pyro_model_guide_6.py

def print_progress():
  alphas = pyro.param("alphas")
  betas = pyro.param("betas")
  
  if torch.cuda.is_available():
    alphas.cuda()
    betas.cuda()

  means = alphas / (alphas + betas)
  normalized_means = means / torch.sum(means)
  factors = betas / (alphas * (1.0 + alphas + betas))
  stdevs = normalized_means * torch.sqrt(factors)

  tiger_pays_string = "probability Tiger pays: {0:.3f} +/- {1:.2f}".format(normalized_means[0], stdevs[0])
  jason_pays_string = "probability Jason pays: {0:.3f} +/- {1:.2f}".format(normalized_means[1], stdevs[1])
  james_pays_string = "probability James pays: {0:.3f} +/- {1:.2f}".format(normalized_means[2], stdevs[2])
  print("[", step, "|", tiger_pays_string, "|", jason_pays_string, "|", james_pays_string, "]")