2

I have am working on a project that requires the use of previously created model objects. They have all been built with scikit-learn but with different library versions (0.23.1, 0.23.2, 1.0.2, and 1.2.2). I have always been under the assumption that if you are using models to make predictions you should be using the version of scikit-learn that the model was built with. Since I only need to make predictions with these model objects I was considering converting them all to onnx objects in order to get around rebuilding all the model objects using the most recent version of scikit-learn. Is this the correct approach or is there a better solution?

magladde
  • 614
  • 5
  • 23
  • 1
    I dont know if there is a better way but your approach will work. Since ONNX was especially build to standardize ML and AI models I would consider this to be a good way – Lenntror Jun 21 '23 at 09:12

0 Answers0