Arm backend: Convert remaining asserts in operators to raise errors (pytorch#10945)

Asserts are converted to proper raises to ensure graph integrity.
Improve error message.

Signed-off-by: Sebastian Larsson <[email protected]>

Author: Sebastian-Larsson

backends/arm/operators/op_sub.py

@@ -163,7 +163,11 @@ def define_node(
         validate_same_dtype(self.target, [*inputs, output])
 
         # Handle int8 (quantized) and int32
-        assert inputs[0].dtype in [ts.DType.INT8, ts.DType.INT32]
+        supported_dtypes = [ts.DType.INT8, ts.DType.INT32]
+        if inputs[0].dtype not in supported_dtypes:
+            raise TypeError(
+                f'IO data type needs to be {supported_dtypes}, got "{inputs[0].dtype}"'
+            )
 
         scale_back = 1.0
         if inputs[0].dtype == ts.DType.INT8:
@@ -228,8 +232,15 @@ def define_node(
             super().define_node(node, tosa_graph, inputs, output)
         else:
             # FP32 Sub lowering
-            assert inputs[0].dtype == ts.DType.FP32
-            assert output.dtype == ts.DType.FP32
+            if (
+                inputs[0].dtype != ts.DType.FP32
+                or inputs[1].dtype != ts.DType.FP32
+                or output.dtype != ts.DType.FP32
+            ):
+                raise TypeError(
+                    f"All IO needs to have data type fp32. Got: {inputs[0].dtype}, "
+                    f"input 2: {inputs[1].dtype} and output: {output.dtype}"
+                )
 
             # MI lowering
             tosa_graph.addOperator(

backends/arm/operators/op_upsample_bilinear2d.py

@@ -153,9 +153,12 @@ def define_node(
         def in_int16_range(x):
             return torch.all(x >= -(2**15)) and torch.all(x <= 2**15 - 1)
 
-        assert in_int16_range(scale_n_yx)
-        assert in_int16_range(scale_d_yx)
-        assert in_int16_range(border_yx)
+        if not in_int16_range(scale_n_yx):
+            raise ValueError("scale_n_yx is out of the int16 range")
+        if not in_int16_range(scale_d_yx):
+            raise ValueError("scale_d_yx is out of the int16 range")
+        if not in_int16_range(border_yx):
+            raise ValueError("border_yx is out of the int16 range")
 
         scales = [scale_n_yx[0], scale_d_yx[0], scale_n_yx[1], scale_d_yx[1]]
 

backends/arm/operators/op_upsample_nearest2d.py

@@ -102,9 +102,8 @@ def define_node(
         validate_num_inputs(self.target, inputs, 3)
         validate_same_dtype(self.target, [inputs[0], output])
 
-        assert (
-            inputs[0].shape is not None and output.shape is not None
-        ), "Only static shapes are supported"
+        if inputs[0].shape is None or output.shape is None:
+            raise ValueError("Only static shapes are supported")
 
         # tosa_shape output is NHWC, take HW
         input_size_yx = torch.tensor(
@@ -121,9 +120,12 @@ def define_node(
         def in_int16_range(x):
             return torch.all(x >= -(2**15)) and torch.all(x <= 2**15 - 1)
 
-        assert in_int16_range(scale_n_yx)
-        assert in_int16_range(scale_d_yx)
-        assert in_int16_range(border_yx)
+        if not in_int16_range(scale_n_yx):
+            raise ValueError("scale_n_yx is out of the int16 range")
+        if not in_int16_range(scale_d_yx):
+            raise ValueError("scale_d_yx is out of the int16 range")
+        if not in_int16_range(border_yx):
+            raise ValueError("border_yx is out of the int16 range")
 
         scales = [scale_n_yx[0], scale_d_yx[0], scale_n_yx[1], scale_d_yx[1]]
         scales_tensor = tosa_graph.addConst(