Stanford Alpaca is an open-source project that demonstrates the capabilities of an instruction-following LLaMA model. Developed by a team of researchers at Stanford University, Alpaca is designed to understand and execute tasks based on user instructions. The project provides a dataset, data generation process, and fine-tuning code for reproducibility.
| General | |
|---|---|
| Repository | https://github.com/tatsu-lab/stanford_alpaca |
| Type | Instruction-Following LLaMA Model |
Alpaca - Resources
Resources to get started with Stanford Alpaca:
- Stanford Alpaca GitHub Repository
- Release Blog Post Detailed information about the Alpaca model, its potential harm and limitations, and the team's thought process behind releasing a reproducible model.
Alpaca - General Information
- Fine-tuned from a 7B LLaMA model on 52K instruction-following data
- Built using techniques from the Self-Instruct paper with modifications
- Capable of understanding and executing tasks based on user instructions
- Provides a dataset, data generation process, and fine-tuning code for reproducibility
Alpaca - Setup
Setup instructions for Stanford Alpaca:
- Data Generation Process Instructions on generating the instruction-following dataset used for fine-tuning the Alpaca model.
- Fine-tuning A guide to fine-tune LLaMA and OPT models using the provided dataset and code.
- Recovering Alpaca Weights Instructions for recovering the Alpaca-7B weights from the released weight diff.
.png&w=96&q=75){kind=small}
.png&w=1920&q=75)
