You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For learning rate schedulers, there is a code that requires steps_per_epoch entry. The current code looks like this:
extra_logs[hyperparam] = float(value((epoch+1) * self.config_info["steps_per_epoch"])) ## here replace with self.params.get('steps')
Replace this part with a call that enables to get the steps_per_epoch term. See if self.params.get("steps") works.
The text was updated successfully, but these errors were encountered:
johnypark
changed the title
Train_Manager.SaveModelHistory(tf.keras.callbacks.Callback) change steps_per_epoch term
Change Train_Manager.SaveModelHistory(tf.keras.callbacks.Callback) steps_per_epoch term
May 4, 2022
For learning rate schedulers, there is a code that requires steps_per_epoch entry. The current code looks like this:
Replace this part with a call that enables to get the steps_per_epoch term. See if self.params.get("steps") works.
The text was updated successfully, but these errors were encountered: