I am doing it from google colab and this is what I do:
%tensorflow_version 1.x
!git clone https://github.com/mtbsteve/redtail
with tf.Session() as sess:
new_saver = tf.train.import_meta_graph('./redtail/stereoDNN/models/ResNet-18_2D/TensorFlow/model-inference-513x257-0.meta')
new_saver.restore(sess, tf.train.latest_checkpoint('./redtail/stereoDNN/models/ResNet-18_2D/TensorFlow/'))
print(sess.run('w1:0'))
Throws the following error when run restore:
ValueError: Can't load save_path when it is None.
According to an internet search it is due to the absence of checkpoint files in the folder.
Can you share the file with the checkpoints or the model in Keras '.h5' format?
How to restore the DNN stero model to test it?
I am doing it from google colab and this is what I do:
Throws the following error when run restore:
ValueError: Can't load save_path when it is None.
According to an internet search it is due to the absence of checkpoint files in the folder.Can you share the file with the checkpoints or the model in Keras '.h5' format?