tatsu-lab/stanford_alpaca
Code and documentation to train Stanford's Alpaca models, and generate the data.
Generates 52K instruction-following training data using text-davinci-003 via aggressive batch decoding (20 examples at once), reducing costs to under $500. Fine-tunes LLaMA-7B/13B using Hugging Face with standard supervised learning on structured instruction-input-output triples. Includes complete reproducible pipeline: data generation code, dataset, training recipes, and weight diff recovery for reconstructing model checkpoints from LLaMA base weights.
30,267 stars. No commits in the last 6 months.
Stars
30,267
Forks
4,011
Language
Python
License
Apache-2.0
Category
Last pushed
Jul 17, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/tatsu-lab/stanford_alpaca"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
limix-ldm-ai/LimiX
LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligence...
XXO47OXX/layer-scan
Automated LLM layer duplication config scanner — find the optimal (i,j) for any model + task
google-research/plur
PLUR (Programming-Language Understanding and Repair) is a collection of source code datasets...
thuml/LogME
Code release for "LogME: Practical Assessment of Pre-trained Models for Transfer Learning" (ICML...
Kakz/prometheus-llm
PrometheusLLM is a unique transformer architecture inspired by dignity and recursion. This...