project_image

Bosch's Model Extraction Attack For Video Classification

Folder Structure
BlackBox_P1: Contains scripts for training the extracted model from SwinT victim under the Black Box setting. Refer to setup_readme within the folder for detailed instructions on running the code for training.
BlackBox_P2: Contains scripts for training the extracted model from MoViNet victim under the Black Box setting. Refer to setup_readme within the folder for detailed instructions on running the code for training.
GreyBox_P1: Contains scripts for training the extracted model from SwinT victim under the Grey Box setting. Refer to setup_readme within the folder for detailed instructions on running the code for training.
GreyBox_P2: Contains scripts for training the extracted model from MoViNet victim under the Grey Box setting. Refer to setup_readme within the folder for detailed instructions on running the code for training.
Evaluation_All - Contains Jupyter Notebooks for evaluation. Refer to eval_readme within the folder to get information on the setup and run environment.
Running Environment
Training codes were run on a single server equipped with NVIDIA V100 GPU. Evaluations codes were run on Google Colaboratory with a Pro Plus subscription.
project_image