Groma

model
ByteDance
Multimodal model specialized in localized visual grounding and reasoning.
GitHubPaper (arXiv)Project Page

Library

GitHub Repository

multimodalvision

Notes

Published at ECCV 2024.