COMPMID-516 Increase tolerance rate of Scale, Conv, fully connected and GEMM

This patch also fix the scale kernel issue where it was calcuated the
scale factor inside the gpu but now in the CPU. The GPU and CPU gave
different result for simple float division operation

Change-Id: Ib6709cb6c41dcf4fc0fa4eb79e481430695bf40e
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/87266
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com>
diff --git a/tests/datasets/ShapeDatasets.h b/tests/datasets/ShapeDatasets.h
index aa653ac..6b3b5c7 100644
--- a/tests/datasets/ShapeDatasets.h
+++ b/tests/datasets/ShapeDatasets.h
@@ -114,6 +114,27 @@
     }
 };
 
+/** Data set containing medium tensor shapes. */
+class MediumShapes final : public ShapeDataset
+{
+public:
+    MediumShapes()
+        : ShapeDataset("Shape",
+    {
+        // Batch size 1
+        TensorShape{ 37U, 37U },
+                     TensorShape{ 27U, 33U, 2U },
+                     TensorShape{ 128U, 64U, 1U, 3U },
+                     // Batch size 4
+                     TensorShape{ 37U, 37U, 3U, 4U },
+                     TensorShape{ 27U, 33U, 2U, 4U },
+                     // Arbitrary batch size
+                     TensorShape{ 37U, 37U, 3U, 5U }
+    })
+    {
+    }
+};
+
 /** Data set containing large tensor shapes. */
 class LargeShapes final : public ShapeDataset
 {
@@ -185,17 +206,13 @@
         : ShapeDataset("InputShape",
     {
         // Batch size 1
-        TensorShape{ 5U, 5U, 3U },
+        TensorShape{ 35U, 35U, 3U },
                      TensorShape{ 32U, 37U, 3U },
-                     TensorShape{ 13U, 15U, 8U },
                      // Batch size 4
-                     TensorShape{ 5U, 5U, 3U, 4U },
                      TensorShape{ 32U, 37U, 3U, 4U },
-                     TensorShape{ 13U, 15U, 8U, 4U },
                      // Batch size 8
-                     TensorShape{ 5U, 5U, 3U, 8U },
                      TensorShape{ 32U, 37U, 3U, 8U },
-                     TensorShape{ 13U, 15U, 8U, 8U },
+                     TensorShape{ 33U, 35U, 8U, 8U },
                      // Arbitrary batch size
                      TensorShape{ 32U, 37U, 3U, 8U }
     })