NCMNext-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation

(scalingintelligence.stanford.edu)

2 points | by matt_d 8 hours ago ago

No comments yet.