New methods and scripts for dataset generation
This merge request changes the way we generate our data:
- Split generation of field configurations and observables (generate configurations first, calculate observables and labels afterwards)
- Fixed random seeds for reproducibility
- Addition of conjugated plaquettes and Polyakov loops (as well as their conjugates) to the set of
\mathcal{W}variables - Minor improvements to the core of the simulation (memory allocation)
- Restore original normalization of topological charge density
Edited by David Mueller