PyTorchCRF is an implementation of Conditional Random Field using the PyTorch library
This is a work-in-progress repository to develop a stand-alone lightweight CRF Layer in PyTorch. The base code is factored out of AllenNLP library's CRF implementation.
We aren't using AllenNLP since we find it to be bloated and hence cumbersome to install due to the innumerable dependencies it come with as of October, 2018.
You can use pip to install this module.