kyo-takano/chinchilla

A toolkit for scaling law research ⚖

42
/ 100
Emerging

This toolkit helps deep learning researchers and practitioners efficiently scale the training of large models like LLMs or Vision Transformers. You provide data from multiple model training runs (compute, parameters, data, and loss), and it outputs optimized configurations for model size and data usage to achieve the best performance within a specific compute budget. It's for anyone pushing the boundaries of large-scale deep learning model development.

No commits in the last 6 months. Available on PyPI.

Use this if you are developing large deep learning models and need to optimize your compute resources to achieve the best possible model performance.

Not ideal if you are working with fine-tuning tasks or in domains with scarce data.

deep-learning-research large-language-models model-scaling compute-optimization machine-learning-engineering
Stale 6m
Maintenance 0 / 25
Adoption 8 / 25
Maturity 25 / 25
Community 9 / 25

How are scores calculated?

Stars

57

Forks

4

Language

Python

License

Apache-2.0

Last pushed

Jan 27, 2025

Commits (30d)

0

Dependencies

8

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/kyo-takano/chinchilla"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.