anonymousgrey/investigating-selection-strategies-for-malware-datasets-using-a-machine-learning-approach ? reverse-engineered prompt

Build me a clean Jupyter notebook project for my malware research work that compares different dataset selection strategies using machine learning.

I want a notebook based workflow that can load a malware image dataset, prepare the data, train a few different models, and compare their results in a simple way. Please include separate notebook sections or notebooks for the main model variants I’m interested in, like DenseNet121, ResNet50, EfficientNet B0, CvT 13, and a hybrid approach. Use the same kind of pipeline across them so the results are easy to compare.

Make it easy to run step by step, with clear data loading, preprocessing, training, evaluation, and charts for accuracy and loss. If anything needs current library guidance or best practice, look up the latest docs online. Keep the code readable, research friendly, and suitable for a cyber clinic project writeup.