Issue spinning up docker container with local model

[ceinstaller]

Hey all, I'm new to honcho and am trying to spin up a local docker container using a local model. I've followed the directions here:

https://docs.honcho.dev/v3/contributing/self-hosting#llm-setup

The 4 docker containers spin up, but the API container doesn't start the local service. I see this error in the logs:

pydantic_core._pydantic_core.ValidationError: 10 validation errors for DialecticSettings
LEVELS.max.thinking_budget_tokens
  Field required [type=missing, input_value={'model': 'glm-flash-30B-...', 'provider': 'custom'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.12/v/missing

I'm not sure if this is related to a larger issue but though I'd ask in case someone has any tips.

Thanks in advance!

[jessegonzalez]

You'll need to set the appropriate env var for each validation error logged. If you copied .env.template to .env and are using docker compose, uncomment and set to your desired thinking budget for each. For the error you shared, you'd uncomment and set:

# DIALECTIC_LEVELS__max__THINKING_BUDGET_TOKENS=0