From 8c33614c9b4d362ad31c6d90dd28d3c235987b0e Mon Sep 17 00:00:00 2001
From: Zeeshan Siddiqui <mzs@microsoft.com>
Date: Fri, 1 Nov 2019 13:45:31 -0700
Subject: [PATCH 1/2] Do not pass validation set to Image Classification API.

---
 .../TrainerExtensions/MultiTrainerExtensions.cs                  | 1 -
 1 file changed, 1 deletion(-)

diff --git a/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs b/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs
index 4d917d035f..4803693ced 100644
--- a/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs
+++ b/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs
@@ -240,7 +240,6 @@ public ITrainerEstimator CreateInstance(MLContext mlContext, IEnumerable<Sweepab
             ColumnInformation columnInfo, IDataView validationSet)
         {
             var options = TrainerExtensionUtil.CreateOptions<Options>(null, columnInfo.LabelColumnName);
-            options.ValidationSet = validationSet;
             var logger = ((IChannelProvider)mlContext).Start(nameof(ImageClassificationExtension));
             options.MetricsCallback = (ImageClassificationMetrics metric) => { logger.Trace(metric.ToString()); } ;
             return mlContext.MulticlassClassification.Trainers.ImageClassification(options);

From 2c7962120efb201dac7100e3f8fcd9a81ea417f1 Mon Sep 17 00:00:00 2001
From: Zeeshan Siddiqui <mzs@microsoft.com>
Date: Fri, 1 Nov 2019 14:12:34 -0700
Subject: [PATCH 2/2] remove metric callback since it is set to default in
 ImageClassification Trainer.

---
 .../TrainerExtensions/MultiTrainerExtensions.cs                 | 2 --
 1 file changed, 2 deletions(-)

diff --git a/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs b/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs
index 4803693ced..583e9c81a0 100644
--- a/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs
+++ b/src/Microsoft.ML.AutoML/TrainerExtensions/MultiTrainerExtensions.cs
@@ -240,8 +240,6 @@ public ITrainerEstimator CreateInstance(MLContext mlContext, IEnumerable<Sweepab
             ColumnInformation columnInfo, IDataView validationSet)
         {
             var options = TrainerExtensionUtil.CreateOptions<Options>(null, columnInfo.LabelColumnName);
-            var logger = ((IChannelProvider)mlContext).Start(nameof(ImageClassificationExtension));
-            options.MetricsCallback = (ImageClassificationMetrics metric) => { logger.Trace(metric.ToString()); } ;
             return mlContext.MulticlassClassification.Trainers.ImageClassification(options);
         }