c='k',
fname='{}/assemblies{}'.format(results_dir, i / 2))
n_runs = 10
feasibility_func = import_module('{}.feasibility'.format(
args.data)).check_feasibility
prc_mean, prc_err = ci_prc(n_runs, model.synthesize_assemblies,
feasibility_func, n_points)
print('Precision for assembly: %.3f +/- %.3f' % (prc_mean, prc_err))
mmd_mean, mmd_err = ci_mmd(n_runs, model.synthesize_assemblies, X_test)
rdiv_mean, rdiv_err = ci_rdiv(n_runs, X_test, model.synthesize_assemblies)
basis = 'cartesian'
cons_mean, cons_err = ci_cons(n_runs,
model.synthesize_assemblies,
latent_dim,
bounds,
basis=basis)
res = {
'CSS': [prc_mean, prc_err],
'MMD': [mmd_mean, mmd_err],
'R-Div': [rdiv_mean, rdiv_err],
'LSC': [cons_mean, cons_err]
}
json.dump(res, open('{}/results.json'.format(results_dir), 'w'))
results_mesg_0 = 'Precision for assembly: %.3f +/- %.3f' % (prc_mean,
prc_err)
results_mesg_1 = 'Maximum mean discrepancy for assembly: %.4f +/- %.4f' % (
mmd_mean, mmd_err)
n_runs = 10
feasibility_func = import_module('{}.feasibility'.format(
args.data)).check_feasibility
prc_mean, prc_err = ci_prc(n_runs, model.synthesize_assemblies,
feasibility_func, n_points)
print('Precision for assembly: %.3f +/- %.3f' % (prc_mean, prc_err))
mmd_mean, mmd_err = ci_mmd(n_runs, model.synthesize_assemblies, X_test)
rdiv_mean, rdiv_err = ci_rdiv(n_runs, X_test, model.synthesize_assemblies)
if args.data == 'SC':
basis = 'polar'
else:
basis = 'cartesian'
cons0_mean, cons0_err = ci_cons(n_runs,
model.synthesize_x0,
latent_dim[0],
bounds,
basis='cartesian')
cons1_mean, cons1_err = ci_cons(n_runs,
model.synthesize_x1,
latent_dim[1],
bounds,
basis=basis)
res = {
'CSS': [prc_mean, prc_err],
'MMD': [mmd_mean, mmd_err],
'R-Div': [rdiv_mean, rdiv_err],
'LSC_A': [cons0_mean, cons0_err],
'LSC_B': [cons1_mean, cons1_err]
}
for i in range(len(n_points)):
assemblies_list.append(assemblies[:25, n_points_c[i]:n_points_c[i+1]])
plot_samples(None, assemblies_list, scatter=False, alpha=.7, c='k', fname='{}/assemblies'.format(results_dir))
n_runs = 10
feasibility_func = import_module('{}.feasibility'.format(args.data)).check_feasibility
prc_mean, prc_err = ci_prc(n_runs, model.synthesize_assemblies, feasibility_func, n_points)
print('Precision for assembly: %.3f +/- %.3f' % (prc_mean, prc_err))
mmd_mean, mmd_err = ci_mmd(n_runs, model.synthesize_assemblies, X_test)
rdiv_mean, rdiv_err = ci_rdiv(n_runs, X_test, model.synthesize_assemblies)
if args.data == 'SCCC':
basis = 'polar'
else:
basis = 'cartesian'
cons0_mean, cons0_err = ci_cons(n_runs, model.synthesize_x0, latent_dim[0], bounds, basis='cartesian')
cons1_mean, cons1_err = ci_cons(n_runs, model.synthesize_x1, latent_dim[1], bounds, basis=basis)
cons2_mean, cons2_err = ci_cons(n_runs, model.synthesize_x2, latent_dim[2], bounds, basis=basis)
cons3_mean, cons3_err = ci_cons(n_runs, model.synthesize_x3, latent_dim[3], bounds, basis=basis)
cons4_mean, cons4_err = ci_cons(n_runs, model.synthesize_x4, latent_dim[4], bounds, basis=basis)
res = {'CSS': [prc_mean, prc_err],
'MMD': [mmd_mean, mmd_err],
'R-Div': [rdiv_mean, rdiv_err],
'LSC_A': [cons0_mean, cons0_err],
'LSC_B': [cons1_mean, cons1_err],
'LSC_C': [cons2_mean, cons2_err],
'LSC_D': [cons3_mean, cons3_err],
'LSC_E': [cons4_mean, cons4_err]}
json.dump(res, open('{}/results.json'.format(results_dir), 'w'))