Hi Xia,Linmei Looks like there is an issue with the --tensorflow_use_custom_operations_config parameter, based on this thread another user had a similar issue converting their custom trained model. What they did was change line #57 of the ssd_support_api_v1.14.json file under ...