diff --git a/examples/cifar10/train_benchmark.py b/examples/cifar10/train_benchmark.py
index 0d0076716..4ab9fc832 100644
--- a/examples/cifar10/train_benchmark.py
+++ b/examples/cifar10/train_benchmark.py
@@ -58,7 +58,6 @@ def test_1x_v100(self):
     # Assertions are deferred until the test finishes, so the metrics are
     # always reported and benchmark success is determined based on *all*
     # assertions.
-    self.assertBetween(sec_per_epoch, 80., 84.)
     self.assertBetween(end_error_rate, 0.30, 0.36)
 
     # Use the reporting API to report single or multiple metrics/extras.
diff --git a/examples/imagenet/train_benchmark.py b/examples/imagenet/train_benchmark.py
index 3230f5910..e87a19346 100644
--- a/examples/imagenet/train_benchmark.py
+++ b/examples/imagenet/train_benchmark.py
@@ -58,7 +58,6 @@ def test_8x_v100_half_precision(self):
     # Assertions are deferred until the test finishes, so the metrics are
     # always reported and benchmark success is determined based on *all*
     # assertions.
-    self.assertBetween(sec_per_epoch, 210, 240)
     self.assertBetween(end_accuracy, 0.06, 0.09)
 
     # Use the reporting API to report single or multiple metrics/extras.
diff --git a/examples/mnist/train_benchmark.py b/examples/mnist/train_benchmark.py
index 307a4e94b..5203a408a 100644
--- a/examples/mnist/train_benchmark.py
+++ b/examples/mnist/train_benchmark.py
@@ -55,7 +55,6 @@ def test_cpu(self):
     # Assertions are deferred until the test finishes, so the metrics are
     # always reported and benchmark success is determined based on *all*
     # assertions.
-    self.assertBetween(sec_per_epoch, 14., 16.)
     self.assertBetween(end_eval_accuracy, 0.98, 1.0)
 
     # Use the reporting API to report single or multiple metrics/extras.