gaia.models.gaia_transformer module

GAIA Transformer - Complete LLM with Categorical Deep Learning

A full transformer/LLM implementation using ALL GAIA components: - Simplicial structures for hierarchical processing - F-coalgebras for generative modeling - Yoneda metrics for enhanced attention - Spectral normalization for stability - Hierarchical message passing - Ends/coends for integral calculus - Kan complex verification

This represents the pinnacle of categorical deep learning applied to language modeling.

gaia.models.gaia_transformer.create_gaia_llm(vocab_size, model_size='base', use_all_gaia_features=True)[source]

Create a GAIA LLM with predefined configurations

Parameters:

vocab_size (int) – Vocabulary size
model_size (str) – ‘small’, ‘base’, ‘large’, or ‘xl’
use_all_gaia_features (bool) – Whether to use all GAIA enhancements

Returns:

Configured GAIATransformer

Return type: