Optimized for fast, high-quality generation with minimal inference steps. Delivers strong results at turbo speed.
Distilled Transformer
Natural language
Fast inference, quality at low step counts