You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Serializing and deserializing a model to and from ONNX not only changes the names of the output (#2980) but also changes the types of the output. For example, a Score column produced by a model (R4/float) becomes a Vec<R4, 1, 1>.
Thanks for investigating @singlis ! I'll close the issue now. We can just add a comments to the text and samples so that people will know what to expect.
Activity
singlis commentedon Apr 9, 2019
See #2980, this is also by design as ONNX does deal with vector types and will create a float[1] for a ML.Net floating point output.
rogancarr commentedon Apr 9, 2019
Thanks for investigating @singlis ! I'll close the issue now. We can just add a comments to the text and samples so that people will know what to expect.
Missing values default to a `StringDataFrameColumn` (dotnet#2982)
Move DataFrame to machinelearning (#5641)
Merge from main repository (#1)
Actually merge from main (#2)
Update editor config and format samples (#5893)