北京——互联网安全巨头360公司近日在开源社区GitHub上发布了一项重大举措，开放了其自主研发的360智脑7B大模型。这款模型包含70亿个参数，刷新了国产开源语言模型的规模纪录，为人工智能领域的研究和应用提供了新的工具。
360智脑7B模型在训练过程中，利用了规模庞大的3.4万亿Tokens的语料库，涵盖中文、英文及代码等多种语言内容，显示了其跨语言处理的强大能力。尤其引人注目的是，该模型支持的文本输入长度最高可达360K，相当于约50万汉字，这在当前的国产开源模型中堪称最长，为长文本理解和生成的应用场景提供了广阔的可能性。
据360公司透露，他们提供了4K、32K和360K三种不同长度的文本输入选项，以适应不同开发者和研究者的实际需求。这一开源行动不仅彰显了360公司在人工智能领域的技术实力，也为全球开发者和学术界提供了一个共享、学习和创新的平台，有望推动自然语言处理技术的进一步发展。
这一举措无疑将激发更多创新应用的诞生，无论是智能客服、机器翻译，还是文本分析、代码智能，都将受益于360智脑7B模型的开源。360公司此举无疑在国产AI技术的开放共享道路上树立了新的里程碑，预示着中国在人工智能领域的自主研发和创新能力正不断攀升。
英语如下：
**News Title:** “360 Open Sources 7 Billion-Parameter Intelligent Brain Model, Setting New Record for Longest Text Input in Domestic Models”
**Keywords:** 360 Intelligent Brain, 7B Parameters, Open-Source Model
**News Content:**
**Title:** 360 Company Launches 7B Parameter Intelligent Brain Model, Pioneering a New Era for Domestic Large Models
**Beijing** — Internet security giant 360 Company recently made a significant move on the open-source platform GitHub, releasing its self-developed 360 Intelligent Brain 7B model. With 7 billion parameters, this model sets a new record for the scale of domestic open-source language models, offering new tools for research and application in the field of artificial intelligence.
The 360 Intelligent Brain 7B model, trained on a vast 3.4 trillion Tokens corpus encompassing Chinese, English, and code, demonstrates its strong cross-lingual processing capabilities. Notably, it supports text input lengths of up to 360K, equivalent to approximately 500,000 Chinese characters, the longest among current domestic open-source models. This feature opens up vast possibilities for long-text understanding and generation applications.
360 Company has disclosed that it offers text input options of 4K, 32K, and 360K lengths to cater to the diverse needs of developers and researchers. By open-sourcing the model, 360 not only showcases its technical prowess in AI but also provides a platform for global developers and academia to collaborate, learn, and innovate, potentially advancing the development of natural language processing technology.
This move is expected to fuel the creation of more innovative applications, benefiting sectors such as intelligent customer service, machine translation, text analysis, and code intelligence. 360 Company’s action sets a new milestone in the open sharing of domestic AI technology, indicating a continuous rise in China’s indigenous research and development capabilities in the field of artificial intelligence.
【来源】https://www.ithome.com/0/761/531.htm
Views: 1