为纯文本模型"看图“设计更好的视觉工具箱和技能,支持多图理解,图片问答,前端UI还原、GUI 自动化等,并可选无缝接入多个主流agent,直接识别粘贴图片| A vision toolkit and skill designed ...
DeepSeek V4 × J-Space capability realization report — benchmark evidence that J-Space reduces capability-realization los...
TongFlow — multimodal workflow studio and engine (canvas + Python plugin engine) and dsh-tongflow, the DeepSeek Harness ...