Overview

🚀LLaMA2-Accessory is an open-source toolkit for pretraining, finetuning and deployment of Large Language Models (LLMs) and mutlimodal LLMs. This repo is mainly inherited from LLaMA-Adapter with more advanced features.🧠

Get started

• Installation Guide
  • 1. Basic Setup
  • 2. Optional: Install Flash-Attention
  • 3. Optional: Install Apex
  • 4. Install LLaMA2-Accessory as Python Packege

Tutorials

• Pretrain
  • Vanilla Dataset
  • Packed Dataset
• Finetune
  • Prerequisites
  • Language-only Full-Parameter Finetuning
  • Language-only Parameter-Efficient Finetuning
  • Multi-Modal Full-Parameter Finetuning
  • Finetuning with Quantization
• Inference
  • Preparation
  • General Pipeline
  • Host Local Demos
  • Model Zoo
• Evaluation
  • Evaluation with OpenCompass
• Light-eval
  • Multimodel evaluation
  • Language Model evaluation

Projects

• SPHINX
• mixtral-8x7b
  • Features
  • Model Implementation
  • Install
  • Prepare Checkpoint
  • Inference
  • Finetuning
  • Acknowledgement

Help

• Frequently Asked Questions (FAQ)
  • Install & Preperation
  • Model