Recently, research teams from several Chinese scientific institutions have launched the super large-scale multimodal dataset named Infinity-MM, and based on this dataset, they have trained an outstanding AI model called Aquila-VL-2B. This breakthrough injects new momentum into the development of multimodal AI. The Infinity-MM dataset is astonishing in size, containing four major types of data: 10 million image descriptions, 24.4 million general visual instruction data, 6 million selected high-quality instruction data, and 3 million generated by G

A research team from China has collaboratively developed an innovative multi-view dataset called 'FreeMan', aiming to address the limitations of existing 3D human pose estimation datasets. The dataset contains 11 million frames of video captured from 8 smartphones, covering both indoor and outdoor environments with diverse lighting conditions, providing a rich resource for real-world scenario diversity. The researchers generated accurate 3D annotations through an automated labeling process, which can be utilized for various tasks such as single 2D to 3D conversion, multi-view 3D estimation, and neural rendering.

{title:The Popular 'Video Version of ControlNet' is Open Source! Precisely Change Art Styles with Prompt Words, Created by a Fully Chinese Team list: The latest video processing algorithm CoDeF, developed by a fully Chinese team, has rapidly gone viral online after its release. CoDeF can precisely control the art style of videos based on prompt words while maintaining lip sync. CoDeF achieves video style transfer tasks through content deformation fields and possesses good cross-frame consistency}

With mathematical capabilities surpassing ChatGPT, Claude Instant 1, and PaLM 2-540B, the WizardMath model developed by Microsoft's all-Chinese team has gained popularity for its ability to enhance large model capabilities generated by AI instructions.