Hi, could you clarify whether the training dataset used locally is the same as the one used when deploying the model to the cloud? If so, is it possible for me to train the model on my local machine and then only use the cloud for inference?
Yes the training set is the same. However for the out-of-sample we might update it by either adding more data or replacing it. More details is available here: FAQs | CrunchDAO Docs V3
What we recommand is to still include your training code but don’t enable the training when creating the run. Your preference will be remembered for the out of sample.
1 Like