Download
InferSent is a supervised sentence embedding method that learns universal representations from Natural Language Inference data and transfers well to many downstream tasks. It uses a BiLSTM encoder with max-pooling to produce fixed-length sentence vectors that capture semantics beyond bag-of-words statistics. Trained on large NLI datasets, the embeddings generalize across tasks like sentiment analysis, entailment, paraphrase detection, and semantic similarity with simple linear classifiers. The repository provides pretrained vectors, training scripts, and clear examples for evaluating transfer on a wide suite of benchmarks. Because the encoder is compact and language-agnostic at the interface level, it’s easy to drop into production pipelines that need robust semantic features. InferSent helped popularize the idea that supervised objectives (like NLI) can yield strong general-purpose sentence encoders, and it remains a reliable baseline against which to compare newer models.
Features
- BiLSTM-max sentence encoder producing fixed-size embeddings
- Pretrained vectors ready for transfer learning across NLP tasks
- Simple evaluation recipes for classification and similarity benchmarks
- Training scripts and datasets for reproducing and fine-tuning
- Language-agnostic integration via plain sentence-to-vector APIs
- Strong baselines that remain competitive for lightweight deployments
Project Samples
