Codesota · Models · MonoT5-3BCastorini (Waterloo)1 results · 1 benchmarks
Model card

MonoT5-3B.

Castorini (Waterloo)open-source3B paramsT5-3B (seq2seq reranker)

Document Ranking with Pretrained Seq2Seq Model. EMNLP 2020 Findings.

§ 01 · Card

Model card,
inline.

Rendered server-side from the upstream README on Hugging Face — same content as the source repo, with editorial typography. The full card, sample weights, and revision history live on HF.

Source
castorini/monot5-3b-msmarco-10k

This model is a T5-3B reranker fine-tuned on the MS MARCO passage dataset for 10k steps (or 1 epoch).

For more details on how to use it, check pygaggle.ai

Paper describing the model: Document Ranking with a Pretrained Sequence-to-Sequence Model

This model is also the state of the art on the BEIR Benchmark.

Card content reproduced from huggingface.co/castorini/monot5-3b-msmarco-10k under the upstream license. Rendering trims fenced HTML, raw widgets and tables for safety; tap the link for the untouched original.
§ 02 · Benchmarks

Every benchmark MonoT5-3B has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01MS MARCONatural Language Processing · Text Rankingmrr@1039.0%#4/42020-11-01source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area

Where MonoT5-3B actually performs.

Natural Language Processing
1
benchmark
avg rank #4.0
§ 04 · Papers

1 paper with results for MonoT5-3B.

  1. 2020-11-01· Natural Language Processing· 1 result

    Document Ranking with a Pretrained Sequence-to-Sequence Model

§ 05 · Related models

Other Castorini (Waterloo) models scored on Codesota.

RankLLaMA-7B
7B params · 0 results
§ 06 · Sources & freshness

Where these numbers come from.

arxiv
1
result
1 of 1 rows marked verified.