Stories by charleshmartin

Setol: SemiEmpirical Theory of (Deep) Learning

Setol: A SemiEmpirical of (Deep) Learning

The Magic of Mistral, a Story of Dragon Kings

Evaluating Fine-Tuned LLMs with WeightWatcher

Deep Learning and Effective Correlation Spaces

Weightwatcher: Data-Free Diagnostics for Deep Learning

Better than BERT: Pick your best model

Predicting trends in the quality of state-of-the-art neural networks

How to tell if you have trained your model with enough data

Predicting trends in the quality of state-of-the-art neural networks

Is Your Model Overtrained?

WeightWatcher: Empirical Quality Metrics for Deep Neural Networks

Towards a New Theory of Learning: Statistical Mechanics of Deep Neural Networks