Human's raw data file format from .cool to .mcool

[wy52940]

When I try to use "cooler zoomify --balance filename.cool" to convert .cool files into .mcool files, I met the error: "INFO:cooler.balance:variance is nan". I think it's the reason why I got this result after I ran PrepareData_linear.py:
The test target length is:torch.Size([970, 1, 64, 64])
tensor([[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan],
...,
[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan]])

The test data length is:torch.Size([970, 1, 64, 64])
tensor([[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan],
...,
[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan],
[nan, nan, nan, ..., nan, nan, nan]])

And I cannot perform other pretrain .py files which need call PrepareData_linear.py.
So could I ask what kind of preprocessing steps have you performed on cell1(2,3)_10kb_contacts.cool to handle this NAN value.