Closed e42101ex closed 2 years ago
If the name of the input/output OP contains :0
, TensorFlow will error. However, the :0
is automatically added by TensorFlow itself. Therefore, to avoid problems with TensorFlow behavior, the tflite file must be edited to remove all :0
.
The tflite file can be rewritten by the IDE.
I have created a tool to make it easier for you guys to edit tflite files. Please refer to the following repository for a tutorial. https://github.com/PINTO0309/tflite2json2tflite
Issue Type
Bug
OS
Ubuntu
OS architecture
x86_64
Programming Language
Python
Framework
TensorFlow
Download URL for tflite file
nms_yolact_edge.zip
Convert Script
python convert_onnx.py
onnxsim yolact_resnet50_62_17262.onnx yolact_resnet50_62_17262.onnx onnxsim yolact_resnet50_62_17262.onnx yolact_resnet50_62_17262.onnx onnxsim yolact_resnet50_62_17262.onnx yolact_resnet50_62_17262.onnx onnxsim yolact_resnet50_62_17262.onnx yolact_resnet50_62_17262.onnx
python nms_yolact_edge.py
tflite2tensorflow \ --model_path nms_yolact_edge.tflite \ --flatc_path ../flatc \ --schema_path ../schema.fbs \ --output_pb \ --optimizing_for_openvino_and_myriad
Description
Hi @PINTO0309, I was trying to convert yolact-edge with backbone resnet50.
Then I got error while trying the following steps.
Error message shows 'serving_default_input_1:0:0' is invalid. Perhaps it should be 'serving_default_input_1:0'?
Would you please help check it? Thanks you.
Relevant Log Output
Source code for simple inference testing code
No response