Bugs Please fill the below form if possible.
When creating an extended PDF, the normalization and observable range seem to be changed
Current Behaviour
I would expect only the yield to be attached to the PDF, instead the normalization range seems to also be changed
import zfit
obs = zfit.Space('x', limits=(-10, 10))
nrm = zfit.Space('x', limits=( 0, 10))
mu = zfit.Parameter('mu', 0.0)
sigma = zfit.Parameter('sigma', 1.0)
yield_ = zfit.Parameter('yield', 1000.0)
gauss = zfit.pdf.Gauss(mu=mu, sigma=sigma, obs=obs, norm=nrm)
print(gauss.space)
print(gauss.norm )
gauss_extended = gauss.create_extended(yield_)
print(gauss_extended.space)
print(gauss_extended.norm )
Expected Behaviour
The PDF should be the same, except for the normalization
Context (Environment)
- zfit version: 0.28
- Python version: 3.13
- Are you using conda, pipenv, etc? : micromamba
- Operating System: Almalinux9
- Tensorflow version: 2.20.0
Possible Solution/Implementation
set_yield is a workaround, but the behavior is wrong and will give the wrong results in the user code.
Bugs Please fill the below form if possible.
When creating an extended PDF, the normalization and observable range seem to be changed
Current Behaviour
I would expect only the yield to be attached to the PDF, instead the normalization range seems to also be changed
Expected Behaviour
The PDF should be the same, except for the normalization
Context (Environment)
Possible Solution/Implementation
set_yieldis a workaround, but the behavior is wrong and will give the wrong results in the user code.