cross-encoder-RoBERTa-DistillRankNET

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

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("FacebookAI/roberta-base")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-RoBERTa-DistillRankNET")

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	35.81	42.34
trec2019	92.13	72.11
trec2020	95.43	73.13
fever	79.50	79.69
arguana	17.94	26.98
climate_fever	30.81	22.50
dbpedia	76.74	46.32
fiqa	47.91	39.47
hotpotqa	85.40	68.07
nfcorpus	55.25	33.79
nq	53.82	58.72
quora	79.55	81.33
scidocs	27.23	15.38
scifact	65.37	68.74
touche	62.74	34.90
trec_covid	83.20	65.28
robust04	71.57	48.06
lotte_writing	68.54	58.58
lotte_recreation	60.93	55.18
lotte_science	46.92	38.96
lotte_technology	52.04	43.24
lotte_lifestyle	72.52	62.88
Mean In Domain	74.46	62.53
BEIR 13	58.88	49.32
LoTTE (OOD)	62.09	51.15