：今日，由知名创业家李开复创办的国内大模型独角兽公司——零一万物，宣布其Yi大模型API开放平台正式上线。这一举措标志着国内人工智能技术在应用层面的又一重大突破，为开发者和企业提供了强大的工具，以提升智能化服务的效率和质量。
零一万物API开放平台（https://platform.lingyiwanwu.com/）首批推出了三个高性能模型，可处理高达30万汉字的复杂任务。其中，Yi-34B-Chat-0205模型以其通用性备受瞩目，它涵盖了聊天、问答、对话、写作和翻译等多种功能，有望广泛应用于日常生活和工作中，提高人机交互的自然度和效率。
此外，Yi-34B-Chat-200K模型则专精于长篇幅的上下文理解和多文档阅读，能够处理超长知识库的构建，对于需要深度学习和理解的场景，如学术研究、信息检索等领域，提供了强大的支持。
最后，多模态模型Yi-VL-Plus成为了视觉与文本结合的典范。该模型在中文图表理解上的表现超越了GPT-4V，为视觉信息的处理和理解提供了新的可能，将极大地推动多媒体内容的生成和解析。
此次零一万物API开放平台的发布，不仅展示了公司在大模型研发上的技术实力，也为各行业提供了更为智能、高效的解决方案，预示着人工智能技术在服务社会、助力产业升级方面将发挥更大的作用。这一事件无疑将对中国乃至全球的人工智能领域产生深远影响，开启大模型应用的新篇章。
英语如下：
**News Title:** “Innovative Company Lingyi Wanwu, Founded by Kai-Fu Lee, Launches Yi Large Model API: Multimodal Processing of 300,000 Chinese Characters, Pioneering a New Era of Intelligent Interaction”
**Keywords:** Lingyi Wanwu, API Open, Yi Large Model
**News Content:**
**Title:** Lingyi Wanwu’s API Open Platform Officially Launched, Paving the Way for a New Era in Large Model Applications
**Body:** Today, Lingyi Wanwu, a unicorn domestic large model company founded by renowned entrepreneur Kai-Fu Lee, announced the official launch of its Yi Large Model API open platform. This move signifies another significant breakthrough in the application of artificial intelligence (AI) technology in China, providing developers and businesses with powerful tools to enhance the efficiency and quality of their intelligent services.
The Lingyi Wanwu API Open Platform (https://platform.lingyiwanwu.com/) initially introduces three high-performance models capable of handling complex tasks involving up to 300,000 Chinese characters. The Yi-34B-Chat-0205 model, with its versatility, draws particular attention as it encompasses chatting, questioning, dialogue, writing, and translation, poised for wide application in daily life and work, improving the naturality and efficiency of human-machine interaction.
The Yi-34B-Chat-200K model, on the other hand, specializes in long-context understanding and multi-document reading, enabling the processing of extensive knowledge bases. It offers robust support for scenarios requiring deep learning and understanding, such as academic research and information retrieval.
Furthermore, the multimodal model Yi-VL-Plus sets a benchmark for the integration of vision and text. Outperforming GPT-4V in Chinese chart comprehension, it opens new possibilities for visual information processing and understanding, significantly advancing the generation and interpretation of multimedia content.
The launch of the Lingyi Wanwu API Open Platform not only demonstrates the company’s technological prowess in large model development but also offers more intelligent and efficient solutions across industries. It forecasts a greater role for AI technology in serving society and facilitating industrial upgrading. This event is set to have a profound impact on China’s and the global AI landscape, ushering in a new chapter in large model applications.
【来源】https://mp.weixin.qq.com/s/qmjEe56WtQeUgV5iyTx4pg
Views: 1