官方纯文本模型也能贴图(dsh-plugin-multimodal) #1709
shinjiyu
started this conversation in
Show and tell
Replies: 4 comments
|
官方线路 重启 仓库:https://github.com/shinjiyu/dsh-plugin-multimodal |
0 replies
|
官方线路 重启 |
0 replies
|
官方纯文本贴图会被拒。这个插件补贴图准入,不是 vision_* 工具箱。 重启 |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
官方线路是纯文本,Web 贴图会被拒:
当前模型不支持图片。这个插件补的是贴图准入,不是视觉工具箱。Official DeepSeek routes reject pasted images (
MODEL_DOES_NOT_SUPPORT_IMAGES). This plugin lets the GUI accept the paste. It is not avision_*toolbox.对应需求:#588
做什么
see_imageAnionex/
dsh-vision-toolkit是另一条路:给 agent 一堆工具,要模型自己去调。本插件让粘贴不被拒。安装
然后重启
dsh web。旧会话的工具表不会更新。配置
DSH_VISION_MODEL必须是真能看图的模型,不要填deepseek-v4-flash/GLM-5.2-FP8。30 秒验收
主模型仍用官方纯文本,配好 sidecar,在 Web 里贴一张报错截图,问「红字说了什么」。过关:不再弹「当前模型不支持图片」。
仓库 / WeChat
All reactions