Task $2$: Time series classification using deep learning $2$

as follows:

Convolutional neural networks for time series classification $(2017)$

Time Series Classification from Scratch with Deep Neural Networks: A Strong Baseline $(2017)$

You should use at least two Tensorflow callbacks when you fit your model. These can be built$-$in ones or your personalized callback. If you use Torch, explain how you implement the equivalent operations.

You should run the model at least $10$ epochs.

You can use the same processed datasets in Task $1.$ For CNNClassifier, as you cannot use tf$.$ Data, you may put the training set directly.

For CNNClassifier, you can run it with the default parameters or reduce the number of epoch $($default is $2000).$

Please use the predefined test dataset to report the test scores.

$vv$

Task $3$: Time series classification using deep learning $3$

Next, you can try to further improve your model by selecting two of the following ideas:

Check this post to get inspired.

report $($UCR repository has a specific dataset type such as AUDIO or MOTION$).$ You could try to crop or pad the time series if you would like to match the sizes.

chose.