Data classification&Value predication: Training failed when choosing "Split or File" validate strategy, "grid" tuner and enable "Subsample option".

[v-Hailishi]

System Information (please complete the following information):
Windows OS: Windows-11-Enterprise-22H2
ML.Net Model Builder 2022: 17.17.0.2332602 (Main Build)
Microsoft Visual Studio Enterprise: 2022(17.5.5)
.Net: 6.0, 7.0

Describe the bug

• On which step of the process did you run into an issue:
  Training failed when choosing "Split or File" validate strategy, "grid" tuner and enable "Subsample option".

TestMatrix
https://testpass.blob.core.windows.net/test-pass-data/wikipedia-detox-250-line-data.tsv
https://testpass.blob.core.windows.net/test-pass-data/taxi-fare.csv

To Reproduce
Steps to reproduce the behavior:

1. Select Create a new project from the Visual Studio start window.
2. Choose the C# Console App (.NET Core) project template.
3. Add model builder by right click on the project.
4. Select Data classification or Value predication scenario.
5. On Data page, choose the "Tabular File" or SQL server data source.
6. On the Train page, click the "Advanced training options..." link.
7. Go to "Tuners" tab, only choose "grid" tuner, go to "Sample strategy" tab, enable "Subsample option" and save it.
8. The training would be failed with error "Must be greater than zero.".

Expected behavior
Training successfully.

Screenshot

Additional context

[LittleLittleCloud]

This is because when enabling subsampling, a fraction will be added to search space which will be used to pull a part from the entire training dataset. The error is thrown because in a small dataset fraction * row count of dataset might be less than 0.

I'll fix this on ML.Net side. In the meantime, the work-around is to avoid using grid search + subsample at the same time

[v-Hailishi]

The bug has been fixed on the latest main build 17.17.0.2361101.