You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We apply Large Concept Models to image captioning via a prefix-based framework, outperforming GPT-2 baselines on all major metrics (BLEU, METEOR, ROUGE, CIDEr, BERTScore) under frozen language model training