diff --git a/tank/arbitrary-image-stylization-v1-256/arbitrary-image-stylization-v1-256_tflite_test.py b/tank/arbitrary-image-stylization-v1-256/arbitrary-image-stylization-v1-256_tflite_test.py
index 19ea0781..ddc41c46 100644
--- a/tank/arbitrary-image-stylization-v1-256/arbitrary-image-stylization-v1-256_tflite_test.py
+++ b/tank/arbitrary-image-stylization-v1-256/arbitrary-image-stylization-v1-256_tflite_test.py
@@ -77,9 +77,10 @@ class ArbitraryImageStylizationV1TfliteModuleTest(unittest.TestCase):
         )
         self.module_tester.save_mlir = self.save_mlir
 
-    @pytest.mark.skip(
-        reason="known macos tflite install issue & "
-        "'tosa.conv2d' op attribute 'quantization_info' failed "
+    import sys
+
+    @pytest.mark.xfail(
+        reason="'tosa.conv2d' op attribute 'quantization_info' failed ",
     )
     def test_module_static_cpu(self):
         self.module_tester.dynamic = False
diff --git a/tank/birds_V1/birds_V1_tflite_test.py b/tank/birds_V1/birds_V1_tflite_test.py
index 3dd597fa..78392325 100644
--- a/tank/birds_V1/birds_V1_tflite_test.py
+++ b/tank/birds_V1/birds_V1_tflite_test.py
@@ -111,9 +111,10 @@ class BirdsV1TfliteModuleTest(unittest.TestCase):
         self.module_tester = BirdsV1TfliteModuleTester(self)
         self.module_tester.save_mlir = self.save_mlir
 
-    @pytest.mark.skip(
-        reason="known macos tflite install issue & "
-        "'tosa.conv2d' op attribute 'quantization_info' failed "
+    import sys
+
+    @pytest.mark.xfail(
+        reason="'tosa.conv2d' op attribute 'quantization_info' failed ",
     )
     def test_module_static_cpu(self):
         self.module_tester.dynamic = False
diff --git a/tank/gpt2-64/gpt2-64_tflite_test.py b/tank/gpt2-64/gpt2-64_tflite_test.py
index 6c458a91..46500978 100644
--- a/tank/gpt2-64/gpt2-64_tflite_test.py
+++ b/tank/gpt2-64/gpt2-64_tflite_test.py
@@ -32,7 +32,7 @@ def compare_results(mlir_results, tflite_results):
         tflite_result = tflite_results[i]
         mlir_result = mlir_result.astype(np.single)
         tflite_result = tflite_result.astype(np.single)
-        mlir_result = np.expand_dims(mlir_result, axis=0)
+        # mlir_result = np.expand_dims(mlir_result, axis=0)
         print("mlir_result.shape", mlir_result.shape)
         print("tflite_result.shape", tflite_result.shape)
         assert mlir_result.shape == tflite_result.shape, "shape doesnot match"
@@ -104,7 +104,6 @@ class GptTfliteModuleTest(unittest.TestCase):
         self.module_tester = GptTfliteModuleTester(self)
         self.module_tester.save_mlir = self.save_mlir
 
-    @pytest.mark.skip(reason="gpt2-64.tflite model too big")
     def test_module_static_cpu(self):
         self.module_tester.dynamic = False
         self.module_tester.device = "cpu"
diff --git a/tank/nasnet/nasnet_tflite_test.py b/tank/nasnet/nasnet_tflite_test.py
index e8e53f2a..415b3f31 100644
--- a/tank/nasnet/nasnet_tflite_test.py
+++ b/tank/nasnet/nasnet_tflite_test.py
@@ -73,7 +73,6 @@ class NasnetTfliteModuleTest(unittest.TestCase):
         self.module_tester = NasnetTfliteModuleTester(self)
         self.module_tester.save_mlir = self.save_mlir
 
-    @pytest.mark.skip(reason="nasnet tflite model too big")
     def test_module_static_cpu(self):
         self.module_tester.dynamic = False
         self.module_tester.device = "cpu"