This repository implements a BERT-BiLSTM-CRF architecture for sequence labeling in Named Entity Recognition tasks. It uses BERT as a feature extractor, BiLSTM to capture sequential dependencies, and a CRF layer to enforce valid tag sequences. The model processes text in BIO format and identifies entity spans across categories like names, addresses, companies, and organizations. Evaluation results on the CLUENER2020 Chinese dataset are provided in the README.