Data
The .bytes file were extracted from the data. There are 10868 of them. Those were divided into 6 parts, each 2000 files except the last one 868 The names of those parts are bytes_names.1,.2,.3,.4,.5,.6 The actual files are in 7z format bytes.1.7z,....
CSV files
the names of the files were combined with the corresonding classes extracted from the trainLabels.csv to obtain the data.1.csv,....
Resized images
The images were resized using the resized-data.py and stored as follows
the 9000 images from 1,2,3,5 and the second half of 4 are stored as training data the 1868 images from 6 and the first part of 4 were stored as test data. Both train and test were compressed into the resized.7z file
Code
- keras_cifar10.ipynb example of image classification using CNN
- keras_cifar10-generator.ipynb the same as the above but the images are loaded on the fly
- malware-resized-yuan.ipynb malware classification of resized images which are assume to be on local disk
- malware-yuan-kaggle.ipynb same as the above but the resized.7z is download from the google drive
- malware-yuan.ipynb resizes images on the fly
- malware.ipynb old attempt