prabalpatkar558412-collab commented on issue #37093: URL: https://github.com/apache/beam/issues/37093#issuecomment-3648536359
Hi @damccorm, Thanks for the detailed audit — this was very helpful to read through. I agree with the conclusion that the base ModelHandler behavior around inference_args doesn’t really match reality anymore, given how many handlers already pass or consume them. Allowing inference args by default and pushing validation down into the handlers that actually need it (e.g., TensorRT, and fixing sklearn) seems like a cleaner and more consistent model. I’d be interested in helping with this, especially: updating the base validate_inference_args behavior, adding explicit validation where needed, and making sure the affected handlers behave consistently, along with the CHANGES.md note. If no one is working on this yet, I’d be happy to take it on. Thanks! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
