train_gen loads indices of all time series instead of batch-size if self.permute is False

[aparna-aketi]

In the train_gen function of TimeSeriesdata class (data_loader.py), the tsidx is set to np.arange(num_ts) instead of something like tsidx = np.arange(batch_idx, min(batch_start + self.batch_size, num_ts)) when self.permute = False. Is this intentional or a bug?