CoViPAL: Layer-wise Contextualized Visual Token Pruning for Large Vision-Language Models

Open in new window