From cf4974dfecf6aa0390a3fc2b7e35b2612e746cac Mon Sep 17 00:00:00 2001
From: Michael Twohey <mtwohey2@yahoo.com>
Date: Sat, 23 Nov 2024 12:07:00 -0800
Subject: [PATCH] Update SeqClassification.cs

Showing the probabilities of an answer
---
 Seq2SeqSharp/Applications/SeqClassification.cs | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/Seq2SeqSharp/Applications/SeqClassification.cs b/Seq2SeqSharp/Applications/SeqClassification.cs
index ae67891..3ecfab1 100644
--- a/Seq2SeqSharp/Applications/SeqClassification.cs
+++ b/Seq2SeqSharp/Applications/SeqClassification.cs
@@ -161,12 +161,21 @@ public override List<NetworkResult> RunForwardOnSingleDevice(IComputeGraph compu
                 using var targetIdxTensor = computeGraph.Argmax(probs, 1);
                 float[] targetIdx = targetIdxTensor.ToWeightArray();
                 List<string> targetWords = m_modelMetaData.TgtVocab.ConvertIdsToString(targetIdx.ToList());
+
+                List<string> allWords = m_modelMetaData.TgtVocab.IndexToWord.Values.Select(s => s).ToList();
+                var combinedOutput = probs.ToWeightArray().Zip(allWords, (prob, word) => $"{word.PadRight(15)}: {prob,10:F8}");
+                Logger.WriteLine(Logger.Level.info, $"\nWrds w/Probs\n{string.Join("\n", combinedOutput)}");
+                                
                 nr.Output.Add(new List<List<string>>());
 
                 for (int k = 0; k < batchSize; k++)
                 {
                     nr.Output[0].Add(new List<string>());
-                    nr.Output[0][k].Add(targetWords[k]);
+                    
+                    // Fetch the corresponding probability for the predicted target index
+                    float probAtTargetIdx = probs.GetWeightAt([0, (long)targetIdx[k]]);
+
+                    nr.Output[0][k].Add($"{targetWords[k]} {probAtTargetIdx:F8}");
                 }
             }