Skip to content

Ollama models configuration for kag developer mode #619

Description

@Bernardo93-bp

Hi,
I am trying to replicate the example of kag explained in the user guide (developer mode) https://openspg.github.io/v2/docs_en. I want to use the Ollama models that I have in the docker container ollamaGPU instead of the models of deepseek or Siliconflow. I wrote the following configuration in the example_config.yaml file:

openie_llm: &openie_llm
base_url: http://ollamaGPU:11434
model: qwen2.5:3b
type: openai

chat_llm: &chat_llm
base_url: http://ollamaGPU:11434
model: qwen2.5:3b
type: openai

---------------- EMBEDDINGS ----------------

vectorize_model: &vectorize_model
base_url: http://ollamaGPU:11434/v1
api_key: ollama
model: qwen3-embedding:8b
type: openai
vector_dimensions: 1024
vectorizer: *vectorize_model

When I execute the command: $ knext project create --config_path ./example_config.yaml
I obtain: Error: invalid llm config: {'base_url': 'http://ollamaGPU:11434', 'model': 'qwen2.5:3b', 'type': 'openai'}, for details: Connection error.

Do you know which is the correct configuration for Ollama models?
Note: in product mode the url worked correctly for both all the models.
Thank you for your help!

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions