《Google：Gemini 技术报告（英文版）（90页）.pdf》由会员分享，可在线阅读，更多相关《Google：Gemini 技术报告（英文版）（90页）.pdf（90页珍藏版）》请在三个皮匠报告上搜索。

1、Gemini:A Family of Highly CapableMultimodal ModelsGemini Team,Google1This report introduces a new family of multimodal models,Gemini,that exhibit remarkable capabilitiesacross image,audio,video,and text understanding.The Gemini family consists of Ultra,Pro,and Nanosizes,suitable for applications ran

2、ging from complex reasoning tasks to on-device memory-constraineduse-cases.Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra modeladvances the state of the art in 30 of 32 of these benchmarks notably being the first model to achievehuman-expert performance on the wel

3、l-studied exam benchmark MMLU,and improving the state of theart in every one of the 20 multimodal benchmarks we examined.We believe that the new capabilities ofthe Gemini family in cross-modal reasoning and language understanding will enable a wide variety ofuse cases.We discuss our approach toward

4、post-training and deploying Gemini models responsibly tousers through services including Gemini,Gemini Advanced,Google AI Studio,and Cloud Vertex AI.1.IntroductionWe present Gemini,a family of highly capable multimodal models developed at Google.We trainedGemini models jointly across image,audio,vid

5、eo,and text data for the purpose of building a modelwith both strong generalist capabilities across modalities alongside cutting-edge understanding andreasoning performance in each respective domain.Gemini 1.0,our first version,comes in three sizes:Ultra for highly-complex tasks,Pro for enhancedperf

6、ormance and deployability at scale,and Nano for on-device applications.Each size is specificallytailored to address different computational limitations and application requirements.After large-scale pre-training,we post-train our models to improve overall quality,enhance targetcapabilities,and ensur