using LocalOllama with deepseek model failed, how to use local deepseek model?

[yaozr]

1. LLM pr select "LocalOllama"
2. API key: leave empty
3. Base URL: http://192.168.12.11::11434 (different of my localhost but stay in same net)
4. Model: I cannot see options such as :deepseek-r1:32b but I'm sure it is there

[yaozr]

I try with "LocalOllama" or "Deepseek" at LLM pr, and change local model address, now it could see the model name and I could select the model "seekseej-r1:32b" but then when I wrote "draw 5 balls pls" and click "Run" nothing happen

[KarryCharon]

@yaozr you should use some LLM models that support function calling.

[NanSanOBA]

@yaozr Because the local model is "dumb" and does not return functions correctly. Using blender console you can find out. If you have an api, you will find that the functions it returns are not the same as the functions returned by the local model. That's why it's wrong. Here are my test results.