Phi
Microsoft's small models trained mainly on curated textbook-quality data.
Aperçu
Phi is a family of small language models trained on a deliberately narrow, high-quality corpus of textbooks and exercises rather than on a web-scale crawl. It solves the assumption that capability requires scale, showing that data selection can substitute for a large part of it. It exists because it reframes the question from how much data to what kind of data, and the answer is useful when the budget is a single GPU.
Graphe des technologies
Comment Phi se rattache au reste de la pile.
Utilise et dépend de
- Transformer architecture Algorithms Phi based on Transformer High ✓ source
Outils
Logiciels qui implémentent Phi ou s'y intègrent.
Infrastructure
Alternatives
Comparaisons
Comparaisons détaillées avec les alternatives ci-dessus.