Distinguishing between high entropy bit streams

This dataset contains the curated files, classified by type and extension so that other researchers can compute their features and replicate outcomes. 

A total of 5 datasets (i.e., one according to each file size denoted as 64, 128, 256, 512, and 1024) each one consisting of exactly 50% encrypted and 50% compressed files.

-The encryption algorithms used to generate the files were:

AES(128 / 192 / 256) and Camelia(128 / 192 / 256

-In the case of compressed files:

ZIP RAR BZIP2 GZIP

-Different source files were considered to generate the encrypted and compressed files. 

COCO Dataset (http://cocodataset.org/home) 
Microsoft Research (https://www.microsoft.com/en-us/research/project/rgb-d-dataset-7-scenes/)
ArXiv (https://arxiv.org/)
Project Gutenberg (https://www.gutenberg.org/)
Several classical music symphonies in MP3 format 
YouTube-8M dataset
Binaries extracted from system32 in Win10 x64 and sbin from Ubuntu 16.04