Azure Foundry "Chat Completion" Step: Does Step-Level Model Selection Override Azure Foundry Setting
Hello,
When using the Azure Foundry module's "Chat Completion" step in Decisions, does specifying a model in the step input or agent configuration override the model and deployment defined in System > Settings > Azure Foundry Settings?
From my testing, it seems like the step always uses the deployment/model set in the global settings, regardless of what I specify in the step or agent. Is this the intended behavior, or am I missing something? Is there a way to use different models for different steps?
When using the Azure Foundry module's "Chat Completion" step in Decisions, does specifying a model in the step input or agent configuration override the model and deployment defined in System > Settings > Azure Foundry Settings?
From my testing, it seems like the step always uses the deployment/model set in the global settings, regardless of what I specify in the step or agent. Is this the intended behavior, or am I missing something? Is there a way to use different models for different steps?
0
Comments
-
Hello,
This is the expected behavior. The "Chat Completion" step always uses the model and deployment defined in Azure Foundry Settings (System > Settings), regardless of what you specify in the step input or agent configuration. Step-level or agent-level model selections are currently ignored.
Hope this helps!0
Howdy, Stranger!