Skip to content
Discussion options

You must be logged in to vote

三方模型和多模态现在都能用;即使主模型本身是 text-only,也可以给它装一条独立视觉链:

dsh plugin --profile web add pi2dsh@0.10.0
dsh plugin --profile web add @kassing/pi-vision

安装后每条文本模型 route 会出现 <route>-vision 伴生项。Web 直接粘贴图片:视觉插件先用你配置的 vision endpoint 分析,结果进入上下文,再由原来的 DeepSeek/GPT/Kimi 等文本 route 回答;像素不会误发给 text-only provider。

真实 DSH Web 已用红、绿、蓝三张图跑通图片准入、分析注入和最终回答。也有按需工具路线 pi-vision-tool,让 Agent 自己调用 describe_image。配置与截图: https://github.com/weijiafu14/pi2dsh/tree/main/examples/vision-bridge

Replies: 3 comments

Comment options

You must be logged in to vote
0 replies
Comment options

You must be logged in to vote
0 replies
Comment options

You must be logged in to vote
0 replies
Answer selected by henshenry2026-lab
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
4 participants