Gte Reranker Modernbert Base

🚀 gte-reranker-modernbert-base

We're thrilled to introduce the gte-modernbert series of models, which are built on the latest modernBERT pre-trained encoder-only foundation models. The gte-modernbert series includes both text embedding models and rerank models.

These models show competitive performance in several text embedding and text retrieval evaluation tasks compared to similar-scale models from the current open-source community. This includes assessments like MTEB, LoCO, and COIR evaluations.

✨ Features

• Developed by Tongyi Lab, Alibaba Group.
• Capable of handling long input sequences with a max input length of 8192 tokens.
• Demonstrates strong performance in multiple text evaluation tasks.

📦 Installation

[!TIP] For transformers and sentence-transformers, if your GPU supports it, the efficient Flash Attention 2 will be used automatically if you have flash_attn installed. It is not mandatory.

pip install flash_attn

If you want to use sentence-transformers, you need to install it:

pip install sentence-transformers

💻 Usage Examples

Basic Usage

You can use the model with different libraries as follows:

Use with transformers

# Requires transformers>=4.48.0
import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer

model_name_or_path = "Alibaba-NLP/gte-reranker-modernbert-base"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForSequenceClassification.from_pretrained(
    model_name_or_path,
    torch_dtype=torch.float16,
)
model.eval()

pairs = [
    ["what is the capital of China?", "Beijing"],
    ["how to implement quick sort in python?", "Introduction of quick sort"],
    ["how to implement quick sort in python?", "The weather is nice today"],
]

with torch.no_grad():
    inputs = tokenizer(pairs, padding=True, truncation=True, return_tensors='pt', max_length=512)
    scores = model(**inputs, return_dict=True).logits.view(-1, ).float()
    print(scores)

# tensor([ 2.1387,  2.4609, -1.6729])

Use with sentence-transformers

# Requires transformers>=4.48.0
from sentence_transformers import CrossEncoder

model = CrossEncoder(
    "Alibaba-NLP/gte-reranker-modernbert-base",
    automodel_args={"torch_dtype": "auto"},
)

pairs = [
    ["what is the capital of China?", "Beijing"],
    ["how to implement quick sort in python?","Introduction of quick sort"],
    ["how to implement quick sort in python?", "The weather is nice today"],
]

scores = model.predict(pairs)
print(scores)
# [0.8945664  0.9213594  0.15742092]
# NOTE: Sentence Transformers calls Softmax over the outputs by default, hence the scores are in [0, 1] range.

Use with transformers.js

import {
  AutoTokenizer,
  AutoModelForSequenceClassification,
} from "@huggingface/transformers";

const model_id = "Alibaba-NLP/gte-reranker-modernbert-base";
const model = await AutoModelForSequenceClassification.from_pretrained(
  model_id,
  { dtype: "fp32" }, // Supported options: "fp32", "fp16", "q8", "q4", "q4f16"
);
const tokenizer = await AutoTokenizer.from_pretrained(model_id);

const pairs = [
  ["what is the capital of China?", "Beijing"],
  ["how to implement quick sort in python?", "Introduction of quick sort"],
  ["how to implement quick sort in python?", "The weather is nice today"],
];
const inputs = tokenizer(
  pairs.map((x) => x[0]),
  {
    text_pair: pairs.map((x) => x[1]),
    padding: true,
    truncation: true,
  },
);
const { logits } = await model(inputs);
console.log(logits.tolist()); // [[2.138258218765259], [2.4609625339508057], [-1.6775450706481934]]

📚 Documentation

Model Overview

Model list

Training Details

The gte-modernbert series of models follows the training scheme of the previous GTE models, with the only difference being that the pre-training language model base has been replaced from GTE-MLM to ModernBert. For more training details, please refer to our paper: mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval

Evaluation

MTEB

The results of other models are retrieved from MTEB leaderboard. Given that all models in the gte-modernbert series have a size of less than 1B parameters, we focused exclusively on the results of models under 1B from the MTEB leaderboard.

LoCo (Long Document Retrieval)

COIR (Code Retrieval Task)

BEIR

🔧 Technical Details

The gte-modernbert series of models follows the training scheme of the previous GTE models, with the only difference being that the pre-training language model base has been replaced from GTE-MLM to ModernBert.

📄 License

This project is licensed under the Apache-2.0 license.

Hiring

We have open positions for Research Interns and Full-Time Researchers to join our team at Tongyi Lab. We are seeking passionate individuals with expertise in representation learning, LLM-driven information retrieval, Retrieval-Augmented Generation (RAG), and agent-based systems. Our team is located in the vibrant cities of Beijing and Hangzhou. If you are driven by curiosity and eager to make a meaningful impact through your work, we would love to hear from you. Please submit your resume along with a brief introduction to dingkun.ldk@alibaba-inc.com.

Citation

If you find our paper or models helpful, feel free to give us a cite.

@inproceedings{zhang2024mgte,
  title={mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval},
  author={Zhang, Xin and Zhang, Yanzhao and Long, Dingkun and Xie, Wen and Dai, Ziqi and Tang, Jialong and Lin, Huan and Yang, Baosong and Xie, Pengjun and Huang, Fei and others},
  booktitle={Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track},
  pages={1393--1412},
  year={2024}
}

@article{li2023towards,
  title={Towards general text embeddings with multi-stage contrastive learning},
  author={Li, Zehan and Zhang, Xin and Zhang, Yanzhao and Long, Dingkun and Xie, Pengjun and Zhang, Meishan},
  journal={arXiv preprint arXiv:2308.03281},
  year={2023}
}