AI Engineering Institute: 3Fine-tuning (fine-tuning of large language models)
📚 Structure of the database
| Models/Catalog | Description and content |
|---|---|
| Axolotl | A framework for fine-tuning language models |
| Gemma | Google's latest implementation of the Big Language Model |
- finetune-gemma.ipynb - gemma-sft.py - Gemma_finetuning_notebook.ipynb | Fine-tuning notebooks and scripts |
| LLama2 | Meta's Open Source Large Language Model |
- generate_response_stream.py - Llama2_finetuning_notebook.ipynb - Llama_2_Fine_Tuning_using_QLora.ipynb | Implementation and fine-tuning guidelines |
| Llama3 | Upcoming Meta Large Language Modeling Experiments |
- Llama3_finetuning_notebook.ipynb | Initial fine-tuning experiments |
| LlamaFactory | A Framework for Training and Deployment of Large Language Models |
| LLMArchitecture/ParameterCount | Technical details of the model architecture |
| Mistral-7b | Mistral AI The 7 billion parameter model |
- LLM_evaluation_harness_for_Arc_Easy_and_SST.ipynb - Mistral_Colab_Finetune_ipynb_Colab_Final.ipynb - notebooks_chatml_inference.ipynb - notebooks_DPO_fine_tuning.ipynb - notebooks_SFTTrainer TRL.ipynb - SFT.py | Integrated notebook for assessment, fine-tuning and reasoning |
| Mixtral | Mixtral's Expert Mixing Model |
- Mixtral_fine_tuning.ipynb | Fine-tuning Realization |
| VLM | visual language model |
- Florence2_finetuning_notebook.ipynb - PaliGemma_finetuning_notebook.ipynb | Visual language model implementation |
🎯 Module Overview
1. LLM architecture
- Explore the following model implementations:
- Llama2 (Meta's open source model)
- Mistral-7b (efficient 7 billion parameter model)
- Mixtral (expert hybrid architecture)
- Gemma (Google's latest contribution)
- Llama3 (upcoming experiment)
2. 🛠️ fine-tuning technology
- implementation strategy
- The LoRA (Low Rank Adaptation) approach
- Advanced Optimization Methods
3. 🏗️ model architecture analysis
- An in-depth study of the model structure
- Parameter calculation method
- Scalability Considerations
4. 🔧 Professional realization
- Code Llama for programming tasks
- Visual language modeling:
- Florence2
- PaliGemma
5. 💻 Practical applications
- Integrated Jupyter Notebook
- Response Generation Pipeline
- Reasoning Implementation Guide
6. 🚀 Advanced Themes
- DPO (Direct Preference Optimization)
- SFT (supervised fine tuning)
- Assessment methodology
© Copyright notes
The copyright of the article belongs to the author, please do not reprint without permission.
Related posts
![[转载]QwQ-32B 的工具调用能力及 Agentic RAG 应用](https://oss.sharenet.ai/wp-content/uploads/2025/03/b04be76812d1a15.jpg)
No comments...
English 
简体中文 
日本語 
한국어 
Español 
Deutsch 
Français 
Português do Brasil