# bbruceyuan/LLMs-Zero-to-Hero

从无名小卒到大模型（LLM）大英雄~ 欢迎关注后续！！！

Repository: https://github.com/bbruceyuan/LLMs-Zero-to-Hero
Canonical: https://ross.abutalabs.com/products/llms-zero-to-hero
Homepage: https://yuanchaofa.com/llms-zero-to-hero/
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: llm, llm101, llm-from-zero-to-hero, llm-zero-to-hero
Last push: 2026-08-16T11:31:16+00:00

## Health v2 (maintenance only)
Score: 65/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 35, longevity 43
- inputs: {"age_days": 605, "days_push": 17, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2278, forks 155 (observed 2026-08-28T04:06:33.671099+00:00)

## What it is
A Chinese-language tutorial series and accompanying book that teaches large language models by building them from scratch in Jupyter notebooks, inspired by Andrej Karpathy's Zero to Hero style. It covers pre-training, SFT, DPO, RLHF, MoE architectures, multimodal models, and deployment with runnable code and video lectures.

## Use cases
- learn how large language models work by implementing them from scratch
- train a mini GPT from zero on a single consumer GPU
- understand LLM fine-tuning techniques like SFT, DPO, and RLHF
- build a transformer and MoE model without relying on transformers library
- find a structured hands-on LLM course with videos and notebooks
- learn LLM deployment, inference optimization, and quantization

## When to choose
- you want to deeply understand LLM internals by writing every line yourself rather than calling APIs
- you prefer Chinese-language tutorials with video walkthroughs and runnable notebooks
- you have a single RTX 3090/4090 and want to train small models end to end

## When to avoid
- you just want to use LLM APIs or the transformers library in production
- you need English-only learning materials
- you need a production-ready training framework rather than educational code

## Facets
- artifact type: learning-resource
- maturity: active
- function: llm-training, machine-learning, deep-learning
- domain: large-language-models, deep-learning, tutorials, education
- platform: python
- tags: jupyter-notebooks, hands-on, karpathy-style, pretraining, fine-tuning, sft, dpo, rlhf, chinese-content, nanogpt, gpu

## Member repositories
- bbruceyuan/LLMs-Zero-to-Hero (main) score 65

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:33.671099+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:41:24.591863+00:00, confidence not recorded.
  - readme: https://github.com/bbruceyuan/LLMs-Zero-to-Hero (fetched 2026-08-28T04:06:33.671099+00:00, sha e41108d88272)
  - homepage: https://yuanchaofa.com/llms-zero-to-hero/ (fetched 2026-08-29T10:21:45.443620+00:00, sha a9e8a8b139f9)
- Data as of 2026-08-30T08:39:29.467469+00:00.