Fix decoding special tokens in SentencePiece tokenizer

[tarekgh]

There was corner case while decoding special tokens Ids in Sentence Piece tokenizer and the special tokens ids show up in the beginning of the decoding list. The change here addresses that and ensures the special tokens will get decoded as expected.

[tarekgh]

CC @LittleLittleCloud

[codecov]

Codecov Report

Attention: Patch coverage is 77.90698% with 19 lines in your changes missing coverage. Please review.

Project coverage is 68.83%. Comparing base (4e364e4) to head (50a86c0).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...t.ML.Tokenizers/Model/SentencePieceBpeTokenizer.cs	75.64%	14 Missing and 5 partials ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #7233   +/-   ##
=======================================
  Coverage   68.82%   68.83%           
=======================================
  Files        1453     1453           
  Lines      271527   271562   +35     
  Branches    28094    28094           
=======================================
+ Hits       186885   186929   +44     
+ Misses      77424    77411   -13     
- Partials     7218     7222    +4     

Flag	Coverage Δ
Debug	68.83% <77.90%> (+<0.01%)	⬆️
production	63.35% <75.64%> (+<0.01%)	⬆️
test	89.03% <100.00%> (+<0.01%)	⬆️

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

Files with missing lines	Coverage Δ
test/Microsoft.ML.Tokenizers.Tests/LlamaTests.cs	99.85% <100.00%> (+<0.01%)	⬆️
...t.ML.Tokenizers/Model/SentencePieceBpeTokenizer.cs	76.86% <75.64%> (+1.20%)	⬆️

... and 7 files with indirect coverage changes