Reranker ModernBERT Large Gooaq Bce

🚀 在GooAQ上訓練的ModernBERT-large模型

這是一個基於answerdotai/ModernBERT-large，使用sentence-transformers庫微調得到的交叉編碼器模型。它可以計算文本對的得分，可用於文本重排序和語義搜索。

訓練腳本請參考training_gooaq_bce.py，該腳本僅將基礎模型從answerdotai/ModernBERT-base更新為answerdotai/ModernBERT-large。此腳本也在交叉編碼器>訓練概述文檔和使用Sentence Transformers v4訓練和微調重排序模型博客文章中有所描述。

🚀 快速開始

直接使用（Sentence Transformers）

首先安裝Sentence Transformers庫：

pip install -U sentence-transformers

然後，你可以加載此模型並進行推理：

from sentence_transformers import CrossEncoder

# 從🤗 Hub下載
model = CrossEncoder("tomaarsen/reranker-ModernBERT-large-gooaq-bce")
# 獲取文本對的得分
pairs = [
    ['what are the characteristics and elements of poetry?', 'The elements of poetry include meter, rhyme, form, sound, and rhythm (timing). Different poets use these elements in many different ways.'],
    ['what are the characteristics and elements of poetry?', "What's the first rule of writing poetry? That there are no rules — it's all up to you! Of course there are different poetic forms and devices, and free verse poems are one of the many poetic styles; they have no structure when it comes to format or even rhyming."],
    ['what are the characteristics and elements of poetry?', "['Blank verse. Blank verse is poetry written with a precise meter—almost always iambic pentameter—that does not rhyme. ... ', 'Rhymed poetry. In contrast to blank verse, rhymed poems rhyme by definition, although their scheme varies. ... ', 'Free verse. ... ', 'Epics. ... ', 'Narrative poetry. ... ', 'Haiku. ... ', 'Pastoral poetry. ... ', 'Sonnet.']"],
    ['what are the characteristics and elements of poetry?', 'The main component of poetry is its meter (the regular pattern of strong and weak stress). When a poem has a recognizable but varying pattern of stressed and unstressed syllables, the poetry is written in verse. ... There are many possible patterns of verse, and the basic pattern of each unit is called a foot.'],
    ['what are the characteristics and elements of poetry?', "Some poetry may not make sense to you. But that's because poets don't write to be understood by others. They write because they must. The feelings and emotions that reside within them need to be expressed."],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# 或者根據與單個文本的相似度對不同文本進行排序
ranks = model.rank(
    'what are the characteristics and elements of poetry?',
    [
        'The elements of poetry include meter, rhyme, form, sound, and rhythm (timing). Different poets use these elements in many different ways.',
        "What's the first rule of writing poetry? That there are no rules — it's all up to you! Of course there are different poetic forms and devices, and free verse poems are one of the many poetic styles; they have no structure when it comes to format or even rhyming.",
        "['Blank verse. Blank verse is poetry written with a precise meter—almost always iambic pentameter—that does not rhyme. ... ', 'Rhymed poetry. In contrast to blank verse, rhymed poems rhyme by definition, although their scheme varies. ... ', 'Free verse. ... ', 'Epics. ... ', 'Narrative poetry. ... ', 'Haiku. ... ', 'Pastoral poetry. ... ', 'Sonnet.']",
        'The main component of poetry is its meter (the regular pattern of strong and weak stress). When a poem has a recognizable but varying pattern of stressed and unstressed syllables, the poetry is written in verse. ... There are many possible patterns of verse, and the basic pattern of each unit is called a foot.',
        "Some poetry may not make sense to you. But that's because poets don't write to be understood by others. They write because they must. The feelings and emotions that reside within them need to be expressed.",
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

✨ 主要特性

• 基於answerdotai/ModernBERT-large模型微調，具備強大的文本理解能力。
• 作為交叉編碼器，能夠計算文本對的得分，適用於文本重排序和語義搜索任務。
• 支持長序列輸入，最大序列長度可達8192個標記。

📦 安裝指南

安裝Sentence Transformers庫：

pip install -U sentence-transformers

💻 使用示例

基礎用法

from sentence_transformers import CrossEncoder

# 從🤗 Hub下載
model = CrossEncoder("tomaarsen/reranker-ModernBERT-large-gooaq-bce")
# 獲取文本對的得分
pairs = [
    ['what are the characteristics and elements of poetry?', 'The elements of poetry include meter, rhyme, form, sound, and rhythm (timing). Different poets use these elements in many different ways.'],
    ['what are the characteristics and elements of poetry?', "What's the first rule of writing poetry? That there are no rules — it's all up to you! Of course there are different poetic forms and devices, and free verse poems are one of the many poetic styles; they have no structure when it comes to format or even rhyming."],
    ['what are the characteristics and elements of poetry?', "['Blank verse. Blank verse is poetry written with a precise meter—almost always iambic pentameter—that does not rhyme. ... ', 'Rhymed poetry. In contrast to blank verse, rhymed poems rhyme by definition, although their scheme varies. ... ', 'Free verse. ... ', 'Epics. ... ', 'Narrative poetry. ... ', 'Haiku. ... ', 'Pastoral poetry. ... ', 'Sonnet.']"],
    ['what are the characteristics and elements of poetry?', 'The main component of poetry is its meter (the regular pattern of strong and weak stress). When a poem has a recognizable but varying pattern of stressed and unstressed syllables, the poetry is written in verse. ... There are many possible patterns of verse, and the basic pattern of each unit is called a foot.'],
    ['what are the characteristics and elements of poetry?', "Some poetry may not make sense to you. But that's because poets don't write to be understood by others. They write because they must. The feelings and emotions that reside within them need to be expressed."],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

高級用法

# 根據與單個文本的相似度對不同文本進行排序
ranks = model.rank(
    'what are the characteristics and elements of poetry?',
    [
        'The elements of poetry include meter, rhyme, form, sound, and rhythm (timing). Different poets use these elements in many different ways.',
        "What's the first rule of writing poetry? That there are no rules — it's all up to you! Of course there are different poetic forms and devices, and free verse poems are one of the many poetic styles; they have no structure when it comes to format or even rhyming.",
        "['Blank verse. Blank verse is poetry written with a precise meter—almost always iambic pentameter—that does not rhyme. ... ', 'Rhymed poetry. In contrast to blank verse, rhymed poems rhyme by definition, although their scheme varies. ... ', 'Free verse. ... ', 'Epics. ... ', 'Narrative poetry. ... ', 'Haiku. ... ', 'Pastoral poetry. ... ', 'Sonnet.']",
        'The main component of poetry is its meter (the regular pattern of strong and weak stress). When a poem has a recognizable but varying pattern of stressed and unstressed syllables, the poetry is written in verse. ... There are many possible patterns of verse, and the basic pattern of each unit is called a foot.',
        "Some poetry may not make sense to you. But that's because poets don't write to be understood by others. They write because they must. The feelings and emotions that reside within them need to be expressed.",
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

📚 詳細文檔

模型詳情

模型描述

模型來源

• 文檔：Sentence Transformers文檔
• 文檔：交叉編碼器文檔
• 倉庫：GitHub上的Sentence Transformers
• Hugging Face：Hugging Face上的交叉編碼器

評估

指標

交叉編碼器重排序（gooaq-dev數據集，always_rerank_positives=False）

使用CrossEncoderRerankingEvaluator進行評估，參數如下：

{
    "at_k": 10,
    "always_rerank_positives": false
}

交叉編碼器重排序（gooaq-dev數據集，always_rerank_positives=True）

使用CrossEncoderRerankingEvaluator進行評估，參數如下：

{
    "at_k": 10,
    "always_rerank_positives": true
}

交叉編碼器重排序（NanoMSMARCO_R100、NanoNFCorpus_R100和NanoNQ_R100數據集，always_rerank_positives=True）

使用CrossEncoderRerankingEvaluator進行評估，參數如下：

{
    "at_k": 10,
    "always_rerank_positives": true
}

交叉編碼器Nano BEIR（NanoBEIR_R100_mean數據集）

使用CrossEncoderNanoBEIREvaluator進行評估，參數如下：

{
    "dataset_names": [
        "msmarco",
        "nfcorpus",
        "nq"
    ],
    "rerank_k": 100,
    "at_k": 10,
    "always_rerank_positives": true
}

訓練詳情

訓練數據集

未命名數據集

• 大小：578,402個訓練樣本
• 列：question、answer和label
• 基於前1000個樣本的近似統計信息： | | 問題 | 答案 | 標籤 | |------|------|------|------| | 類型 | 字符串 | 字符串 | 整數 | | 詳情 |
  • 最小：22個字符
  • 平均：43.99個字符
  • 最大：93個字符
  |
  • 最小：51個字符
  • 平均：252.75個字符
  • 最大：378個字符
  |
  • 0：約82.30%
  • 1：約17.70%
  |
• 樣本： | 問題 | 答案 | 標籤 | |------|------|------| | what are the characteristics and elements of poetry? | The elements of poetry include meter, rhyme, form, sound, and rhythm (timing). Different poets use these elements in many different ways. | 1 | | what are the characteristics and elements of poetry? | What's the first rule of writing poetry? That there are no rules — it's all up to you! Of course there are different poetic forms and devices, and free verse poems are one of the many poetic styles; they have no structure when it comes to format or even rhyming. | 0 | | what are the characteristics and elements of poetry? | ['Blank verse. Blank verse is poetry written with a precise meter—almost always iambic pentameter—that does not rhyme. ... ', 'Rhymed poetry. In contrast to blank verse, rhymed poems rhyme by definition, although their scheme varies. ... ', 'Free verse. ... ', 'Epics. ... ', 'Narrative poetry. ... ', 'Haiku. ... ', 'Pastoral poetry. ... ', 'Sonnet.'] | 0 |
• 損失函數：BinaryCrossEntropyLoss，參數如下：

{
    "activation_fn": "torch.nn.modules.linear.Identity",
    "pos_weight": 5
}

訓練超參數

非默認超參數

• eval_strategy：steps
• per_device_train_batch_size：64
• per_device_eval_batch_size：64
• learning_rate：2e-05
• num_train_epochs：1
• warmup_ratio：0.1
• seed：12
• bf16：True
• dataloader_num_workers：4
• load_best_model_at_end：True

所有超參數

訓練日誌

加粗行表示保存的檢查點。

框架版本

• Python：3.11.10
• Sentence Transformers：3.5.0.dev0
• Transformers：4.49.0
• PyTorch：2.5.1+cu124
• Accelerate：1.5.2
• Datasets：2.21.0
• Tokenizers：0.21.0

🔧 技術細節

該模型基於answerdotai/ModernBERT-large進行微調，使用BinaryCrossEntropyLoss作為損失函數，在訓練過程中採用了一系列超參數進行優化，如學習率、批次大小、訓練輪數等。通過在特定數據集上的訓練，模型能夠學習到文本對之間的語義關係，從而實現文本重排序和語義搜索的功能。

📄 許可證

本模型使用Apache-2.0許可證。

📖 引用

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}