Sea-Snell/grokking

unofficial re-implementation of "Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets"

42
/ 100
Emerging

This project helps machine learning researchers and students explore an unusual deep learning phenomenon called "grokking." It takes configuration settings for neural network training and outputs visualizations of model performance over time, showing how a model can suddenly generalize well long after it appeared to overfit. This tool is designed for those studying generalization in artificial intelligence.

No commits in the last 6 months.

Use this if you are a machine learning researcher or student interested in replicating and experimenting with the 'grokking' phenomenon described in the Power et al. paper.

Not ideal if you are looking for a tool to solve a practical, real-world machine learning problem or to train production models.

deep-learning-research neural-network-training model-generalization machine-learning-phenomena ai-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 17 / 25

How are scores calculated?

Stars

86

Forks

16

Language

Python

License

MIT

Last pushed

Jul 04, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/Sea-Snell/grokking"

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