AI model Zeta to expand use of Tibetan language
A new Tibetan language AI model named Zeta was unveiled in Beijing, offering enhanced support for sectors like healthcare, education, and cultural preservation. Zeta integrates Tibetan, standard Chinese, and English within a multilingual framework and is supported by a 150-gigabyte Tibetan text corpus.
A Tibetan language AI model, Zeta, was launched in Beijing, expanding digital access and public services across Tibetan-speaking regions. Developed by the State Key Laboratory of Tibetan Intelligence at Qinghai Normal University, Zeta is Qinghai's first large-scale model for full multimodal processing of the Tibetan language. It has been filed under national regulations for generative AI services and entered public testing. Zeta takes a comprehensive approach to language processing, spanning listening, speaking, reading, writing, and translating across three main Tibetan dialects. The model incorporates innovations like mixed-language document recognition and intelligent subtitle transcription. It is supported by a 150-gigabyte Tibetan text corpus, 87 million multilingual parallel sentences, and 30,000 hours of multi-dialect audio recordings. Zeta is designed to tackle challenges like linguistic translation, voice recognition, and literature preservation, with versions available in 7 billion, 50 billion, and 122 billion parameters.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.