**昆仑万维自研大模型“天工3.0”即将开启公测，同时AI音乐生成大模型SkyMusic启动邀测**
近日，昆仑万维通过官方消息宣布，其自研的大语言模型“天工3.0”即将于4月17日正式开启公测。据公开信息透露，“天工3.0”采用了创新的4千亿级参数MoE混合专家模型技术，标志着我国在自然语言处理领域的一大突破。
值得一提的是，“天工3.0”不仅在语言模型上有显著进展，其AI音乐生成大模型SkyMusic也备受瞩目。SkyMusic将在4月2日率先面向社会开启邀请测试，这将为音乐创作领域带来前所未有的创新力量。开源同步的选择展示了昆仑万维推动技术交流与发展的决心。
业内专家表示，此次昆仑万维的大模型公测将加速自然语言处理和人工智能音乐生成领域的技术迭代与创新竞争。对于消费者而言，这有望带来更为智能、便捷的内容创作体验。期待未来昆仑万维继续在这一领域深耕细作，为社会带来更多科技创新的惊喜。新浪财经讯。
英语如下：
News Title: KunlunTech’s Self-developed Big Model “Tiankong 3.0” to Be Open for Public Testing Soon, SkyMusic, an AI Music Generation Model, Starts Its Invite Test
Keywords: KunlunTech, Tiankong 3. 0, Public Testing Starts
News Content: **KunlunTech’s Self-developed Big Model “Tiankong 3.0” to Start Public Testing Soon, and Its AI Music Generation Model SkyMusic Starts Its Invite Test**
Recently, KunlunTech announced through official channels that its self-developed big language model “Tiankong 3.0” will officially start public testing on April 17th. According to public information, “Tiankong 3.0” adopts innovative 400 billion parameter MoE hybrid expert model technology, marking a breakthrough in the field of natural language processing in China.
It is worth mentioning that SkyMusic, an AI music generation model of “Tiankong 3.0”, is also attracting attention. SkyMusic will be the first to open an invitation test to the public on April 2, which will bring unprecedented innovation to music creation. The choice of open source synchronization shows KunlunTech’s determination to promote technology exchange and development.
Industry experts say that the public testing of KunlunTech’s big model will accelerate technology iteration and innovation competition in the fields of natural language processing and AI music generation. For consumers, this is expected to bring a more intelligent and convenient content creation experience. Looking forward to KunlunTech’s continuous deep exploration in this field and bringing more surprises of technological innovation to the society. Xinlang Finance News.
【来源】https://finance.sina.com.cn/7×24/2024-04-01/doc-inaqhtzx7331878.shtml
Views: 1