Want to create an interactive transcript for this episode?
Podcast: Arxiv Papers
Episode: GenRecal: Generation after Recalibration from Large to Small Vision-Language Models
Description: GenRecal is a novel distillation framework for vision-language models that enhances knowledge transfer across diverse architectures, improving performance on resource-constrained devices while outperforming large-scale VLMs.https://arxiv.org/abs//2506.15681YouTube: https://www.youtube.com/@ArxivPapersTikTok: https://www.tiktok.com/@arxiv_papersApple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers