BARM - Official Trained Weights (ECCV 2024)

Background Adaptation with Residual Modeling for Exemplar-Free Class-Incremental Semantic Segmentation
Anqi Zhang¹, Guangyu Gao¹

¹ _{School of Computer Science and Technology, Beijing Institute of Technology, Beijing, China}

Accepted in ECCV 2024

Download

The official weights of trained models for BARM.

You can download the .pth files by using the SDK of modelscope, or simply use git.

Download by SDK

# Install ModelScope
pip install modelscope

# Download by SDK
from modelscope import snapshot_download
model_dir = snapshot_download('UltraDoughnut/BARM')

Download by Git

# Download by Git
git clone https://www.modelscope.cn/UltraDoughnut/BARM.git

Citation

@inproceedings{zhang2024background,
  title={Background Adaptation with Residual Modeling for Exemplar-Free Class-Incremental Semantic Segmentation},
  author={Zhang, Anqi and Gao, Guangyu},
  journal={ECCV},
  year={2024}
}