dotnet / machinelearning-modelbuilder

Simple UI tool to build custom machine learning models.
Creative Commons Attribution 4.0 International
269 stars 59 forks source link

Question Answering: The "Answer Index" column's data type is not "Int" in default when choosing SQL Server data. #2805

Closed v-Hailishi closed 8 months ago

v-Hailishi commented 11 months ago

System Information (please complete the following information): Windows OS: Windows-11-Enterprise-22H2 ML.Net Model Builder 2022: 17.17.0.2360601 (Main Build) Microsoft Visual Studio Enterprise: 2022(17.7.7) .Net: 6.0,7.0

Describe the bug

TestMatrix Smallsuad.tsv.txt

To Reproduce Steps to reproduce the behavior:

  1. Using attached file to create Microsoft SQL Server data.
  2. Select Create a new project from the Visual Studio start window.
  3. Choose the C# Console App (.NET Core) project template.
  4. Add model builder by right click on the project.
  5. Select "Question Answering" scenario.
  6. On Data page, choose the "SQL Server" data source.
  7. Click the "Advanced data options..." link, check the "Answer Index" column's data type.

Expected behavior The "Answer Index" column's data type should be "Int" in default when choosing SQL Server data.

Screenshot image

Additional context The bug is not repro when choosing file data. image

zewditu commented 10 months ago

The Fix for this will be we have to display what is detected by prose, but when we load into Loader option we load it as an INT.

v-Hailishi commented 8 months ago

Verified on the latest main build 17.18.2.2416101, "Answer Index" column is forced to Int instead of single if the user chooses single data format in advanced data options. in other words, the training can be successfully now when chooses single data format for "Answer Index" column in advanced data options. image