研究团队发布塔吉克语嵌入模型Saidzoda-Tajik-E5-Large
Start here
2026年11月3日,研究团队在arXiv发布了塔吉克语文本嵌入模型Saidzoda-Tajik-E5-Large。该模型基于560M参数的multilingual-e5-large-instruct进行微调,并在235,140个问答对上完成了训练。实验结果显示,该模型在未见过的法律和税务文档检索任务中,top-10命中率分别提升至93.6%和94.5%;在同风格测试中,其表现追平了参数规模为其5倍的Giga-Embeddings-3B,并且保留了对俄语与英语的处理能力。
AI-generated from reports · Updated 1 day ago
Report timeline
Follow the reports to see different sides of the story.
- arXiv 自然语言处理Saidzoda-Tajik-E5-Large:微调 560M 嵌入模型用于塔吉克语检索
Saidzoda-Tajik-E5-Large 是基于 560M 参数 multilingual-e5-large-instruct 微调的塔吉克语文本嵌入模型。该模型在 235,140 个问答对上训练,使未见法律和税务文档检索的 top-10 命中率分别提升至 93.6% 与 94.5%。在同风格测试中,其表现追平 5 倍参数规模的 Giga-Embeddings-3B,且保留了俄语与英语能力。
Story heat trend
Current heat 5·Comparable peak 9(Nov 3 18:00)·Comparable change, last 24 hours–
Trends compare only the same parties observed continuously, so the scope can be narrower than current heat counts. Move the pointer or tap the chart to see hourly heat; use the left and right arrow keys on a keyboard.