Can I execute a Model on a Stream in Foundry?

[VincentF]

I know this is possible to execute a call to AIP in Pipeline Builder in streaming mode.

I want to execute a model (not AIP/not an LLM) as part of a streaming pipeline.
This model is a “Model” in Foundry (it could be a sentiment analysis BERT model, an audio transcription model, a remote model available via API call…)

How can I do this ? Is it possible ?

[pchristofi]

There is currently no first class that support of this.

[9eab477d6eb76c06c4f6]

Done a quick trial on that using Compute Modules.

A Video stream + YOLO model in a container and it was working as expected.