DSH Image Vision
xiaoyuink/dsh-image-vision · model-adapters
Photo Identification Plugin for DeepSeek Harness: Autojudge current model recognition capability to support multi-supplier visual model management and testing
★ 3GitHub stars
0Forks
2026-09-04Last updated
JavaScriptLanguage
—License
Key features
- 7 built-in professional presetspathology / cell / autopsy / statistical chart / large combination / clinical / general.
- Current modelTake the current provider/mode by anentDefaultModel.currentSelement(), and use llm.resolveModelInfo() to determine whether inputModalities contain image.
- Current model visualizationwithout calling the plugin visual model, hand pictures over as image block directly to the current model and analyse them by preset hint.
- The "Add Picture" button appears in the dialogue box box toolbar after installing and opening plugins
- Enter the frame original imagePaste / Upload / Drag and Drag pictures, send auto-literate maps.
Install command
dsh plugin --profile web add github:xiaoyuink/dsh-image-vision