You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi! It's a good start of the awesome project! Thank you for your work.
When I call /models it gives me only single model from config.yaml, not list of available models on Ollama server.
And other thing:
It's named openwebui-telegram, but work with Ollama api /ollama/api/v1 instead of /api (/api/chat/completions)
Also we can use models/models list directly from openwebui! It will be the best implementation, and know why? Because we can have 1 model in Ollama and 1000 models in openwebui used this one model, but with different settings and SYSTEM prompts.
Why is important? Because I very interested in using of tools, especially for generating images from my selfhosted automatic1111 instance.
Now it works perfect in openwebui with a tool named "image_gen".
I know, of cource we can use directly /images/api/v1/ but it sounds wrong, when we have openwebui as a router to all the things
Anyway, I hope you can implement image generation - it would be nice to have this functional!.
Hi! It's a good start of the awesome project! Thank you for your work.
When I call
/models
it gives me only single model from config.yaml, not list of available models on Ollama server.And other thing:
It's named openwebui-telegram, but work with Ollama api
/ollama/api/v1
instead of/api
(/api/chat/completions)Also we can use models/models list directly from openwebui! It will be the best implementation, and know why? Because we can have 1 model in Ollama and 1000 models in openwebui used this one model, but with different settings and SYSTEM prompts.
Why is important? Because I very interested in using of tools, especially for generating images from my selfhosted automatic1111 instance.
Now it works perfect in openwebui with a tool named "image_gen".
I know, of cource we can use directly
/images/api/v1/
but it sounds wrong, when we have openwebui as a router to all the thingsAnyway, I hope you can implement image generation - it would be nice to have this functional!.
Make your bot the best of the best with full fuctional!
Upd. Created a small PR
Tested and it works like with Ollama, but will be helpful to implement other things.
Thank you so much!
The text was updated successfully, but these errors were encountered: