Allow different variant of ADD kernel (#1139)

This corresponds to cl/9257066. This allows some variant
ADD kernel optimization.

BUG=https://b/231368244Co-authored-by: Nmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

tensorflow/lite/micro/micro_mutable_op_resolver.h

@@ -120,8 +120,8 @@ class MicroMutableOpResolver : public MicroOpResolver {
                       ParseAbs);
   }
 
-  TfLiteStatus AddAdd() {
-    return AddBuiltin(BuiltinOperator_ADD, tflite::Register_ADD(), ParseAdd);
+  TfLiteStatus AddAdd(const TfLiteRegistration& registration = Register_ADD()) {
+    return AddBuiltin(BuiltinOperator_ADD, registration, ParseAdd);
   }
 
   TfLiteStatus AddAddN() {