Linkup Research Releases SPARSEUP: A 149M-Parameter Open-Source Sparse Embedding Model

TL;DR

Linkup Research has unveiled SPARSEUP, an open-source sparse embedding model with 149 million parameters that outperforms other models in its class on the BEIR-13 benchmark, highlighting advancements in efficient large-scale language modeling.

Detailed Summary

Linkup Research has released SPARSEUP, an open-source sparse embedding model with 149 million parameters based on ModernBERT, achieving the best known result of 56.4 nDCG@10 on BEIR-13. The broader impact includes advancing the field of sparse models under 150M parameters and providing researchers and developers with a powerful tool for information retrieval tasks.

Key Points

• • SPARSEUP is an open-source sparse embedding model.
• • It has 149 million parameters and is based on ModernBERT.
• • SPARSEUP scores 56.4 nDCG@10 on BEIR-13.
• • Linkup claims it's the best public sparse encoder under 150M parameters.