km1994/NLP-Interview-Notes

Organized hierarchical Q&A reference covering HMM, CRF, BiLSTM-CRF, and deep learning-based sequence labeling architectures for NER; word embeddings (TF-IDF, Word2Vec, FastText, ELMo); and information extraction tasks including relation and event extraction. Structures progressive difficulty from foundational probability models through modern transformer-based approaches, with practical implementation tricks for handling Chinese NER, nested entities, and low-resource scenarios.

2,581 stars. No commits in the last 6 months.