VCoder LLaVA-1.5-7b

VCoder LLaVA-1.5-7b was trained on COST training dataset in December 2023. It uses the pretrained LLaVA-1.5-7b model weights. It was introduced by Jain et al. in this repository.

VCoder is an adapter for improving existing Multimodal LLMs at object-level perception tasks with the use of perception modalities as control inputs while retaining performance on other tasks.

Citation

@article{jain2023vcoder,
    title={{VCoder: Versatile Vision Encoders for Multimodal Large Language Models}},
    author={Jitesh Jain and Jianwei Yang and Humphrey Shi},
    journal={arXiv},
    year={2023}
}