Hi all, The training parameters like "cutoff" are case-sensitive when training models in the parameters file, so cutoff != CUTOFF != Cutoff.
I made a PR to make them not be case sensitive and if anybody has any comments please check out the PR: https://github.com/apache/opennlp/pull/423 The idea was to try to remove some of the "gotchas" when training a model that a new user might run into. Thanks, Jeff
