Closed arm-ser closed 4 months ago
Interesting, I will look into it, it is probably something that I may be able to do if ollama allows for something like this. I will keep you posted
On Wed, Apr 10, 2024, 15:49 Observant4678 @.***> wrote:
Is your feature request related to a problem? Please describe. If I understand correctly, the current setup for Lexido and Ollama requires running both applications on the same server, however, I have minipcs as servers on my network , and a gaming machine with powerfull graphics card. When trying to run ollama on minipc servers, I encounter very slow performance because of hardware. But the gaming machine runs llms just fine.
Describe the solution you'd like I would like to request a new feature that allows users to set a remote host where Ollama is running, instead of having to run it on the same server as Lexido. This for example can be implemented through a new tag similar to "--setModel", such as "--setOllama (to set Ollama's IP and port).
Describe alternatives you've considered I tried relatively small and simple models (such as qwen:0.5b) but even those models are slow and also not capable of doing the required task.
Additional context I think that providing a way to set a remote host for Ollama would greatly improve Lexidos performance and since ollama can be run as a docker container, will help to setup more secure network.
— Reply to this email directly, view it on GitHub https://github.com/micr0-dev/lexido/issues/49, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGJEUKVDQJZ2VFUWSC35UNLY4WJS7AVCNFSM6AAAAABGBBJ4NGVHI2DSMVQWIX3LMV43ASLTON2WKOZSGIZTMMZRGU4TANI . You are receiving this because you are subscribed to this thread.Message ID: @.***>
I want to update you that I am working on this and that soon lexido will support all REST API LLMs. Including remote local LLMs or even ChatGPT, Claude, and more!
Amazing. Cannot wait to test it out. Thank you!!
Just posted version 1.4! Be sure to read the README and the section on remote LLMs If you have any questions feel free to ask, good luck!
Re-open this if there are issues
Is your feature request related to a problem? Please describe. If I understand correctly, the current setup for Lexido and Ollama requires running both applications on the same server, however, I have minipcs as servers on my network , and a gaming machine with powerfull graphics card. When trying to run ollama on minipc servers, I encounter very slow performance because of hardware. But the gaming machine runs llms just fine.
Describe the solution you'd like I would like to request a new feature that allows users to set a remote host where Ollama is running, instead of having to run it on the same server as Lexido. This for example can be implemented through a new tag similar to "--setModel", such as "--setOllama (to set Ollama's IP and port).
Describe alternatives you've considered I tried relatively small and simple models (such as qwen:0.5b) but even those models are slow and also not capable of doing the required task.
Additional context I think that providing a way to set a remote host for Ollama would greatly improve Lexidos performance and since ollama can be run as a docker container, will help to setup more secure network.