Training not working with example midi files

[tonfeld]

Hi, hope somebody can help?
the code from the Train notebook ( converted to python)
gives an error, when executing following line:
data = MusicDataBunch.from_files(midi_files, data_path, processors=processors, bs=2, bptt=12)
numpy is version 2.0.1 ( as dependency of fastai 1.0.61)

setting an array element with a sequence. The requested array has an inhomogeneous shape after 2 dimensions. The detected shape was (18, 2) + inhomogeneous part.
Stack trace:

File "D:\data\vs23\midgen\musicautobot\music_transformer\dataloader.py", line 94, in process (Current frame)
super().process(ds)
File "D:\data\vs23\midgen\musicautobot\music_transformer\dataloader.py", line 39, in from_files
.label_const(label_cls=LMLabelList))
File "D:\data\vs23\midgen\notebooks\music_transformer\Train.py", line 55, in
data = MusicDataBunch.from_files(midi_files, data_path, processors=processors, bs=2, bptt=12)
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 2 dimensions. The detected shape was (18, 2) + inhomogeneous part.

Thanks in advance!