examples

Quick Demo

Run inference with fixed hyperparameters for quick demonstration:

./infer.py --hyper-h 0.1 --hyper-sigma 0.02 --num-samples 100000 --output demo

Plot the results with only a CMB posterior applied:

./plot.py --input demo.0 --posterior CMB --show --zoom

Parallel Inference

Use the multi package to run parallel jobs to calculate the dark-energy evolution with higher statistics at fixed hyperparameters:

~/multi/multi --nohup --split 0:10:1 --run "./infer.py --seed NNN --num-samples 10000000 --output de_NNN --dark-energy"

Combine the parallel inferences:

./combine.py --input 'de_*' --output de

Plot the dark-energy evolution with all posteriors applied:

./plot.py --input de --posterior H0-LRG-Lya-CMB --show --dark-energy

Full Marginalization Calculation

Run inferences on a grid of hyperparameter values:

~/multi/multi --nohup --split 0:400:20 --run "./infer.py --seed NNN --hyper-num-h 20 --hyper-num-sigma 20 --hyper-index NNN --hyper-count 20 --output hyper20_NNN --num-samples 5000000"

Combine inferences to marginalize over hyperparameters:

./combine.py --input 'hyper20_*' --output combined20

Generate plots:

./plot.py --input combined20 --nlp --full --zoom --output plots20/