Log_scale

[Aseman7]

Excuse me, I have a question about log_scale too. When Model is called, the output cell gives a value for the log_scale of the fit too. Does it matter whether or not its value is near zero in order to have a good fit? What my model gives is: 

SpectrumModel

Data: Spectrum
Emulator: F_SPEX_emu
Log Likelihood: 14126.359545518464

Parameters
T: 2970.0
logg: 5.1
Z: 0.0
log_scale: -24.216672955323997 (fit)