commit 3584a9cfdf0bcf0e75d38b78ec39e5b083947e19
author Tim Hall <tim.hall@arm.com> Thu Nov 18 22:05:17 2021 +0000
committer tim.hall <tim.hall@arm.com> Thu Nov 25 14:36:26 2021 +0000
tree 00a08343ce1c3a95b9ec8eb3d4414d0d6e58daf7
parent f88771bcdca9bbd7582ba3183d698f5d8f83b7ac

MLBEDSW-3602: Output mismatch on some mobilenet_v1 int8 and int16

 - The failing tests contain operations with dynamic tensors which
are not supported and therefore they should be placed on the CPU.
However, a bug in the removal of RESHAPEs which contain a dynamic
shape prevented this happening.
 - This change adds a check to make sure that RESHAPE ops with a
dynamic shape tensor are not removed and instead are placed on the
CPU.

Signed-off-by: Tim Hall <tim.hall@arm.com>
Change-Id: I2d7481f7f80f99a0f01df100d956933777e6875a

ethosu/vela/tflite_model_semantic.py

diff --git a/ethosu/vela/tflite_model_semantic.py b/ethosu/vela/tflite_model_semantic.py
index 51d1f07..b783bb7 100644
--- a/ethosu/vela/tflite_model_semantic.py
+++ b/ethosu/vela/tflite_model_semantic.py
@@ -160,7 +160,7 @@
             valid, extra = constraint(op)
             if not valid:
                 print(
-                    f"Warning: unsupported TensorFlow Lite semantics for {ext_type} '{op.name}'. Placing on CPU instead"
+                    f"Warning: Unsupported TensorFlow Lite semantics for {ext_type} '{op.name}'. Placing on CPU instead"
                 )
                 print(f" - {constraint.__doc__}")
                 if extra:

ethosu/vela/tflite_supported_operators.py

diff --git a/ethosu/vela/tflite_supported_operators.py b/ethosu/vela/tflite_supported_operators.py
index a3c0dd8..5c7fd51 100644
--- a/ethosu/vela/tflite_supported_operators.py
+++ b/ethosu/vela/tflite_supported_operators.py
@@ -210,6 +210,9 @@
         self.specific_constraints[Op.Mean].append(TFLiteSupportedOperators.constraint_mean_height_width_product)
         self.specific_constraints[Op.Mean].append(TFLiteSupportedOperators.constraint_mean_height_width_product_int8)
 
+        # Reshape specific checks:
+        self.specific_constraints[Op.Reshape].append(TFLiteSupportedOperators.constraint_reshape_shape_constant)
+
     def is_operator_supported(self, op):
         ext_type = optype_to_builtintype(op.type)
         if op.type not in TFLiteSupportedOperators.supported_operators:
@@ -682,3 +685,20 @@
         h, w = shape[hi : hi + 2]
         max_prod = cls.mean_kernel_product_int8
         return h * w <= max_prod, f"Product of height and width is {h * w}"
+
+    @staticmethod
+    def constraint_reshape_shape_constant(op):
+        "Shape must be constant"
+        valid = True
+        extra = []
+
+        reshape_tens = op.inputs[1]
+        if reshape_tens is not None:
+            # constant inputs have either no driving operator or a const one
+            # create a list of non-constant inputs
+            if not (len(reshape_tens.ops) == 0 or reshape_tens.ops[0].type == Op.Const):
+                valid = False
+                extra.append(reshape_tens.name)
+        extra = ", ".join(extra)
+
+        return valid, f"Op has non-const input(s): {extra}"