**360集团创始人周鸿祎宣布开源大规模预训练模型，引领技术创新潮流**
360集团创始人、知名企业家周鸿祎在最近的一次直播活动中宣布，360集团即将开源其先进的360智脑70亿参数模型，这一模型将支持高达50万字的长文本输入，创下了国内开源长文本能力的新纪录。周鸿祎在直播中提到，这一举措是为了响应业界对于大模型文本长度不断提升的趋势，同时他也表达了自己对开源精神的坚定信仰，认为开源能够推动技术的快速发展，避免行业内的重复劳动。
“我们看到，大模型行业正在竞相提升文本处理的长度，100万字可能很快会成为新的标准。因此，360决定开源我们的70亿参数模型，它的输入长度设定在360K，即50万字，寓意着‘360全方面覆盖’。”周鸿祎解释道，这一决定不仅仅是技术上的突破，更是一种对开源文化的致敬。
360智脑70亿参数模型的开源，将为开发者和研究者提供一个强大的工具，他们可以利用这个模型进行各种自然语言处理任务，如文本生成、问答系统和机器翻译等。这一开源行动不仅彰显了360集团在人工智能领域的技术实力，也体现了其对开源社区的贡献和对技术创新的推动。
据金融界消息，360k的文本处理能力在目前国内已开源的模型中是最长的，这无疑将为国内的AI研发带来新的机遇和挑战。周鸿祎的开源决定，无疑将激发更多开发者和研究者在人工智能领域的创新活力，进一步推动中国AI技术的快速发展。
英语如下：
**News Title:** “Zhou Hongyi Announces: 360 Brain 7B Model Open Sourced, Pioneering the Era of 500,000-Character Long Text Input”
**Keywords:** Zhou Hongyi, 360 Brain, Open-Source Model
**News Content:**
360 Group’s founder, Zhou Hongyi, announced the open sourcing of a large-scale pre-training model, spearheading the trend of technological innovation.
In a recent live streaming event, Zhou Hongyi, the founder of 360 Group and a renowned entrepreneur, declared that the group would open-source its advanced 360 Brain model with 7 billion parameters. This model supports text input of up to 500,000 characters, setting a new domestic record for open-source long-text capabilities. Zhou mentioned in the live stream that this move responds to the industry’s trend of increasing model text lengths and underscores his steadfast belief in the open-source spirit, which he believes fosters rapid technological advancement and avoids redundant efforts in the industry.
“We see the big model industry racing to enhance text processing lengths, with 1 million characters possibly becoming the new standard soon. Hence, 360 has decided to open-source our 7 billion parameter model, with an input length of 360K, or 500,000 characters, symbolizing ‘360 all-round coverage’,” Zhou explained. This decision, he added, is not only a technological breakthrough but also a tribute to the open-source culture.
The open-sourcing of the 360 Brain 7 billion parameter model will provide developers and researchers with a powerful tool for various natural language processing tasks, such as text generation, question-answering systems, and machine translation. This open-source initiative not only showcases 360 Group’s technological prowess in artificial intelligence but also highlights its contribution to the open-source community and its drive for technological innovation.
According to financial circles, the 360K text processing capacity is currently the longest among domestically open-sourced models, presenting new opportunities and challenges for AI development in China. Zhou’s open-source decision is expected to ignite a surge of innovative energy among developers and researchers in the AI field, further accelerating the rapid development of China’s AI technology.
【来源】https://finance.jrj.com.cn/2024/03/29115840050454.shtml
Views: 1