deep-dive
      • 03-parallel-and-streaming-tool-calls
      • 04-structured-output
      • 05-tool-schema-design
      • 06-mcp-fundamentals
      • 07-building-an-mcp-server
      • 08-building-an-mcp-client
      • 09-mcp-transports
      • 10-mcp-resources-and-prompts
      • 11-mcp-sampling
      • 12-mcp-roots-and-elicitation
      • 13-mcp-async-tasks
      • 14-mcp-apps
      • 15-mcp-security-tool-poisoning
      • 16-mcp-security-oauth-2-1
      • 17-mcp-gateways-and-registries
      • 18-mcp-auth-production
      • 19-a2a-protocol
      • 20-opentelemetry-genai
      • 21-llm-routing-layer
      • 22-skills-and-agent-sdks
      • 23-capstone-tool-ecosystem
    • 14-agent-engineering
      • 01-the-agent-loop
      • 02-rewoo-plan-and-execute
      • 03-reflexion-verbal-rl
      • 04-tree-of-thoughts-lats
      • 05-self-refine-and-critic
      • 06-tool-use-and-function-calling
      • 07-memory-virtual-context-memgpt
      • 08-memory-blocks-sleep-time-compute
      • 09-hybrid-memory-mem0
      • 10-skill-libraries-voyager
      • 11-planning-htn-and-evolutionary
      • 12-anthropic-workflow-patterns
      • 13-langgraph-stateful-graphs
      • 14-autogen-actor-model
      • 15-crewai-role-based-crews
      • 16-openai-agents-sdk
      • 17-claude-agent-sdk
      • 18-agno-and-mastra-runtimes
      • 19-benchmarks-swebench-gaia
      • 20-benchmarks-webarena-osworld
      • 21-computer-use-agents
      • 22-voice-agents-pipecat-livekit
      • 23-otel-genai-conventions
      • 24-agent-observability-platforms
      • 25-multi-agent-debate
      • 26-failure-modes-agentic
      • 27-prompt-injection-defense
      • 28-orchestration-patterns
      • 29-production-runtimes
      • 30-eval-driven-agent-development
      • 31-agent-workbench-why-models-fail
      • 32-minimal-agent-workbench
      • 33-instructions-as-executable-constraints
      • 34-repo-memory-and-state
      • 35-initialization-scripts
      • 36-scope-contracts
      • 37-runtime-feedback-loops
      • 38-verification-gates
      • 39-reviewer-agent
      • 40-multi-session-handoff
      • 41-workbench-for-real-repos
      • 42-agent-workbench-capstone
    • 15-autonomous-systems
      • 01-long-horizon-agents
      • 02-star-family-reasoning
      • 03-alphaevolve-evolutionary-coding
      • 04-darwin-godel-machine
      • 05-ai-scientist-v2
      • 06-automated-alignment-research
      • 07-recursive-self-improvement
      • 08-bounded-self-improvement
      • 09-coding-agent-landscape
      • 10-claude-code-permission-modes
      • 11-browser-agents
      • 12-durable-execution
      • 13-cost-governors
      • 14-kill-switches-canaries
      • 15-propose-then-commit
      • 16-checkpoints-rollback
      • 17-constitutional-ai
      • 18-llama-guard
      • 19-anthropic-rsp
      • 20-openai-preparedness-deepmind-fsf
      • 21-metr-external-evaluation
      • 22-cais-caisi-societal-risk
    • 16-multi-agent-and-swarms
      • 01-why-multi-agent
      • 02-fipa-acl-heritage
      • 03-communication-protocols
      • 04-primitive-model
      • 05-supervisor-orchestrator-pattern
      • 06-hierarchical-architecture
      • 07-society-of-mind-debate
      • 08-role-specialization
      • 09-parallel-swarm-networks
      • 10-group-chat-speaker-selection
      • 11-handoffs-and-routines
      • 12-a2a-protocol
      • 13-shared-memory-blackboard
      • 14-consensus-and-bft
      • 15-voting-debate-topology
      • 16-negotiation-bargaining
      • 17-generative-agents-simulation
      • 18-theory-of-mind-coordination
      • 19-swarm-optimization-pso-aco
      • 20-marl-maddpg-qmix-mappo
      • 21-agent-economies
      • 22-production-scaling-queues-checkpoints
      • 23-failure-modes-mast-groupthink
      • 24-evaluation-coordination-benchmarks
      • 25-case-studies-2026-sota
    • 17-infrastructure-and-production
      • 01-managed-llm-platforms
      • 02-inference-platform-economics
      • 03-gpu-autoscaling-kubernetes
      • 04-vllm-serving-internals
      • 05-eagle3-speculative-decoding
      • 06-sglang-radixattention
      • 07-tensorrt-llm-blackwell
      • 08-inference-metrics-goodput
      • 09-production-quantization
      • 10-cold-start-mitigation
      • 11-multi-region-kv-locality
      • 12-edge-inference
      • 13-llm-observability
      • 14-prompt-semantic-caching
      • 15-batch-apis
      • 16-model-routing
      • 17-disaggregated-prefill-decode
      • 18-vllm-production-stack-lmcache
      • 19-ai-gateways
      • 20-shadow-canary-progressive
      • 21-ab-testing-llm-features
      • 22-load-testing-llm-apis
      • 23-sre-for-ai
      • 24-chaos-engineering-llm
      • 25-security-secrets-audit
      • 26-compliance-frameworks
      • 27-finops-llms
      • 28-self-hosted-serving-selection
    • 18-ethics-safety-alignment
      • 01-instruction-following-alignment-signal
      • 02-reward-hacking-goodhart
      • 03-direct-preference-optimization-family
      • 04-sycophancy-rlhf-amplification
      • 05-constitutional-ai-rlaif
      • 06-mesa-optimization-deceptive-alignment
      • 07-sleeper-agents-persistent-deception
      • 08-in-context-scheming-frontier-models
      • 09-alignment-faking
      • 10-ai-control-subversion
      • 11-scalable-oversight-weak-to-strong
      • 12-red-teaming-pair-automated-attacks
      • 13-many-shot-jailbreaking
      • 14-ascii-art-visual-jailbreaks
      • 15-indirect-prompt-injection
      • 16-red-team-tooling-garak-llamaguard-pyrit
      • 17-wmdp-dual-use-evaluation
      • 18-frontier-safety-frameworks-rsp-pf-fsf
      • 19-model-welfare-research
      • 20-bias-representational-harm
      • 21-fairness-criteria-group-individual-counterfactual
      • 22-differential-privacy-for-llms
      • 23-watermarking-synthid-stable-signature-c2pa
      • 24-regulatory-frameworks-eu-us-uk-korea
      • 25-echoleak-cves-for-ai
      • 26-model-system-dataset-cards
      • 27-data-provenance-training-governance
      • 28-alignment-research-ecosystem
      • 29-moderation-systems-openai-perspective-llamaguard
      • 30-dual-use-risk-cyber-bio-chem-nuclear
    • 19-capstone-projects
      • 01-terminal-native-coding-agent
      • 02-rag-over-codebase
      • 03-realtime-voice-assistant
      • 04-multimodal-document-qa
      • 05-autonomous-research-agent
      • 06-devops-troubleshooting-agent
      • 07-end-to-end-fine-tuning-pipeline
      • 08-production-rag-chatbot
      • 09-code-migration-agent
      • 10-multi-agent-software-team
      • 11-llm-observability-dashboard
      • 12-video-understanding-pipeline
      • 13-mcp-server-with-registry
      • 14-speculative-decoding-server
      • 15-constitutional-safety-harness
      • 16-github-issue-to-pr-agent
      • 17-personal-ai-tutor
      • 20-agent-harness-loop-contract
      • 21-tool-registry-schema-validation
      • 22-jsonrpc-stdio-transport
      • 23-function-call-dispatcher
      • 24-plan-execute-control-flow
      • 25-verification-gates-observation-budget
      • 26-sandbox-runner-denylist
      • 27-eval-harness-fixture-tasks
      • 28-observability-otel-traces
      • 29-end-to-end-coding-task-demo
      • 30-bpe-tokenizer-from-scratch
      • 31-tokenized-dataset-sliding-window
      • 32-token-positional-embeddings
      • 33-multihead-self-attention
      • 34-transformer-block
      • 35-gpt-model-assembly
      • 36-training-loop-eval
      • 37-loading-pretrained-weights
      • 38-classifier-finetuning
      • 39-instruction-tuning-sft
      • 40-dpo-from-scratch
      • 41-eval-pipeline
      • 42-large-corpus-downloader
      • 43-hdf5-tokenized-corpus
      • 44-cosine-lr-warmup
      • 45-gradient-clipping-amp
      • 46-gradient-accumulation
      • 47-checkpoint-save-resume
      • 48-distributed-fsdp-ddp
      • 49-lm-eval-harness
      • 50-hypothesis-generator
      • 51-literature-retrieval
      • 52-experiment-runner
      • 53-result-evaluator
      • 54-paper-writer
      • 55-critic-loop
      • 56-iteration-scheduler
      • 57-end-to-end-research-demo
      • 58-vision-encoder-patches
      • 59-vit-transformer
      • 60-projection-layer-modality-align
      • 61-cross-attention-fusion
      • 62-vision-language-pretraining
      • 63-multimodal-eval
      • 64-chunking-strategies-advanced
      • 65-hybrid-retrieval-bm25-dense
      • 66-reranker-cross-encoder
      • 67-query-rewriting-hyde
      • 68-rag-eval-precision-recall
      • 69-end-to-end-rag-system

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

AGENTS_cn.md

Lines changed: 218 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,218 @@
1+
# AGENTS.md
2+
3+
贡献者和 AI 代理参与本仓库的操作手册。在提交 PR 前请先阅读。
4+
5+
本仓库是一个课程体系,而非 SaaS 应用。课程内容才是核心产品。以下每一条规则都是为了确保 485 门课程长期保持一致性。
6+
7+
---
8+
9+
## 理念
10+
11+
485 门课程,20 个阶段。在引入任何框架之前,每个算法都从原始数学开始构建。你需要用 Python、TypeScript、Rust 或 Julia 手写反向传播、分词器、注意力机制和代理循环。然后用生产库运行同样的操作,让框架不再是黑盒。"从零构建 / 使用框架"是整个课程的脊梁。每门课程都产出一个可复用的制品,你可以直接集成到日常工作中。
12+
13+
---
14+
15+
## 仓库结构
16+
17+
```
18+
phases/
19+
NN-phase-slug/
20+
NN-lesson-slug/
21+
docs/en.md # 课程讲解文档
22+
code/ # 实现代码 + 测试
23+
quiz.json # 6 道测验题
24+
outputs/ # 可复用制品(技能 / 提示词 / 代理 / MCP 服务器)
25+
README.md # 门面文件;课程数量自动同步
26+
ROADMAP.md # 阶段/课程状态
27+
glossary/terms.md # 规范术语定义
28+
site/
29+
build.js # 解析 README + ROADMAP + 术语表 -> data.js
30+
data.js # 自动生成;CI 在 main 分支推送时重建
31+
scripts/ # 自动化脚本
32+
.github/workflows/
33+
curriculum.yml # 不变量检查 + 自动同步工作流
34+
```
35+
36+
---
37+
38+
## 硬性规则
39+
40+
1. **每个课程目录一个提交。** 绝不将多个课程打包进一个提交。10 个课程的 PR 应有 10 个提交。
41+
2. **约定式提交信息** 不超过 72 字符:`feat(phase-NN/MM): <slug>`。正文解释原因,而非描述内容。
42+
3. **图表仅使用 Mermaid 或 SVG。** 禁止 ASCII / Unicode 方框图。
43+
4. **每个围栏代码块必须标注语言。** 适当使用 `text``json``python``typescript``rust``julia``bash``console``mermaid``yaml`
44+
5. **仅限原创实现。** 不要在文档、代码注释或提交信息中引用外部课程仓库。引用 RFC、官方规范和学术论文作为规范来源。
45+
6. **依赖白名单**(见下文`依赖`)。优先使用标准库。
46+
7. **绝不提交生成文件**`catalog.json` 被 gitignore,`site/data.js` 由 CI 重建,`package-lock.json` 永不跟踪。
47+
48+
---
49+
50+
## 依赖
51+
52+
| 语言 | 允许使用的依赖 |
53+
|------------|---------------------------------------------------------------------------|
54+
| Python | `numpy``torch``h5py``zstandard``safetensors`、标准库 |
55+
| TypeScript | `hono``zod``ws`(仅在需要 WebSockets 时)、`@hono/node-server`、Node 20+ 标准库 |
56+
| Rust | 仅标准库(单文件 `rustc --edition 2021`|
57+
| Julia | `Random``Statistics``LinearAlgebra``Printf`(Julia 标准库) |
58+
59+
如果发现引用了被禁止的依赖,跳过并注明"保持标准库优先以确保教学清晰性"。
60+
61+
---
62+
63+
## 课程规范
64+
65+
### docs/en.md 前置元数据
66+
67+
```markdown
68+
# <标题>
69+
70+
> <一句话核心观点>
71+
72+
**类型:** <学习 | 构建 | 参考>
73+
**编程语言:** <与 code/ 中 main.* 文件匹配的逗号分隔列表>
74+
**前置条件:** <上游课程的逗号分隔列表,或"无">
75+
**时间:** ~<预计分钟数>
76+
77+
## 学习目标
78+
- <4-6 个以动词开头的要点>
79+
```
80+
81+
`**编程语言:**` 字段必须与 `code/` 中拥有 `main.*` 文件的语言匹配。
82+
83+
### quiz.json 结构
84+
85+
```json
86+
{
87+
"lesson": "<目录slug>",
88+
"title": "<课程标题>",
89+
"questions": [
90+
{"stage": "pre", "question": "...", "options": ["a","b","c","d"], "correct": 0, "explanation": ""},
91+
{"stage": "check", "question": "...", "options": ["a","b","c","d"], "correct": 1, "explanation": ""},
92+
{"stage": "check", "question": "...", "options": ["a","b","c","d"], "correct": 2, "explanation": ""},
93+
{"stage": "check", "question": "...", "options": ["a","b","c","d"], "correct": 1, "explanation": ""},
94+
{"stage": "post", "question": "...", "options": ["a","b","c","d"], "correct": 3, "explanation": ""},
95+
{"stage": "post", "question": "...", "options": ["a","b","c","d"], "correct": 0, "explanation": ""}
96+
]
97+
}
98+
```
99+
100+
恰好 6 道题:1 道预测试 + 3 道检查题 + 2 道后测试。`correct` 索引从 0 开始。网站渲染器只识别这种结构——旧版 `q/choices/answer` 结构会静默失败。
101+
102+
### code/
103+
104+
- 使用该语言的规范命令运行后能完整执行并退出码为 0。
105+
- 自终止演示。无无限 stdin 循环,不会因缺少 API 密钥而挂起。
106+
- 4-6 行头部注释,引用课程的 `docs/en.md` 路径和任何规范或 RFC 来源。
107+
108+
### code/tests/
109+
110+
- 至少 5 个单元测试。
111+
- 使用该语言的标准库运行器运行(`python3 -m unittest discover``npx tsx --test`、Rust/Julia 内联)。
112+
113+
---
114+
115+
## PR 验证
116+
117+
推送前在本地运行:
118+
119+
```bash
120+
python3 scripts/audit_lessons.py
121+
python3 scripts/check_readme_counts.py # 建议性检查——CI 在合并时修复
122+
123+
# 对每个涉及的课程:
124+
cd phases/NN-phase/MM-lesson/code
125+
python3 main.py && python3 -m unittest discover tests -v # 或对应语言的等效命令
126+
```
127+
128+
CI 门禁(`.github/workflows/curriculum.yml`):
129+
130+
| 任务 | 触发条件 | 行为 |
131+
|-------------------------------------|-------------|-------------------------------------------------------|
132+
| `audit` | push + PR | 运行 `audit_lessons.py`。阻塞性检查。 |
133+
| `readme-counts-sync`(仅 main) | 推送到 main | 重建 catalog + 自动修复 README 计数。 |
134+
| `site-rebuild`(仅 main) | 推送到 main | 重新运行 `node site/build.js`,提交 `site/data.js`|
135+
| `readme-counts-drift` | PR | 仅建议性检查——main 在合并时自动修复。 |
136+
137+
---
138+
139+
## 自动化契约
140+
141+
**CI 自动处理——请勿在你的 PR 中修改:**
142+
143+
| 表面 | 机器人 | 时机 |
144+
|----------------------|------------------------------|---------------------|
145+
| `catalog.json` | 按需重建(被 gitignore) | 每次 CI 任务 |
146+
| `README.md` 计数 | `readme-counts-sync` | 推送到 main 时 |
147+
| `site/data.js` | `site-rebuild` | 推送到 main 时 |
148+
149+
**你需要处理:**
150+
151+
| 表面 | 时机 |
152+
|-----------------------------------|------------------------------------------------------------------|
153+
| `README.md` 课程链接行 | 添加新课程时——链接 `[标题](phases/NN-phase/MM-lesson/)` |
154+
| `ROADMAP.md` 状态 | 标记课程完成或进行中时 |
155+
| `glossary/terms.md` | 引入被多门课程使用的术语时 |
156+
157+
**常见 bug**:如果合并后 `grep -c 'tree/main/phases/NN-' site/data.js` 为 0,则 Phase NN 的 README 行是纯文本,缺少 `[标题](phases/NN-...)` markdown 链接。`site/build.js` 从该链接派生 URL。
158+
159+
---
160+
161+
## 冲突解决
162+
163+
```bash
164+
git fetch origin main
165+
git merge --no-edit origin/main
166+
167+
# Catalog 冲突(仅旧分支——catalog.json 现已被 gitignore):
168+
git rm catalog.json
169+
git commit --no-edit
170+
171+
# README 计数冲突:
172+
git checkout --theirs README.md
173+
python3 scripts/build_catalog.py
174+
python3 scripts/check_readme_counts.py --fix
175+
git add README.md && git commit --no-edit
176+
177+
# site/data.js 冲突:
178+
git checkout --theirs site/data.js
179+
node site/build.js
180+
git add site/data.js && git commit --no-edit
181+
182+
git push origin <你的分支>
183+
```
184+
185+
避免对有未解决评审评论的分支执行 `git push --force`。强制推送会使评论与提交脱离关联。
186+
187+
---
188+
189+
## 新课程入门指南
190+
191+
```bash
192+
mkdir -p phases/NN-phase-slug/MM-new-lesson/{docs,code/tests,outputs}
193+
194+
# 1. 编写 docs/en.md,使用上述前置元数据。
195+
# 2. 编写 code/main.<lang>,包含 4-6 行头部注释。
196+
# 3. 编写 code/tests/test_main.*,包含 5+ 个测试。
197+
# 4. 编写 quiz.json,使用上述结构。
198+
# 5. (可选)如果课程产出技能文件,添加 outputs/skill-<slug>.md。
199+
200+
# 6. 添加到 README.md:
201+
# | MM | [课程标题](phases/NN-phase-slug/MM-new-lesson/) | 类型 | 语言 |
202+
203+
# 7. 更新 ROADMAP.md 状态行。
204+
205+
# 8. 本地验证。
206+
207+
# 9. 原子提交:
208+
git add phases/NN-phase-slug/MM-new-lesson README.md ROADMAP.md
209+
git commit -m "feat(phase-NN/MM): add <slug>"
210+
git push -u origin <你的分支>
211+
gh pr create --title "feat(phase-NN/MM): add <slug>" --body "<5行摘要>"
212+
```
213+
214+
`site/data.js` 在合并时重新生成——交给 CI 处理。
215+
216+
---
217+
218+
最后审阅:2026-05-27。

CHANGELOG_cn.md

Lines changed: 46 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,46 @@
1+
# 更新日志
2+
3+
本课程的最新更新内容,按时间倒序排列。
4+
5+
格式参考 [Keep a Changelog](https://keepachangelog.com/)。每条记录注明阶段、课程和变更内容,方便学习者直接查看差异。
6+
7+
## [未发布]
8+
9+
### 新增
10+
- `scripts/scaffold-lesson.sh` — 课程脚手架脚本,可创建 `phases/NN-phase/NN-lesson/` 完整目录结构,并从 `LESSON_TEMPLATE.md` 预填充 `docs/en.md` 骨架。
11+
- `.github/PULL_REQUEST_TEMPLATE.md` — 贡献者检查清单(代码可运行、无代码注释、从零构建优先、每课程原子提交、markdown 链接 ROADMAP 行)。
12+
- `.github/ISSUE_TEMPLATE/bug_report.md``new_lesson_proposal.md` — 用于 bug 报告和课程提案的结构化模板。
13+
-`CHANGELOG.md` 文件。
14+
15+
## 2026-04 — 阶段 4:计算机视觉 完成
16+
17+
### 新增
18+
- 阶段 4 全部 28 门课程,涵盖图像基础到多模态视觉(VLM、3D、视频、自监督学习)。
19+
- `ROADMAP.md` 中的阶段 4 行已链接为 markdown 链接指向课程目录,使网站能够正确展示。
20+
21+
### 修复
22+
- 阶段 4 跨 15+ 门课程的精度修正:
23+
- `phase-4/02`:形状计算器指定自适应池化、展平和线性层的 RF/步幅处理。
24+
- `phase-4/03`:骨干网络选择器描述列出所有覆盖的系列;为 OCR、医疗、工业场景添加了头部指导。
25+
- `phase-4/04`:分类诊断使用量化阈值针对每种失败模式;`n/a` 用于未定义指标;对少于 3 个类别的情况增加防护。
26+
- `phase-4/06`:检测指标读取器使用 `AP@0.5`(而非 `mAP@0.5`);逐类召回率声明为可选;锚点设计器澄清步幅截断和每层单锚点路径。
27+
- `phase-4/10`:采样器选择器声明 `unet_forward_ms` 作为输入;ControlNet 防护提升为规则 0。
28+
- `phase-4/14`:ViT 检查器与拒绝规则对齐——端口尝试被审计而非背书。
29+
- `phase-4/24`:开放词汇堆栈选择器具有显式规则优先级和许可证过滤语义;概念设计器解决步骤-5/规则-80 冲突。
30+
- `phase-4/25`:VLM 文档中 `_merge` 在占位符不匹配时抛出描述性 `ValueError`;CMER 内部进行归一化。
31+
- `phase-4/27``synthetic_frames` 将 GT 边界框裁剪到帧的高/宽。
32+
- `phase-4/28``rope_3d` 验证维度拆分;从 DiT 块示例中移除了未使用的 `F` 导入。
33+
34+
## 2026-Q1 及更早
35+
36+
### 新增
37+
- 阶段 0(环境搭建与工具):全部 12 门课程。
38+
- 阶段 1(数学基础):全部 22 门课程。
39+
- 阶段 2(机器学习基础):全部 18 门课程。
40+
- 阶段 3(深度学习核心):涵盖感知机、反向传播、优化器的核心课程。
41+
- 内置 Claude Code 技能:`find-your-level`(分班测试)和 `check-understanding`(逐阶段测验)。
42+
- 网站 aiengineeringfromscratch.com:课程目录、逐课程页面、路线图、277 条术语表。
43+
- 所有 20 个阶段的初始脚手架(`phases/00-*``phases/19-*`)。
44+
- `LESSON_TEMPLATE.md``CONTRIBUTING.md``ROADMAP.md``README.md`
45+
46+
[未发布]: https://github.com/rohitg00/ai-engineering-from-scratch/compare/HEAD...HEAD

CODE_OF_CONDUCT_cn.md

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,30 @@
1+
# 行为准则
2+
3+
## 我们的承诺
4+
5+
我们致力于让每个人都能参与本项目,不受年龄、体型、残疾、种族、性别认同与表达、经验水平、教育程度、社会经济地位、国籍、外貌、肤色、宗教、性取向的骚扰影响。
6+
7+
## 我们的标准
8+
9+
**积极行为:**
10+
11+
- 使用友好和包容的语言
12+
- 尊重不同的观点和经验
13+
- 虚心接受建设性批评
14+
- 以社区的最佳利益为重
15+
- 对其他社区成员表现出同理心
16+
17+
**不可接受的行为:**
18+
19+
- 挑衅、侮辱/贬损性评论以及人身或政治攻击
20+
- 公开或私下的骚扰
21+
- 未经明确许可发布他人的私人信息
22+
- 其他可能被合理认为不当的行为
23+
24+
## 执行
25+
26+
如遇虐待、骚扰或其他不可接受的行为,请联系项目维护者 ghumare64@gmail.com 进行举报。所有投诉都将被审查和调查。
27+
28+
## 来源
29+
30+
本行为准则改编自 [Contributor Covenant](https://www.contributor-covenant.org) 2.1 版。

0 commit comments

Comments
 (0)