DocentSzachista / master_degree_project

Master degree thesis

How to launch it?

1. Create virtual environment

    python3 -m venv venv

2. Activate it

    source venv/bin/activate 

3. Install requirements.txt

    pip3 install -r requirments.txt

4. Launch script

    python3 main.py 

Config.json file

Required options:

• "model": name of the model to be processed.

• "model_g_drive": Id of file on Google drive that contains model to be downloaded,

• "augumentations": list of image augmentations to be applied. Currently supported are noise and mixup augumentations

Required fields for augumentation:
    - name: name of the augumentation
    - start_point: starting amount of pixels affected by augumentation
    - finish_point: end amount of pixels affected by augumentation
    - step: how many pixels affected increases

• "dataset": dataset name that we want to work on. For now only supported is "CIFAR10"
• "image_dim": Image dimensions Argument specific for mixup augumentation - "chosen_image": path to image to blend in every image in dataset.

Optional fields:

• "apply_random_on_each_image": for each image generate a new mask
• "chosen_color_chanels": color channels that should be noised, default value: "RGB"
• "save_preprocessing": Save preprocessed images. Defaults to false