Re-Implementation of SPARTA: Efficient Open-Domain Question Answering via Sparse Transformer Matching Retrieval. It is the re-implementation we used for BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models.

Also have a look at our BEIR repository: https://github.com/UKPLab/beir

Have a look at https://github.com/nreimers/beir-sparta for the training and inference code of this SPARTA model