DeepSeek
Open-weight model family notable for efficient training at frontier scale.
Aperçu
DeepSeek is an open-weight model family whose published training methods attracted attention for reaching competitive quality at lower reported cost. It solves the cost question in training rather than in inference, which is where most optimisation effort goes. It exists as evidence that architecture and training-schedule choices matter enough to change what a given budget can produce, and its weights and methods are published.
Graphe des technologies
Comment DeepSeek se rattache au reste de la pile.
Utilise et dépend de
- Transformer architecture Algorithms DeepSeek based on Transformer High ✓ source
Appliqué à
- MoE architecture Algorithms MoE used for DeepSeek High ✓ source
Outils
Logiciels qui implémentent DeepSeek ou s'y intègrent.
Infrastructure
Alternatives
Comparaisons
Comparaisons détaillées avec les alternatives ci-dessus.