Python: Semantic Kernel Planners not working with Huggingface AI Service

[Sruinard]

Objective: better experience for developers leveraging opensource models

Many of the examples leverage (Azure) OpenAI, but seem to have not been tested with other model providers. Planners are one such example. I would expect functions to be executed, results parsed and leveraged in further reasoning. Even a simple example of retrieving current date + time is not working/executing.

I have created an example here: huggingface-with-planners

Additional note:

• I believe the main thing is that we should expose more control through the generation config.
• We probably want to be able to set/overwrite the default execution_settings for an ai-service.
• For the planner, we should be able to control the stopping criteria and how the results of the ai-service are returned (i.e. prompt + generated context or just generated content)

[github-actions[bot]]

This issue is stale because it has been open for 90 days with no activity.

[github-actions[bot]]

This issue was closed because it has been inactive for 14 days since being marked as stale.