AB
AB
AiBoss

All Categories

project

self-llm - An open-source large model tutorial specifically designed for Chinese developers

self-llm (Open Source Large Model Usage Guide) is an open source large model tutorial created by Datawhale specifically for beginners in China. Based on the Linux platform, it provides a complete guide from environment configuration to model deployment and fine-tuning, covering LLaMA...

What is self-llm?

self-llm (Open Source Large Model User Guide) is an open source large model tutorial created by Datawhale specifically for beginners in China. Based on the Linux platform, it provides a complete guide from environment configuration to model deployment and fine-tuning, covering mainstream models such as LLaMA and ChatGLM. The project simplifies the usage process of open source large models, helping students and researchers get started quickly. The tutorial provides detailed tutorials, fine-tuning methods (such as LoRA and ptuning), and application examples for various models, suitable for learners of different levels. The project encourages community participation to jointly improve the content and promote the popularization and application of open source large models.

The main functions of self-llm

  • Environment Configuration GuideProvides a configuration guide for open-source large model environments based on the Linux platform, helping users build a basic environment suitable for running different models.
  • Model Deployment TutorialIt covers local deployment methods for mainstream open-source models (such as LLaMA, ChatGLM, InternLM, etc.) both domestically and internationally, including command-line calls and online demo deployment.
  • Fine-tuning method guideIt provides detailed tutorials on efficient fine-tuning methods such as full fine-tuning, LoRA fine-tuning, and ptuning, helping users to customize and optimize models according to their needs.
  • Application Development GuideIt combines frameworks such as LangChain to guide users on how to integrate open-source large models into practical applications and create domain-specific private domain models.
  • Community co-creation and supportUsers are encouraged to submit issues or contribute code pull requests to improve the tutorial content and form a collaborative and supportive system for the open-source community.

Models supported by self-LLM

  • GLM-4.1-Thinking
    • GLM-4.1V-Thinking vLLM Deployment Invocation
    • GLM-4.1V-Thinking Gradio Deployment
    • GLM-4.1V-Thinking LoRA Fine-tuning and SwanLab Visual Recording
    • GLM-4.1V-Thinking Docker image
  • GLM-4.5-Air
    • GLM-4.5-Air vLLM Deployment Invocation
    • GLM-4.5-Air EvalScope IQ and EQ Assessment
    • GLM-4.5-Air LoRA Fine-tuning
    • GLM-4.5-Air Ucloud Docker image
  • ERNIE-4.5
    • ERNIE-4.5-0.3B-PT LoRA fine-tuning and SwanLab visualization recording
    • ERNIE-4.5-0.3B-PT LoRA Docker image
  • Hunyuan-A13B-Instruct
    • Hunyuan-A13B-Instruct Model Architecture Analysis
    • Hunyuan-A13B-Instruct SGLang Deployment Invocation
    • Hunyuan-A13B-Instruct LoRA SwanLab Visual Fine-tuning
    • Hunyuan-A13B-Instruct LoRA Docker image
  • Qwen3
    • Qwen3 Model Structure Analysis
    • Qwen3-8B vllm deployment and invocation
    • Qwen3-8B Windows LMStudio Deployment and Invocation
    • Qwen3-8B Evalscope IQ and EQ Assessment
    • Qwen3-8B LoRA Fine-tuning and SwanLab Visual Recording
    • Qwen3-30B-A3B Fine-tuning and SwanLab Visual Recording
    • Qwen3 Think Decryption
    • Qwen3-8B Docker image
    • Uses of Qwen3-0.6B small model
    • Qwen3-1.7B Medical Reasoning Dialogue Fine-tuning and SwanLab Visual Recording
    • Qwen3-8B GRPO Fine-tuning and SwanLab Visualization
  • Kimi-VL-A3B
    • Kimi-VL-A3B Technical Report Interpretation
    • Deployment of Kimi-VL-A3B-Thinking WebDemo (Web Dialogue Assistant)
  • Llama4
    • Llama4 Conversation Assistant
    • SpatialLM
    • SpatialLM 3D Point Cloud Understanding and Object Detection Model Deployment
    • Hunyuan3D-2
    • Hunyuan3D-2 series model deployment
    • Hunyuan3D-2 series model code calling
    • Hunyuan3D-2 series models Gradio deployment
    • Hunyuan3D-2 Series Model API Server
    • Hunyuan3D-2 Docker image
  • Gemma3
    • Gemma-3-4b-it FastAPI Deployment and Invocation
    • Gemma-3-4b-it ollama + open-webui deployment
    • Gemma-3-4b-it Evalscope IQ and EQ Assessment
    • Gemma-3-4b-it LoRA fine-tuning
    • Gemma-3-4b-it Docker image
    • Gemma-3-4b-it GRPO fine-tuning and SwanLab visualization
  • DeepSeek-R1-Distill
    • DeepSeek-R1-Distill-Qwen-7B FastAPI Deployment and Invocation
    • DeepSeek-R1-Distill-Qwen-7B Langchain Access
    • Deployment of DeepSeek-R1-Distill-Qwen-7B WebDemo
    • DeepSeek-R1-Distill-Qwen-7B vLLM Deployment and Invocation
    • DeepSeek-R1-0528-Qwen3-8B-GRPO and SwanLab Visualization
  • MiniCPM-o-2_6
    • MiniCPM-o-2.6 FastAPI Deployment and Invocation
    • MiniCPM-o-2.6 WebDemo Deployment
    • MiniCPM-o-2.6 Multimodal Speech Capabilities
    • MiniCPM-o-2.6 Visualization: LaTeX OCR LoRA Fine-tuning
  • InternLM3
    • InternLM3-8b-instruct FastAPI Deployment Invocation
    • InternLM3-8b-instruct Langchain Access
    • InternLM3-8b-instruct WebDemo Deployment
    • InternLM3-8b-instruct LoRA fine-tuning
    • InternLM3-8b-instruct o1-like inference chain implementation
  • phi4
    • Phi4 FastAPI Deployment and Invocation
    • phi4 Langchain integration
    • phi4 WebDemo Deployment
    • phi4 LoRA fine-tuning
    • phi4 LoRA Fine-tuning NER Task: SwanLab Visual Record
    • phi4 GRPO fine-tuning and SwanLab visualization
  • Qwen2.5-Coder
    • Qwen2.5-Coder-7B-Instruct FastApi Deployment Invocation
    • Qwen2.5-Coder-7B-Instruct Langchain Access
    • Deploying Qwen2.5-Coder-7B-Instruct WebDemo
    • Qwen2.5-Coder-7B-Instruct vLLM Deployment
    • Qwen2.5-Coder-7B-Instruct LoRA Fine-tuning
    • Qwen2.5-Coder-7B-Instruct LoRA Fine-tuning SwanLab Visual Record Version
  • Qwen2-vl
    • Qwen2-vl-2B FastAPI Deployment and Invocation
    • Qwen2-vl-2B WebDemo Deployment
    • Qwen2-vl-2B vLLM Deployment
    • Qwen2-vl-2B LoRA fine-tuning
    • Qwen2-vl-2B LoRA Fine-tuning SwanLab Visual Recording Version
    • Qwen2-vl-2B LoRA Fine-tuning Case Study – LaTeXOCR
  • Qwen2.5
    • Qwen2.5-7B-Instruct FastAPI Deployment Invocation
    • Qwen2.5-7B-Instruct Langchain Access
    • Qwen2.5-7B-Instruct vLLM Deployment Invocation
    • Deploying Qwen2.5-7B-Instruct WebDemo
    • Qwen2.5-7B-Instruct LoRA Fine-tuning
    • Qwen2.5-7B-Instruct o1-like inference chain implementation
    • Qwen2.5-7B-Instruct LoRA Fine-tuning SwanLab Visual Record Version
  • Apple OpenELM
    • OpenELM-3B-Instruct FastAPI Deployment Invocation
    • OpenELM-3B-Instruct LoRA Fine-tuning
  • Llama3_1-8B-Instruct
    • Llama3_1-8B-Instruct FastApi Deployment Invocation
    • Llama3_1-8B-Instruct Langchain Access
    • Llama3_1-8B-Instruct WebDemo Deployment
    • Llama3_1-8B-Instruct LoRA Fine-tuning
    • Convert your GGUF model and deploy it locally using Ollama
  • Gemma-2-9b-it
    • Gemma-2-9b-it FastAPI Deployment and Invocation
    • Gemma-2-9b-it Langchain Access
    • Gemma-2-9b-it WebDemo Deployment
    • Gemma-2-9b-it Peft LoRA Fine-tuning
  • Yuan2.0
    • Yuan2.0-2B FastAPI Deployment and Invocation
    • Yuan2.0-2B Langchain Access
    • Deployment of Yuan2.0-2B WebDemo
    • Yuan2.0-2B vLLM Deployment and Invocation
    • Yuan 2.0-2B LoRA Fine-tuning
  • Yuan2.0-M32
    • Deploying and calling Yuan2.0-M32 FastAPI
    • Yuan2.0-M32 Langchain Access
    • Deployment of Yuan2.0-M32 WebDemo
  • DeepSeek-Coder-V2
    • DeepSeek-Coder-V2-Lite-Instruct FastApi Deployment Invocation
    • DeepSeek-Coder-V2-L

self-llm project address

  • GitHub repositoryhttps://github.com/datawhalechina/self-llm

Application scenarios of self-llm

  • Education and LearningIt provides beginners with a complete guide from environment configuration to model fine-tuning, helping students and developers quickly master the use of large open-source models.
  • Enterprise application developmentIt helps enterprises customize and fine-tune open-source models according to their own business needs, and develop exclusive applications such as intelligent customer service and knowledge management.
  • Personal project developmentIt supports individual developers in using open-source large models to develop projects such as intelligent writing assistants and personal intelligent assistants, thereby improving personal productivity.
  • Research and InnovationIt provides researchers with an experimental platform to support research on the optimization of open-source large models, improvement of training methods, and other research work.
Back