Open edangx100 opened 2 months ago
You can use the agent.code.model=...
argument to use a Claude model for the coding step, however the execution feedback step currently only supports the OpenAI API (for function calling) so you will still need to provide an API key for that.
I tried what you described (agent.code.model=..
) at file 'config.yaml'.
Both Claude and OpenAI keys are then exported as suggested from README.
But when i run (Aide), only OpenAI model is used for both (1) coding as well as (2) execution feedback. Claude model was not used.
How to use Claude model here?
I tried what you described (
agent.code.model=..
) at file 'config.yaml'. Both Claude and OpenAI keys are then exported as suggested from README.But when i run (Aide), only OpenAI model is used for both (1) coding as well as (2) execution feedback. Claude model was not used.
How to use Claude model here?
can you plz show me what your config looks like.
config.yaml
filecfg
object in run.py
I had performed the step of export the Anthropic key as stated in README file. But i am still prompted to use OpenAI key. How can i use Anthropic Claude and not OpenAI?