[Bug]: Missing outputs when converting ONNX to IR with ovc about openvino HOT 5 CLOSED

The ONNX model does have outputs.

This was generated with

python3 -m tf2onnx.convert --saved-model saved_model/ --output model.onnx

from openvino.

@blakeblackshear thank you for reaching the OpenVINO!

We'll take a look from both ONNX and TF perspective as it seems that they have a common issue.

Thank you!

from openvino.

Hi @blakeblackshear, how do you check results? Did you try enumerate Result nodes? If you use names, try to add a suffix "/sink_port_0".

https://github.com/openvinotoolkit/openvino/blob/master/src/frontends/onnx/frontend/src/core/graph.cpp#L313

from openvino.

Thank you @gkrivor!

@blakeblackshear I've checked the IR generated for the TF model and there are result nodes named the same as in the original TF model. If the issue is with how outputs are displayed in Netron app then you have to submit the issue to the https://github.com/lutzroeder/netron

from openvino.

Sorry for the hassle. I shouldn't have assumed that Netron was displaying outputs correctly. Thanks for the quick support.

from openvino.