使用lobechat上传图片到多模态大模型，无法识别 #5476
Unanswered
liuyiwei3366
asked this question in
General Question | 普通问题
Replies: 1 comment
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
显然，我已经配好了minio以及lobechat和ollama本地的多模态大模型，我做了很多测试。在本地终端中直接输入本地图片的绝对地址是可以调用多模态大模型去识别图片内容的。但在lobechat中它们似乎不能直接访问url的图片。即使我的minio的lobe桶权限为公开的。实际上我在浏览器中输入图片的url也能正常访问。但是模型似乎只能接收将图片下载下来并使用文件地址来识别。请告诉我我该如何解决？谢谢！！
Beta Was this translation helpful? Give feedback.
All reactions