tourkaman / S3-BERT

Semantic Search With Sentence-BERT. S-BERT, implemented and evaluated on STS, is used for building a semantic search engine.

Semantic Search With Sentence-BERT

The goal of this project is to build a semantic search engine, using S-BERT, after first implementing and evaluating it on the STS benchmark.

This is done as part of the second lab exercise in the course ID2223 Scalable Machine Learning and Deep Learning at KTH.

Progress

So far we have only implemented the regression objective of the S-BERT paper, and evaluated it on the STS benchmark, using the bert-base-uncased pre-trained BERT model from Hugging Face.