Hi, I heard that offline batch inference is not recommended in flyte, and I better to deploy model to end-point and send requests there. Is it true?

where did you hear this from

Great, thank you! It popped up, In our internal conversations, maybe a misunderstanding of something. And I couldn't google any direct answer Just to be completely clear for our internal discussion. Even if the model requires GPU, right?

that is indeed true

Awesome, thank you!