A proposed roadmap for OpenNLP 2.0.x.

Goals

• Backwards compatible with 1.x.
• Support deep learning models trained in other framework, i.e. Tensorflow/PyTorch/MXnet.

Code Changes

• Create an opennlp-common project that contains the interfaces and common classes.
• Change opennlp-tools to have opennlp-common as a dependency.
• Create an opennlp-dl project with implementations for deep learning.

These changes are to maintain backwards compatibility and not introduce any deep learning framework dependencies into opennlp-tools.