Within the callbacks.py module of this repo there is a ServingCheckpoint class that extends Keras' Callback class. The idea would be to implement model checkpointing per epoch, while training.
Within train.py there is a serving_checkpoing defined that instantiated this ServingCheckpoint class, and it is added into the training loop through the array of callbacks that are defined and passed into the model at training time. However this does not function correctly. The tf.Session() closes upon execution of the array of callbacks that includes the model checkpointing.
@edhenry in which group we attach our docker? As I am a beginner to this repo I followed your video and I stuck at docker ps command. As it needs authentication. Please update me to my stupid qeustion.
Within the
callbacks.py
module of this repo there is a ServingCheckpoint class that extends Keras' Callback class. The idea would be to implement model checkpointing per epoch, while training.Within
train.py
there is a serving_checkpoing defined that instantiated this ServingCheckpoint class, and it is added into the training loop through the array of callbacks that are defined and passed into the model at training time. However this does not function correctly. The tf.Session() closes upon execution of the array of callbacks that includes the model checkpointing.