metadata
language: en
license: apache-2.0
library_name: transformers
base_model: microsoft/MiniLM-L12-H384-uncased
model_name: cross-encoder-MiniLM-L12-ADR-MSE
source: https://github.com/xpmir/cross-encoders
paper: http://arxiv.org/abs/2603.03010
tags:
- cross-encoder
- sequence-classification
- tensorboard
datasets:
- msmarco
pipeline_tag: text-classification
cross-encoder-MiniLM-L12-ADR-MSE
This model is a cross-encoder based on microsoft/MiniLM-L12-H384-uncased. It was trained on Ms-Marco using loss ADR 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
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 ADR
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-MiniLM-L12-ADR-MSE")
model = AutoModelForSequenceClassification.from_pretrained("xpmir/cross-encoder-MiniLM-L12-ADR-MSE")
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 | 37.26 | 43.81 |
| trec2019 | 96.12 | 75.26 |
| trec2020 | 94.48 | 73.98 |
| fever | 81.26 | 81.08 |
| arguana | 17.91 | 27.21 |
| climate_fever | 28.27 | 21.19 |
| dbpedia | 77.93 | 46.52 |
| fiqa | 43.97 | 36.31 |
| hotpotqa | 85.51 | 66.88 |
| nfcorpus | 56.62 | 34.20 |
| nq | 53.26 | 58.30 |
| quora | 76.30 | 78.58 |
| scidocs | 28.29 | 15.98 |
| scifact | 66.57 | 69.55 |
| touche | 64.16 | 35.13 |
| trec_covid | 87.73 | 69.22 |
| robust04 | 73.25 | 51.05 |
| lotte_writing | 66.28 | 58.18 |
| lotte_recreation | 60.32 | 54.97 |
| lotte_science | 46.27 | 38.24 |
| lotte_technology | 53.35 | 44.34 |
| lotte_lifestyle | 72.22 | 62.33 |
| Mean In Domain | 75.95 | 64.35 |
| BEIR 13 | 59.06 | 49.24 |
| LoTTE (OOD) | 61.95 | 51.52 |