• This is an image classification system for HW1 of Security and Privacy of Machine Learning.

• Five adversarially trained models are provided for evaluating images uploaded by users.

  • Students need to generate adversarial examples (images) with enough attacking strength to make pre-trained models mis-classify.
  • Each submission is in .zip format containing no more than 500 images.
  • Each student has only 10 submissions per day.

• An csv file of the classification results on each submitted image will be provided after each submission.

  

Log In	Create an account	Authorized page

Password reset	Set new password	Password change

    git clone https://github.com/shiannn/SPML-query-system.git
    cd SPML-query-system

    virtualenv -p <Path to Your Python> env
    source env/bin/activate

    python -m pip install -r requirements.txt

    python source/manage.py migrate

    python source/manage.py collectstatic

    python source/manage.py runserver

The codebase of the registration and login is highly borrowed from Yehor Smoliakov

• https://github.com/egorsmkv/simple-django-login-and-register