cross-encoder-RoBERTa-BCE

This model is a cross-encoder based on FacebookAI/roberta-base. It was trained on Ms-Marco using loss bce 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 bce

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("xpmir/cross-encoder-RoBERTa-BCE")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-RoBERTa-BCE")

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	36.25	42.75
trec2019	93.24	65.77
trec2020	87.72	63.74
fever	76.61	77.40
arguana	21.10	31.40
climate_fever	30.96	22.59
dbpedia	67.43	38.87
fiqa	45.32	37.70
hotpotqa	85.08	69.14
nfcorpus	45.98	26.00
nq	51.33	56.51
quora	72.60	75.70
scidocs	25.04	14.19
scifact	65.32	67.50
touche	59.10	34.03
trec_covid	72.75	55.71
robust04	62.09	40.66
lotte_writing	65.39	56.09
lotte_recreation	59.62	54.72
lotte_science	44.01	36.15
lotte_technology	51.81	43.00
lotte_lifestyle	72.69	62.32
Mean In Domain	72.40	57.42
BEIR 13	55.28	46.67
LoTTE (OOD)	59.27	48.82