cross-encoder-ettin-17m-MarginMSE

This model is a cross-encoder based on jhu-clsp/ettin-encoder-17m. It was trained on Ms-Marco using loss marginMSE 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 marginMSE

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-MarginMSE")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-ettin-17m-MarginMSE")

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	30.19	35.85
trec2019	82.75	59.06
trec2020	84.78	58.81
fever	68.73	70.30
arguana	18.92	28.38
climate_fever	24.94	18.41
dbpedia	60.12	34.31
fiqa	37.39	30.22
hotpotqa	79.72	63.19
nfcorpus	44.37	25.27
nq	42.58	47.43
quora	78.84	80.04
scidocs	23.43	12.92
scifact	61.12	64.06
touche	59.18	31.98
trec_covid	86.01	64.39
robust04	54.84	33.24
lotte_writing	54.54	46.38
lotte_recreation	51.88	46.95
lotte_science	41.42	33.95
lotte_technology	44.98	36.89
lotte_lifestyle	66.51	57.08
Mean In Domain	65.91	51.24
BEIR 13	52.72	43.92
LoTTE (OOD)	52.36	42.42