HanLP

Why we included this project

HanLP covers the full Chinese text processing pipeline: tokenization, part-of-speech tagging, named entity recognition, dependency and constituent parsing, semantic role labeling, and text classification, all through a single Python API. It ships dozens of pretrained models trained on a large multilingual corpus, so teams can get accurate results in 130 languages without training their own. The library is built on PyTorch and TensorFlow 2.x and supports both multi-task and single-task models, which makes it practical to embed into real applications. Developers building Chinese-language products, from search to content analysis to document processing, will find it a dependable base, and researchers get a wide range of tasks and annotation standards to work with.