This repository contains the code for the paper MemoryMamba: Memory-Augmented State Space Model for Defect Recognition by Qianning Wang, He Hu, and Yucheng Zhou.

MemoryMamba is a state-of-the-art framework for defect recognition using a state space model. This repository provides the implementation details and the necessary scripts to reproduce the results presented in the paper.

To install the necessary dependencies, please use the following command:

pip install -r requirements.txt

To run the model, use the following command:

python simple_memorymamba.py

The dataset used for training and evaluation is provided in Huggingface.

If you find this code useful, please consider citing our work:

@article{wang2024memorymamba,
  title={MemoryMamba: Memory-Augmented State Space Model for Defect Recognition},
  author={Wang, Qianning and Hu, He and Zhou, Yucheng},
  journal={arXiv preprint arXiv:2405.03673},
  year={2024}
}

VMamba: the codebase we built upon.