Glove and Elmo embeddings are compared for solving NER task. Using similar model architecure and different embeddings types (Glove and Elmo), NER task was sloved.

Both models (with Glove and Elmo embeddings) achieved accuracy of 0.97 on test set.

https://www.kaggle.com/datasets/c4pt41n2004/named-entity-recognition-ner/data

https://www.kaggle.com/datasets/takuok/glove840b300dtxt

https://drive.google.com/file/d/1scqJxGrRi4VrGX5B1GdyyzZXDc4mPhRB/view?usp=sharing https://s3-us-west-2.amazonaws.com/allennlp/models/elmo/2x4096_512_2048cnn_2xhighway/elmo_2x4096_512_2048cnn_2xhighway_weights.hdf5