How can I custom train the model "ner_ontonotes_bert_mult". I want to train the model with my own dataset which have some different tags. How to train the model and use it?
When I am training my model it accepting only 40 data and after that training stops automatically. How to resolve this.
There are several issues covering your question in DeepPavlov repository:
You need to change
NER_PATHin configuration file such that vocabulary (set of tags) is not re-used and model parameters from Ontonotes are not re-used also (but still model will use pre-trained weights for BERT model).