The trained model is in the form of meta graph. So I loaded it to model optimizer:
$mo_tf.py --input_meta_graph <INPUT_META_GRAPH>.meta
After executing the above commands, I got the error as follows:
Model Optimizer version: 126.96.36.199d067a0
Converted 259 variables to const ops.
[ ERROR ] Exception occurred during running replacer "None" (<class 'extensions.front.tf.assign_elimination.AssignAddElimination'>): Data flow edge coming out of AssignAdd node Adam
Is there anyone encountering the similar issue?
I have the same problem,It seems that only frozen model can be converted to model optimizer directly.So I only succeeded in converting the frozen model of tensorflow(eg.pb file) using "python mo_tf.py --input_model <pb file name>.pb --input_shape <shape>"