7月22日,科技媒体9to5Mac发布博文,深入分析苹果公司名为《Apple Intelligence Foundation Language Models – Tech Report 2025》的AI技术报告。报告详细描述了苹果在AI模型架构、数据处理、预训练、后训练、工具开发、优化和性能基准等方面的进展。博文指出四个关键点:苹果端侧AI模型采用双区块设计,减少约37.5%的内存需求并加快输出速度;云端模型采用Parallel-Track Mixture-of-Experts(PT-MoE)架构,通过激活相关子网络提高响应速度和精确度;苹果显著增强了多语言支持,将外语数据训练比例从8%提高到30%,并增加标记器容量至150K;苹果承诺不会抓取不同意数据被抓取的出版商的数据。
海量资讯、精准解读,尽在新浪财经APP
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.