cross-encoder-ettin-17m-BCE

This model is a cross-encoder based on jhu-clsp/ettin-encoder-17m. 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.

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 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-ettin-17m-BCE")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-ettin-17m-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	23.97	29.44
trec2019	69.26	50.88
trec2020	75.44	51.94
fever	56.09	58.88
arguana	13.07	19.57
climate_fever	16.67	11.92
dbpedia	45.06	23.29
fiqa	32.23	25.35
hotpotqa	67.78	51.21
nfcorpus	36.68	20.49
nq	31.84	37.01
quora	65.88	67.70
scidocs	19.01	10.16
scifact	55.57	58.18
touche	57.47	30.63
trec_covid	74.72	51.14
robust04	43.73	22.74
lotte_writing	44.62	36.79
lotte_recreation	43.38	39.14
lotte_science	34.20	28.39
lotte_technology	31.35	27.11
lotte_lifestyle	56.02	46.79
Mean In Domain	56.22	44.09
BEIR 13	44.01	35.81
LoTTE (OOD)	42.22	33.49