update error explanation in parameters fold in tutorials train_controlnet.ipynb (#1974)

Fixes # .

Description
The current documentation incorrectly states that "fold 0 is used for
training" in the prepare_maisi_controlnet_json_dataloader function. In
reality, the function uses the fold parameter to determine the
validation set based on matching fold indices. Specifically, if a
dataset item's fold value matches the fold value specified in
config_maisi_controlnet_train.json, it is assigned to the validation
set; otherwise, it is used for training.

Changes
Updated Documentation:
Revised the description of the fold parameter to accurately reflect its
functionality.
Clarified that in the current parameter settings, fold 0 is designated
for validation and fold 1 is designated for training.

---------

Signed-off-by: MaybeRichard <lisicheng168@gmail.com>
Co-authored-by: Eric Kerfoot <17726042+ericspod@users.noreply.github.com>