Multilingual E5 Language Detection

🚀 多语言检测模型

本模型支持检测45种语言，它基于 multilingual-e5-base 模型在 common-language 数据集上进行微调。整体准确率达到了98.37%，更多评估结果如下所示。

🚀 快速开始

📦 下载模型

from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained('Mike0307/multilingual-e5-language-detection')
model = AutoModelForSequenceClassification.from_pretrained('Mike0307/multilingual-e5-language-detection', num_labels=45)

💻 使用示例

基础用法

import torch

languages = [
    "Arabic", "Basque", "Breton", "Catalan", "Chinese_China", "Chinese_Hongkong", 
    "Chinese_Taiwan", "Chuvash", "Czech", "Dhivehi", "Dutch", "English", 
    "Esperanto", "Estonian", "French", "Frisian", "Georgian", "German", "Greek", 
    "Hakha_Chin", "Indonesian", "Interlingua", "Italian", "Japanese", "Kabyle", 
    "Kinyarwanda", "Kyrgyz", "Latvian", "Maltese", "Mongolian", "Persian", "Polish", 
    "Portuguese", "Romanian", "Romansh_Sursilvan", "Russian", "Sakha", "Slovenian", 
    "Spanish", "Swedish", "Tamil", "Tatar", "Turkish", "Ukranian", "Welsh"
]

def predict(text, model, tokenizer, device = torch.device('cpu')):
    model.to(device)
    model.eval()
    tokenized = tokenizer(text, padding='max_length', truncation=True, max_length=128, return_tensors="pt")
    input_ids = tokenized['input_ids']
    attention_mask = tokenized['attention_mask']
    with torch.no_grad():
        input_ids = input_ids.to(device)
        attention_mask = attention_mask.to(device)
        outputs = model(input_ids=input_ids, attention_mask=attention_mask)
    logits = outputs.logits
    probabilities = torch.nn.functional.softmax(logits, dim=1)
    return probabilities

def get_topk(probabilities, languages, k=3):
    topk_prob, topk_indices = torch.topk(probabilities, k)
    topk_prob = topk_prob.cpu().numpy()[0].tolist()
    topk_indices = topk_indices.cpu().numpy()[0].tolist()
    topk_labels = [languages[index] for index in topk_indices]
    return topk_prob, topk_labels

text = "你的測試句子"
probabilities = predict(text, model, tokenizer)
topk_prob, topk_labels = get_topk(probabilities, languages)
print(topk_prob, topk_labels)

# [0.999620258808, 0.00025940246996469, 2.7690215574693e-05]
# ['Chinese_Taiwan', 'Chinese_Hongkong', 'Chinese_China']

📚 详细文档

评估结果

测试数据集采用的是 common_language 测试数据集。

📄 许可证

本项目采用 Apache-2.0 许可证。

模型信息

支持语言

本模型支持以下45种语言的检测：

• Arabic、Basque、Breton、Catalan、Chinese_China、Chinese_Hongkong、Chinese_Taiwan、Chuvash、Czech、Dhivehi、Dutch、English、Esperanto、Estonian、French、Frisian、Georgian、German、Greek、Hakha_Chin、Indonesian、Interlingua、Italian、Japanese、Kabyle、Kinyarwanda、Kyrgyz、Latvian、Maltese、Mongolian、Persian、Polish、Portuguese、Romanian、Romansh_Sursilvan、Russian、Sakha、Slovenian、Spanish、Swedish、Tamil、Tatar、Turkish、Ukranian、Welsh

评估指标

本模型使用以下评估指标：

• 精确率（precision）
• 召回率（recall）
• F1分数（f1）
• 准确率（accuracy）

标签

本模型相关标签：

• language-detection
• Frisian
• Dhivehi
• Hakha_Chin
• Kabyle
• Sakha