Adapted from
and
python main.py --hybrid
The original model defined in gluon has batch_size attribute
# note to self: requring batch_size in model definition is sad, not sure how to deal with this otherwise though
My implementation removes this limitation.