I propose four corrections for minor problems that observed in my use of Deformetrica.
The first one interrupts the optimization process before the result are written:
- scipy.optimize.OptimizeResults.message is a string, using "decode('utf-8')" raise an exception that interrupts the execution. It is probably for an old version of scipy (I don't know which one).
The second leads to incorrect results in Shooting:
- the parameters t0, t1 are not set, and only the default values are used.
The third and the fourth only affect the logs and are less critical.
- log files are automatically removed by clean up stage of longitudinal atlas
- a deprecation warning of torch pollutes the logs.
I am of course available for any further clarification, testing, or better formatting.
Thanks to all the deformetrica team!