Too Long; Didn't Read
We present how to improve AllenNLP’s coreference resolution model in order to achieve a more coherent output. We also introduce a couple of ensemble strategies to make use of both Huggingface and Allen NLP models at the same time. All code from the next chapter is available on NeuroSYS GitHub repository. We focus mainly on improving how the libraries resolve found clusters. We’ve decided to implement several minor changes that result in a significant improvement of the final text. We propose the following solutions: Do not consider a cluster at all if it doesn’t contain any noun phrases.