Add in support for 1 unknown dimension for ONNX runtime.

[michaelgsharp]

Adds support so that you can have 1 unknown dimension for the ONNX runtime models (not including the batch input since we set that to 1).

[codecov]

Codecov Report

Merging #6265 (6e65c3d) into main (9f26edd) will increase coverage by 0.04%.
The diff coverage is 69.69%.

@@            Coverage Diff             @@
##             main    #6265      +/-   ##
==========================================
+ Coverage   68.40%   68.44%   +0.04%     
==========================================
  Files        1141     1141              
  Lines      244881   244905      +24     
  Branches    25404    25408       +4     
==========================================
+ Hits       167520   167635     +115     
+ Misses      70707    70603     -104     
- Partials     6654     6667      +13     

Flag	Coverage Δ
Debug	68.44% <69.69%> (+0.04%)	⬆️
production	62.91% <69.69%> (+0.05%)	⬆️
test	88.99% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/Microsoft.ML.OnnxTransformer/OnnxTransform.cs	89.83% <69.69%> (-1.22%)	⬇️
...ML.Transforms/Text/StopWordsRemovingTransformer.cs	86.38% <0.00%> (-0.15%)	⬇️
...StandardTrainers/Standard/LinearModelParameters.cs	66.07% <0.00%> (+0.25%)	⬆️
src/Microsoft.ML.LightGbm/LightGbmTrainerBase.cs	79.75% <0.00%> (+0.55%)	⬆️
src/Microsoft.ML.FastTree/Dataset/IntArray.cs	13.00% <0.00%> (+0.89%)	⬆️
src/Microsoft.ML.Core/Data/ProgressReporter.cs	77.94% <0.00%> (+6.98%)	⬆️
src/Microsoft.ML.FastTree/Dataset/DenseIntArray.cs	44.39% <0.00%> (+17.66%)	⬆️

[tarekgh]

Added a few comments. LGTM otherwise.