Ways to save and convert TF to OpenVINO

save_and_convert.sh

There are several options here. For Keras, just simply load the Keras model and use sess = keras.get_session() to link the TF session variable to the Keras model (model.output and model.input are the output and input ops)

# Option 1: Saved TensorFlow model
tf.saved_model.simple_save(session, export_dir, inputs={"x":x, "y":y}, outputs={"z": z})
mo_tf.py -saved_model_dir export_dir

# Option 2: Saved TensorFlow checkpoint
saver = tf.train.Saver()
saver.save(sess, "./model.ckpt")

mo_tf.py -input_meta_graph model.ckpt.meta

# Option 3: Frozen model
frozen_model = tf.graph_util.convert_variables_to_constants(sess, tf.get_default_graph(), ['output'])
tf.train.write_graph(frozen_model, '.', 'frozen_model.pb', as_text=False)
mo_tf.py -input_model frozen_model.pb