dotnet / dotnet/machinelearning
TensorFlow scoring sample should pass an empty dataview
Open
area-TensorFlow
Priority:1
- Dominant language
- C#
- Stars
- 9.4k
- Forks
- 2k
- Avg merge
- 2d 20h
- Merged PRs (30d)
- 11
Description
This sample is calling Fit() with a dataview generated from one sample: https://github.com/dotnet/machinelearning/blob/master/docs/samples/Microsoft.ML.Samples/Dynamic/TensorFlow/TextClassification.cs
The pipeline contains a pre-trained TensorFlow model, so there is not need (and is misleading) to call Fit with data.
In other places where no dataview is required, we generate an empty one.
Related to #3795
Contributor guide
Assessment
This issue has not been assessed yet.