Replies: 1 comment 3 replies
-
|
Hi, I'm not sure if you're looking for an answer in here in terms of sktime or keras. In Keras, you just need to do something like: early_stopper = EarlyStopping(monitor='val_loss', mode='min', verbose=True, patience=8, restore_best_weights=True)
reduce_lr_callback = ReduceLROnPlateau(monitor='val_loss', factor=0.5, min_lr=1e-8, patience=5, verbose=True)
callbacks = [reduce_lr_callback, early_stopper]
history = autoencoder.fit(callbacks=callbacks)Here I'm invoking EarlyStopping and setting it to monitor 'val_loss'. It will stop training when validation loss hits a minimum. (taken from a simple demo here: https://github.com/BlankAdventure/snippets/blob/main/Python/autoenc_denoise.py) |
Beta Was this translation helpful? Give feedback.
-
|
Hello Thank you for your answer, i was looking for an answer in terms of SKTime. I know that i can monitor val_loss with EarlyStopping from Keras but since the SKTime Neural Network Classifier's fit function does not take validation data as an input, I am little confused on how to do it. This is how I am trying to deal with it at the moment but it would better if I could use validation data for the purpose. I am new at using SKTime and Keras. It would be great if I could see the learning curve. Best |
Beta Was this translation helpful? Give feedback.
-
|
Maybe @fkiraly can comment, but, if some new features are needed for this I would be open to contributing. |
Beta Was this translation helpful? Give feedback.
-
|
Hello, I'm having the same trouble as you. I want to train a InceptionTime Classifier with sktime but the .fit method does not accept validation data as an input, so if you found the answer, I would be happy to see it. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Hello,
I'm looking for guidance on monitoring validation loss when using the SKTime CNN for class methods. I am currently using CNN for multivariate classification. Specifically, I’m trying to find a way to fit the model and track the loss on validation data throughout the epochs. I know I can use Keras EarlyStopping, but it only monitors training loss. Is there a method to obtain the learning curve for validation loss or to prevent the model from overfitting to the training data? The current fit method does not take validation data as a parameter.
Best,
Ece
Beta Was this translation helpful? Give feedback.
All reactions