This is the full, unpreprocessed DIRT dataset referenced by our paper "Augmenting Decompiler Output with Learned Variable Names and Types". For more information see README.txt.