roadmap Feb 2025

[ismael-mendoza]

• ensure SNC trick + JK resampling is working on images #86
• check whether we can force SNC to work if we ran chains for longer with shape noise 0.01 and 1000 galaxies
• run exp with 0.01 shape noise and all parameters free
• redo galaxy samples chain timing with more realistic settings (flux + hlr)
• check distribution of tuned parameters
• redo galaxy sample timing with median tuned parameters, no warmup
• run shear exp with inference of logflux and logsize parameters along with shear (0.01 shape noise)
• run full noise experiment with 10^6 Gaussian galaxies and check if we meet requirement (fixed true prior)
• What image size(s) should we use for experiments?  #75