// the find
youngyangyang04/llm-master
大模型（LLM）全栈学习路线与中文教程🔥：覆盖 Prompt Engineering、RAG、AI Agent、MCP、微调、模型部署、Transformer、AI 编程与大厂面试，从入门到生产实践。
A Chinese-language learning path for application developers who want to move into LLM engineering, covering model API calls, RAG, agents, fine-tuning, deployment, Transformer internals, and interview prep. It is a set of markdown tutorials rather than a codebase, and it is only useful if you read Chinese and are building applications rather than doing research.
On the README's own description, the stages are ordered by dependency, and each names an entry doc and a concrete completion check, such as being able to draw the full request path for one model call. The RAG and agent topics cover the failure side (evaluation, failure modes, recovery) that many tutorials skip. The Transformer track builds from Q/K/V through attention, FFN, LayerNorm and residuals to a tiny implementation, so the code sits at a clear point in the sequence. The interview notes are split by topic (RAG, agents, fine-tuning, Transformer) rather than collected into one list.
There is no project code to run. Apart from snippets inside the markdown, the four resume projects exist only as deliverable checklists, so nothing can be executed or tested end to end. The AI-usage row links to account top-up and API relay repos, which reads as a sales funnel inside a learning resource and should be weighed accordingly. The news directory holds dozens of dated model-release write-ups that will go stale quickly, and the continuous-update promise leans on that churn without any expiry marking. The README is written in a promotional register (emoji feature tables, star calls to action), which says little about the depth of the articles behind it.