snb | Simple hierarchical notebook

I got several compile errors due to the following:

• Mismatched begin/end due to inconsistent indentation.
• Undeclared signals: enabled, q6
• Lack of whitespace around ? operators.

When you don't understand error messages, try your code on other simulators, such as the ones on edaplayground.

This code compiles without errors for me:

One approach uses a correlated subquery to choose the correct non NULL value to report from the SNB_Kurs table:

MCnet model has a generator and a discriminator. Generator is LSTM based and hence there is no problem in loading of the weights by varying the number of timesteps T. However the discriminator, as they've coded it, is convolutional. To apply convolutional layers on video, they are concatenating the frames in channel dimension. With K=4,T=7, you get a video of length 11 with 3 channels. When you concatenate them along channel dimension, you get an image with 33 channels. When they define discriminator, they define the first layer of discriminator to have 33 input channels and hence the weights have similar dimension. But with K=4,T=1, video length is 5 and the final image has 15 channels and so the weights would have 15 channels. This is the mismatch error you're observing. To fix this, you can pick weights from the first 15 channels only (for lack of a better way I can think of). Code below:

You can use @PreDestroy annotation.

The @PreDestroy annotation is used on methods as a callback notification to signal that the instance is in the process of being removed by the container.