Closed
Description
(source issue: zhenhuaw-me/tflite#9)
This change breaks software stacks that depend on the built TFLite model parser, e.g. tvm, tflite2onnx, and etc.
Starting from TensorFlow 2.4.0, the BuiltinOperator switches to int32 from byte. As a result, the OperatorCode.builtin_code is now an int32 too, and code like op_code.BuiltinCode() is broken.
// An OperatorCode can be an enum value (BuiltinOperator) if the operator is a
// builtin, or a string if the operator is custom.
table OperatorCode {
// This field is for backward compatibility. This field will be used when
// the value of the extended builtin_code field has less than
// BulitinOperator_PLACEHOLDER_FOR_GREATER_OP_CODES.
deprecated_builtin_code:byte;
custom_code:string;
// The version of the operator. The version need to be bumped whenever new
// parameters are introduced into an op.
version:int = 1;
// This field is introduced for resolving op builtin code shortage problem
// (the original BuiltinOperator enum field was represented as a byte).
// This field will be used when the value of the extended builtin_code field
// has greater than BulitinOperator_PLACEHOLDER_FOR_GREATER_OP_CODES.
builtin_code:BuiltinOperator;
}
For any code that uses op_code.BuiltinCode(), a workaround like the below (or this PR) is needed.
if op_code.BuiltinCode() < BuiltinOperator.PLACEHOLDER_FOR_GREATER_OP_CODES:
opc = op_code.DeprecatedBuiltinCode()
else:
opc = op_code.BuiltinCode()