Monoelectra Large

🚀 Cross-Encoder for Text Ranking

This model serves as a port of the webis/monoelectra-large model from lightning-ir to Sentence Transformers and Transformers. It can be used as a reranker in a 2 - stage "retrieve - rerank" pipeline, enhancing the efficiency of information retrieval.

✨ Features

• Ported from lightning-ir to Sentence Transformers and Transformers.
• Can be used as a reranker in a "retrieve - rerank" pipeline.
• Based on the research in the paper A Systematic Investigation of Distilling Large Language Models into Cross - Encoders for Passage Re - ranking.

📦 Installation

If you want to use this model with Sentence Transformers, you need to install the sentence-transformers library:

pip install sentence-transformers

💻 Usage Examples

Basic Usage with Sentence Transformers

from sentence_transformers import CrossEncoder

model = CrossEncoder("cross-encoder/monoelectra-large", trust_remote_code=True)
scores = model.predict([
    ("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
    ("How many people live in Berlin?", "Berlin is well known for its museums."),
])
print(scores)
# [ 6.016401  -3.6922567]

Basic Usage with Transformers

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model = AutoModelForSequenceClassification.from_pretrained("cross-encoder/monoelectra-large", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("cross-encoder/monoelectra-large")

features = tokenizer(
    [
        ("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
        ("How many people live in Berlin?", "Berlin is well known for its museums."),
    ],
    padding=True,
    truncation=True,
    return_tensors="pt",
)

model.eval()
with torch.no_grad():
    scores = model(**features).logits.view(-1)
print(scores)
# tensor([ 6.0164, -3.6923])

📚 Documentation

The original model was introduced in the paper A Systematic Investigation of Distilling Large Language Models into Cross - Encoders for Passage Re - ranking. You can find the code used to train the original model at https://github.com/webis-de/rank - distillm.

For more details on using the model as a reranker in a 2 - stage "retrieve - rerank" pipeline, see SBERT.net Retrieve & Re - rank.

📄 License

This project is licensed under the Apache 2.0 license.