150w+点击CSDN: AI生成曾小健
发起项目主要有大模型训练，和大模型代码解析(逐行)两块。还有一些重大项目的微小贡献。
Baichuan微调训练 *** 百川微调训练源码解析
ChatGLM2-6B-Explained *** ChatGLM-Efficient-Tuning-Explained
| Company | Project | Stars | Description |
|---|---|---|---|
 |
llama3 |  |
Meta的最强开源模型 |
| transformers |  |
自然语言理解的基础框架 | |
 |
Qwen |  |
阿里7B/14B SOTA聊天LLM项目 |
 |
Baichuan2 |  |
百川智能第二代7-13B（9.6发布） |
|
swift |  |
阿里通义千问实验室官方出品训练框架，支持35+LLM/MLM |
 |
lmdeploy |  |
上海人工智能实验室出品的大模型部署框架 |
 |
RAGFlow |  |
顶级RAG框架 |
 |
fastllm |  |
大模型加速/量化 |
 |
Awesome-Prompt-Engineering |  |
提示工程集合 |
 |
PaddleNLP |  |
百度通用基础NLP框架 |
 |
Pycaret |  |
顶级自动化机器学习框架 |
 |
AutoX |  |
第四范式出品自动化机器学习框架 |
 |
ProG |  |
图神经/图提示网络统一框架(2023 KDD Best Paper) |
 |
torchlm |  |
Torch人脸/关键点系列框架 |
| Total | ⭐218.8k |
| Project | Description | Stars |
|---|---|---|
| llama3 | Meta's most powerful open-source model | |
| transformers | Fundamental framework for natural language understanding | |
| Qwen | Alibaba's Qwen 7B/14B SOTA chat project | |
| Baichuan2 | Second version of Baichuan 7-13B (released 9.6) | |
| RAGFlow | RAG framework with nearly 10k stars | |
| fastllm | Large model acceleration/quantization | |
| Awesome-Prompt-Engineering | Prompt engineering collection | |
| PaddleNLP | Baidu's general NLP framework | |
| Pycaret | Top automated machine learning framework | |
| AutoX | Automated machine learning | |
| ProG | Graph neural network framework | |
| torchlm | Torch-based framework | |
| Total | ⭐218.8k |
The initiated projects mainly include large model training and large model code analysis (line by line).
There are also minor contributions to major projects.
Baichuan fine-tuning training *** Baichuan fine-tuning training source code analysis
ChatGLM2-6B-Explained ***
ChatGLM-Efficient-Tuning-Explained Source Code Contributor