pytextclassifier

Why we included this project

pytextclassifier is a pragmatic option if you want to add text classification to an application without building a full ML pipeline. It wraps a wide set of algorithms in one uniform API, from scikit-learn-style baselines like logistic regression, SVM, and Naive Bayes to deep models such as TextCNN, TextRNN, FastText, and BERT, with MiniBatchKMeans available for clustering. That lets you start with a fast classical baseline and step up to a neural model as your corpus grows, all through the same train, predict, and evaluate workflow. It handles binary, multi-class, multi-label, and hierarchical tasks at both sentence and document level, and works with English and Chinese text, so it fits sentiment analysis, topic labeling, or risk classification. The example-driven docs even show feature-importance visualization, which helps when you want to see what a trained model actually keys on.