Deep Dissect

Version: 1.0

Summary

deep_dissect is a Python library tailored for performing ablation studies on MMDetection models. It offers tools to investigate and assess the influence of different components within object detection models, with a particular focus on detection transformer models implemented in the MMDetection framework.

Features

• Ablation Study Support: Easily perform ablations by removing or modifying parts of your detection model.
• Customizable: Highly customizable to suit different experimental setups and model variations.

Installation

To install the deep_dissect package, clone the repository and install it using (cuda 10.2 is needed for this installation):

git clone https://repo.git
cd deep_dissect
python install.py

License

This project is licensed under the MIT License.