RT @albarrentine: Releasing ~100GB of openly-licensed labeled data for sequence modeling research (addresses/other geographic strings from… Posted on April 8, 2017 by @thej