cross-encoder-ettin-32m-Hinge

This model is a cross-encoder based on jhu-clsp/ettin-encoder-32m. It was trained on Ms-Marco using loss hingeLoss as part of a reproducibility paper for training cross encoders: "Reproducing and Comparing Distillation Techniques for Cross-Encoders", see the paper for more details.

Contents

• Model Description
• Usage
• Evals

Model Description

This model is intended for re-ranking the top results returned by a retrieval system (like BM25, Bi-Encoders or SPLADE).

• Training Data: MS MARCO Passage
• Language: English
• Loss hingeLoss

Training can be easily reproduced using the assiciated repository. The exact training configuration used for this model is also detailed in config.yaml.

Usage

Quick Start:

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("jhu-clsp/ettin-encoder-32m")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-ettin-32m-Hinge")

features = tokenizer("What is experimaestro ?", "Experimaestro is a powerful framework for ML experiments management...", padding=True, truncation=True, return_tensors="pt")

model.eval()
with torch.no_grad():
    scores = model(**features).logits
    print(scores)

Evaluations

We provide evaluations of this cross-encoder re-ranking the top 1000 documents retrieved by naver/splade-v3-distilbert.

dataset	RR@10	nDCG@10
msmarco_dev	33.98	40.04
trec2019	91.81	66.69
trec2020	90.39	65.52
fever	69.95	70.72
arguana	12.21	18.58
climate_fever	20.13	14.67
dbpedia	60.87	34.71
fiqa	39.77	31.69
hotpotqa	79.42	62.45
nfcorpus	51.64	30.62
nq	42.88	47.91
quora	76.94	78.36
scidocs	24.01	13.22
scifact	61.83	64.79
touche	65.31	32.77
trec_covid	88.00	67.20
robust04	61.43	37.53
lotte_writing	65.38	55.76
lotte_recreation	54.31	49.21
lotte_science	44.43	36.64
lotte_technology	49.41	41.33
lotte_lifestyle	65.28	56.09
Mean In Domain	72.06	57.42
BEIR 13	53.30	43.67
LoTTE (OOD)	56.71	46.09