[SameResize] ensure idt, PaddingStyle correctness

- SameResize_Batch hlslib op always pads with 0, i/odt must support
zero values
- Added assertion to check for PaddingStyle=2, the only tested
value

src/finn/custom_op/fpgadataflow/sameresize_batch.py

@@ -86,6 +86,8 @@ class SameResize_Batch(HLSCustomOp):
         node = self.onnx_node
         # data type stays the same
         dtype = model.get_tensor_datatype(node.input[0])
+        exp_idtype = self.get_input_datatype()
+        assert dtype == exp_idtype, "Unexpected datatype for SameResize_Batch"
         model.set_tensor_datatype(node.output[0], dtype)
 
     def verify_node(self):
@@ -93,11 +95,15 @@ class SameResize_Batch(HLSCustomOp):
 
     def get_input_datatype(self):
         """Returns FINN DataType of input."""
-        return DataType[self.get_nodeattr("inputDataType")]
+        ret = DataType[self.get_nodeattr("inputDataType")]
+        # the hlslib op always pads with zeroes, so ensure that the DataType
+        # is able to represent zeroes
+        assert ret.allowed(0), "SameResize_Batch DataType must support zero"
+        return ret
 
     def get_output_datatype(self):
         """Returns FINN DataType of output. (Same as input datatype)"""
-        return DataType[self.get_nodeattr("inputDataType")]
+        return self.get_input_datatype()
 
     def get_instream_width(self):
         ibits = self.get_input_datatype().bitwidth()
@@ -120,6 +126,7 @@ class SameResize_Batch(HLSCustomOp):
 
     def defines(self, var):
         numReps = 1
+        assert self.get_nodeattr("PaddingStyle") == 2, "Only PaddingStyle=2 supported"
         self.code_gen_dict["$DEFINES$"] = [
             """#define ImgDim1 {}\n #define KernelDim1 {}\n
             #define Stride1 {}\n #define NumChannels1 {}\n
@@ -240,12 +247,7 @@ class SameResize_Batch(HLSCustomOp):
             inp.shape == exp_ishape
         ), """Input shape doesn't
         match expected shape (1, ImgDim, ImgDim, NumChannels)."""
-        if self.get_input_datatype() == DataType.BIPOLAR:
-            # store bipolar activations as binary
-            inp = (inp + 1) / 2
-            export_idt = DataType.BINARY
-        else:
-            export_idt = self.get_input_datatype()
+        export_idt = self.get_input_datatype()
 
         # no reshaping for input since assuming no folding on input
         # make copy before saving array
@@ -290,11 +292,6 @@ class SameResize_Batch(HLSCustomOp):
                     mode
                 )
             )
-        # binary -> bipolar if needed
-        if self.get_output_datatype() == DataType.BIPOLAR:
-            out = context[node.output[0]]
-            out = 2 * out - 1
-            context[node.output[0]] = out
         assert (
             context[node.output[0]].shape == exp_oshape
         ), """Output shape doesn't match expected shape