Hi, SISSO_predict will read in coefficients right after the descriptor, and the error occurs in your case because no coefficients to read after the 1D descriptor in the SISSO.out. I see the SISSO.out output is in a mess. I didn't see this before and don't know what caused this. What if you run SISSO again with just one core?

Model/descriptor for generating residual:
coefficients_001: 0.7195685272E-02
coefficients_001: -0.1425151870E-02
Intercept_001: 0.3607139813E+01
coefficients_001: 0.5149789758E-01
Intercept_001: 0.3651808281E+01
coefficients_001: 0.1518181025E-02
Intercept_001: 0.3558224874E+01
coefficients_001: 0.3887855296E-01
Intercept_001: 0.3705094158E+01
coefficients_001: 0.3067298966E-02
Intercept_001: 0.3662611429E+01

      Intercept_001:     0.3677778981E+01

1D descriptor (model):
Total RMSE,MaxAE: 0.106794 0.340241
@@@descriptor:
7:[((Z_A+Z_B)/rP_B)]
Total number of features in the space phi01: 646
Total number of features in the space phi02: 690942
Size of the SIS-selected subspace from phi02: 100
Wall-clock time (second) for this FC: 0.54

from sisso.

Thanks for the quick response! I noticed an error in my python code that starts SISSO, after fixing that the problem did not appear anymore, so it might be fixed already. If it reappears later despite the adjustments, I will write here and try out running it on one core. Regards

from sisso.