cross-encoder-RoBERTa-infoNCE

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

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

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	38.91	45.72
trec2019	95.35	73.74
trec2020	93.21	72.00
fever	77.82	78.27
arguana	21.78	32.36
climate_fever	26.02	19.42
dbpedia	75.34	44.45
fiqa	48.16	40.59
hotpotqa	86.88	70.72
nfcorpus	54.94	33.58
nq	54.68	59.47
quora	75.73	78.56
scidocs	27.99	15.66
scifact	68.45	71.15
touche	59.24	34.76
trec_covid	91.07	71.85
robust04	70.84	48.48
lotte_writing	70.65	61.44
lotte_recreation	61.61	56.37
lotte_science	46.90	39.01
lotte_technology	56.38	47.20
lotte_lifestyle	74.03	64.38
Mean In Domain	75.82	63.82
BEIR 13	59.08	50.06
LoTTE (OOD)	63.40	52.81