Thank you for building such an amazing application.
I'm trying to convert a tf-slim pretrained model from deeplabcut model zoo, which is used to predict keypoints on animals.
Having looked into the graph with tensorboard I figured I only need the output of sub-graph "pose". Is there a way to set more than one layer when converting the model?
Getting more than one output layer is possible but the method depends on many variables.
You could read more on the model optimizer method for cutting the models.
However, it seems as you are looking for a single output, could you put the final layer of "pose" sub-graph as the output layer?
Thank you for your question. If you need any additional information from Intel, please submit a new question as Intel is no longer monitoring this thread.