[MacOS] Ollama model support discussion

[yvesete]

I am on a apple silicon Macbook (M3) and did install writing tools shortcut control w, had not the issue #105 all seems ok but it does not do anything. No error but ... nothing. Icon is there, settings working, menu coming up but it does nothing. No error only a big nothing.

[dev-x3Ro]

same here... just the spinning cursor for half a sec and nothing.
Maybe API keys a broken? Is there a way to test those?

[yvesete]

Ah no can't be. forgot to mention I am using ollama local. And it does work for sure.

[dev-x3Ro]

When I select Gemini as a model it works. Looks like an issue with openAI / Local LLM for me

[Aryamirsepasi]

Hi, if the cursor spins but there are no results, then the problem must be in the provider settings. How did you set the Ollama options in the app?

[yvesete]

the cursor does not spin no reaction. Ollama setup is correct and ollama is working. I only see the full menu when short cutting and the settings window is open also but no reaction, no cursor spin nothing. only nothing.

[yvesete]

settings api key : ollama
base url: http://localhost:11434/v1
model name llama3.1:8b

as I said ollama is working ok in other applications and in terminal

[gabrielbr]

Just a small suggestion, but did you try connecting to Ollama with the ip instead of localhost?

[yvesete]

yes it is the same and all other apps are working with localhost also. It is strange because it does only nothing, no cpu usage in activity monitor, just nothing.

[Aryamirsepasi]

This is strange. I use llama3.2 myself and it works without any problems. Maybe something happens when using llama3.1:8b. I'll test that and see what goes wrong.

[yvesete]

mmhhh I tested all the following models:

NAME ID SIZE MODIFIED
deepseek-coder-v2:latest 63fb193b3a9b 8.9 GB 5 weeks ago
qwen2.5-coder:latest 2b0496514337 4.7 GB 5 weeks ago
nomic-embed-text:latest 0a109f422b47 274 MB 3 months ago
llama3.2:latest a80c4f17acd5 2.0 GB 3 months ago
llama3.1:latest 62757c860e01 4.7 GB 5 months ago
llama3:8b-instruct-q8_0 1b8e49cece7f 8.5 GB 7 months ago
llama3:latest a6990ed6be41 4.7 GB 9 months ago
mistral:latest 61e88e884507 4.1 GB 11 months ago
llama2:latest 78e26419b446 3.8 GB 11 months ago

only llama3.2:latest and llama3:8b-instruct-q8_0 seem to work somehow. I test it further and let you know.

[Nabuzata]

Hello.

1. Similar situation (MBA M3) : 0 reactions as I tried to summarize something (or the other options), or talking with the chatbot.
   -> I figured by accident that the window was actually open on my Spotify client, which was completely hidden...

2. In my Terminal, I saved my model by giving it a name : "/save llama". Now to run it, I simply type "ollama run llama".
   -> in the writing tools configuration window, the only thing working now for me are these :
   API Key : llama (or whatever)
   Base URL : same as you
   Model Name : llama

Maybe that will help someone.
Cheers.