Closed Tombana closed 3 years ago
The converter is incompatible with flatbuffers==2.0 (specifically the re-serialization of a flatbuffer is broken, when we remove the Quantize/Dequantize ops in python) so in this PR we pin flatbuffers<2.0 just as tensorflow has.
flatbuffers==2.0
flatbuffers<2.0
What do these changes do?
The converter is incompatible with
flatbuffers==2.0(specifically the re-serialization of a flatbuffer is broken, when we remove the Quantize/Dequantize ops in python) so in this PR we pinflatbuffers<2.0just as tensorflow has.