Description
NLP machine-learning components can be built with Thinc as a compact neural network library. Developers and researchers can define models, layers, optimizers, and training loops used by spaCy-style workflows. It may process sensitive text or model data supplied by applications.