How to apply trained CNN model to classify unlabeled set of images?
So I've trained a model designed to classify images of tile-level scratch signatures of but I'm unsure how to now run raw data through this model and have it classify that data. The model itself doesn't show up as a node within the flow either even though the model exists and is trained.
If anyone has a solution/ if I'm missing something trivial please let me know!
Thanks.
Operating system used: Windows 10
Operating system used: Windows 10
Answers
-
JakeA Dataiker, Dataiku DSS Core Designer, Dataiku DSS Adv Designer, Registered Posts: 4 DataikerOptionsHi!
Seems like you need to deploy your model into the flow of your project. This knowledge base document should have a comprehensive guide for you to follow:https://knowledge.dataiku.com/latest/courses/scoring/deploy-model/deploy-model.html#deploy-the-model
Once you've done that, using a scoring recipe would be your best bet for classification:
https://knowledge.dataiku.com/latest/courses/scoring/scored-results/scored-results-summary.html