cross-encoder-bert-base-Hinge

This model is a cross-encoder based on bert-base-uncased. 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.

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("bert-base-uncased")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-bert-base-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	38.05	44.38
trec2019	97.09	73.12
trec2020	93.52	70.83
fever	79.83	79.96
arguana	22.79	33.87
climate_fever	33.40	24.97
dbpedia	73.46	43.13
fiqa	42.20	34.97
hotpotqa	88.41	72.40
nfcorpus	54.88	33.45
nq	51.22	56.13
quora	77.95	79.84
scidocs	26.64	15.26
scifact	66.32	69.18
touche	59.68	32.27
trec_covid	89.62	63.75
robust04	71.38	46.18
lotte_writing	64.17	55.35
lotte_recreation	59.69	54.46
lotte_science	42.39	35.07
lotte_technology	51.80	42.35
lotte_lifestyle	69.76	60.15
Mean In Domain	76.22	62.78
BEIR 13	58.95	49.17
LoTTE (OOD)	59.86	48.93