Maximizing the log likelihood in Training section in Published paper

[sege01]

Hi @kashif ,

I was referring to the paper and in the training section it has the below equation where we maximize the log-likelihood

)
where D is the batch of time series.
Could you please help in understanding why D is used as |D| ? What does this indicate?

Thank you.

[kashif]

ah yes what i mean here is the size of the batch so $\mathcal{D}$ is the batch and $|\mathcal{D}|$ is the batch size... hope that helps? I guess its a fancy way of saying loss.mean() 😄

[sege01]

Thank you so much @kashif . :)