Pinned post
This work (/w @debforit & @macavaney) simplifies large-scale experiments with IR benchmarks; for instance, a recent work I am preparing reduced persistent disk usage from 63TB to 90 GB. Check out the repo and the demo at #ECIR2026! github.com/Parry-Parry/suβ¦
SuiteEval: Simplifying Retrieval BenchmarksView quoted post on X
@MrParryParry et al. present a framework for standardized IR evaluation with automatic e2e benchmarking, dynamic indexing, and built-in support for BEIR, LoTTE, MS MARCO, etc.
π arxiv.org/abs/2602.18107
π¨π½βπ» github.com/Parry-Parry/suβ¦