Despite being spoken by nearly 50 million individuals, the Uzbek language remains underrepresented in Natural Language Processing (NLP). One primary reason is the limited availability of Uzbek linguistic resources. With the rising prominence of the Transformer architecture in NLP, it has overtaken earlier methods like convolutional and recurrent neural networks. The T5 (Text-to-Text Transfer Transformer) standardizes linguistic tasks in English by converting them into a text-to-text format. The mT5, its multilingual version, has shown promising outcomes in various NLP tasks spanning multiple languages. However, the considerable dimensions of the mT5 pose challenges for applications focused on a singular language. In our study, we fine-tuned the mT5 specifically for Uzbek, leading to a more compact T5 model. We compared this tailored model's efficiency with the mT5 on Automatic Text Summarization (ATS) and Named Entity Recognition (NER) tasks using identical protocols and datasets. Our adapted model surpassed the performance of the mT5, indicating the feasibility of developing a more compact pre-trained model with nearly half the size, without compromising results. This streamlined model also benefits from reduced memory usage, faster startup, and swifter processing times. For access to this model, please reach out.
| Mualliflar | Адилова, Ф.Т., Давронов, Р.Р., Кушмуратов, С.И. |
|---|---|
| Jurnal | Рақамли технологияларнинг назарий ва амалий масалалари |
| Nashr sanasi | 2023-10-02 |
| Jild | 5 |
| Son | 3 |
| Betlar | 7-16 |
| Til | Ingliz |
model compression, transformer, pre-trained model, automatic text summarization, named entity recognition, сжатие модели, преобразователь, предварительно обученная модель, автоматическое суммирование текста, распознавание именованных объектов
Fires are one of the most serious and destructive natural and man-made disasters, having dangerous consequences for human life, property and the environment. This article presents the general structure of a hardware and…
In this paper, the issues of analysis and practical application of traffic lights, remote control methods using telecommunication technologies in densely populated urban areas are considered. Practical solutions for…
Shaxsni ovozi asosida identifikatsiyalash tizimlari bugungi kunda amaliyotda keng qoʻllanilmoqda. Jumladan, ma’lumotlar bazasiga kirish, bank, kriminalistik ekspertiza kabilarda sohalarda keng qoʻllanilmoqda. Shaxsni…
The article deals with the problem of business development in Russia after the deterioration of the economic situation in the country in 2022. The topic of the work is very relevant, as currently Russian companies…
In the article reviewed basics of quantization of digital images, discriting non-continuous images sampling and uniform quantization of brightness based on the use of a rectangular raster are practiced. This is due to…
Ushbu maqolada malakali kadrlarni tayyorlash va ularning bilimi, saviyasini oshirish, yurtimizni gullab yashnashi uchun barcha imkoniyatlardan unumli foydalanib, dunyo miqiyosidagi oʻqitish tizimidan yuqori oʻrinlarga…
Currently, the demand for mobile applications for evaluating and analyzing the performance of medical information systems is increasing due to the need of time for digitization. Medical mobile applications are software…
Maqolada koʻchmas mulk savdolari uchun axborot tizimini qanday ishlatish va boshqarish, moliyaviy natijalari va tizimlarning foydalanishidagi imkoniyatlari koʻrib chiqilgan. Samarqand viloyatidagi qurilish…
Language recognition in natural language processing (NLP) aims to determine the specific language of a text or document. As the number of languages increases, this task becomes more complex. This study introduces a…
Ushbu tadqiqot ishining asosiy maqsadi mantiqiy funksiyalarni qisqartirish orqali mantiqiy o‘zgaruvchilarni minimallashtirish bo‘lib, elektron sxemalarda Quine-McCluskey minimallashtirish usuli tahlil qilingan. Mantiqiy…
Рақамли технологияларнинг назарий ва амалий масалалари — barcha maqolalar