1月22日,零一万物公司开源了最新研发的Yi-VL大型语言模型。该模型基于之前开源的Yi语言模型进行开发,包含了两个版本:Yi-VL-34B和Yi-VL-6B。测试结果显示,在多模态任务上的表现优于其他模型,在英文数据集MMMU和中文数据集CMMMU上的结果均属于领先水平。这充分展示了Yi-VL在复杂的跨学科任务上的强大实力。
Title: Baidu releases Yi-VL model, leading in multimodal tasks
Keywords: Baidu, Yi-VL model, multimodal tasks
News content: On January 22nd, Baidu released its latest developed language model – Yi-VL. Based on the previously open-sourced Yi model, Yi-VL contains two versions: Yi-VL-34B and Yi-VL-6B. Test results show that it outperforms other models on multimodal tasks and achieves state-of-the-art results on English benchmark MMMU and Chinese benchmark CMMMU. This fully demonstrates the powerful capability of Yi-VL on complex cross-disciplinary tasks.
【来源】https://www.jiqizhixin.com/articles/2024-01-22-10
Views: 1