Sign up to take part
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
We have built a logistic regression model and a random forest model using Dataiku’s competency. We now want to do away with our original projects and just export the models to be run on new input datasets. Right now we only have the option of sharing the model to our new project for which we will have to retain the original project. Can you please guide us regarding the same? We intend to copy the model (to a less cluttered flow) and then score new datasets based on that.
You can copy the Saved Model and the Training Recipe to a new project by selecting the Saved Model and click on "Copy" at the bottom of the right panel. You will then get prompted with a few parameters regarding the target (current vs. new vs. existing project, keep all versions, etc.) before finally performing the copy and being redirected to the new Flow items created.
Not that you will need to share or copy the initial training Dataset along with it.
Since managing multiple projects for the same use-case can often add overhead, another alternative would be to keep your current Project structure but with a rearrangement of the Flow items using Flow zones. Experience has shown that this feature greatly helps users reach a better readability in their Flows.
Hope this helps!
Hi, thank your for providing this solution: """You can copy the Saved Model and the Training Recipe to a new project by selecting the Saved Model and click on "Copy" at the bottom of the right panel."""
However, has anyone found a programmatic solution for this problem using dataiku's python API ?