Does darpa subT dataset support mini-batch on delora?

[JaySlamer]

Hi, thanks for your great work.
I trained with default hyperparameters, and can not get any sensible result.
I noticed the default batch_size is 1 in config file, which is very likely to cause unreliable and unstable result of training. The comment says:

batch_size > 1 currently only supported if single image dims are used (vertical and horizontal cells)
and
In general: larger batches currently implemented rather primitively

I don't understand what the comment means and whether darpa subT could be trained with larger batch size.
After change batch size to 32, I got error when training 1st epoch to 99.8%
Below is the error message
Traceback (most recent call last):
File "/home/slam/Documents/delora/bin/run_training.py", line 94, in
trainer.train()
File "/home/slam/Documents/delora/src/deploy/trainer.py", line 122, in train
epoch_losses = self.train_epoch(epoch=epoch, dataloader=dataloader)
File "/home/slam/Documents/delora/src/deploy/trainer.py", line 71, in train_epoch
self.step(
File "/home/slam/Documents/delora/src/deploy/deployer.py", line 292, in step
preprocessed_dict = preprocessed_dicts[batch_index]
IndexError: list index out of range