[paper] [project page]

We introduce Instruction-guided Visual Masking (IVM), a new versatile visual grounding model that is compatible with diverse multimodal models, such as LMM and robot model. By constructing visual masks for instruction-irrelevant regions, IVM-enhanced multimodal models can effectively focus on task-relevant image regions to better align with complex instructions. Specifically, we design a visual masking data generation pipeline and create an IVM-Mix-1M dataset with 1 million image-instruction pairs. We further introduce a new learning technique, Discriminator Weighted Supervised Learning (DWSL) for preferential IVM training that prioritizes high-quality data samples. Experimental results on generic multimodal tasks such as VQA and embodied robotic control demonstrate the versatility of IVM, which as a plug-and-play tool, significantly boosts the performance of diverse multimodal models.

Duck on green plate | Red cup on red plate | Red cup on red plate | Red cup on silver pan | Red cup on silver pan

• Quick Start
• Model Zoo
• Datasets

1. Clone this repository and navigate to IVM folder

git clone https://github.com/2toinf/IVM.git
cd IVM

2. Install Package

conda create -n IVM
pip install -e .

from IVM import load
from PIL import Image

model = load(ckpt_path, type="lisa", low_gpu_memory = False) # Set `low_gpu_memory=True` if you don't have enough GPU Memory

image = Image.open("assets/demo.jpg")
instruction = "pick up the red cup"

result = model(image, instruction, threshold = 0.5)
'''
result content:
	    'soft': heatmap
            'hard': segmentation map
            'blur_image': rbg
            'highlight_image': rbg
            'cropped_blur_img': rgb
            'cropped_highlight_img': rgb
            'alpha_image': rgba
'''

Coming Soon

Coming Soon

Policy Learning: https://github.com/Facebear-ljx/BearRobot

Robot Infrastructure: https://github.com/rail-berkeley/bridge_data_robot

Coming Soon

This work is built upon the LLaVA and SAM. And we borrow ideas from LISA

@misc{zheng2024instructionguided,
            title={Instruction-Guided Visual Masking}, 
            author={Jinliang Zheng and Jianxiong Li and Sijie Cheng and Yinan Zheng and Jiaming Li and Jihao Liu and Yu Liu and Jingjing Liu and Xianyuan Zhan},
            year={2024},
            eprint={2405.19783},
            archivePrefix={arXiv},
            primaryClass={cs.CV}
        }