def test_opt_qn_normal(f=hp_normal):
bandit = Bandit(
{'loss': -scope.sum([f('v%i' % ii, 0, 1)
for ii in range(25)]) ** 2},
loss_target=0)
algo = TreeParzenEstimator(bandit,
prior_weight=.5,
n_startup_jobs=0,
n_EI_candidates=1,
gamma=0.15)
trials = Trials()
experiment = Experiment(trials, algo, async=False)
experiment.max_queue_len = 1
experiment.run(40)
print list(sorted(trials.losses()))
idxs, vals = miscs_to_idxs_vals(trials.miscs)
if 1:
import hyperopt.plotting
hyperopt.plotting.main_plot_vars(trials, bandit, do_show=1)
else:
import matplotlib.pyplot as plt
begin = [v[:10] for k, v in vals.items()]
end = [v[-10:] for k, v in vals.items()]
plt.subplot(2, 1, 1)
plt.title('before')
plt.hist(np.asarray(begin).flatten())
plt.subplot(2, 1, 2)
plt.title('after')
plt.hist(np.asarray(end).flatten())
plt.show()
def test_opt_qn_normal(f=hp_normal):
bandit = Bandit(
{'loss': -scope.sum([f('v%i' % ii, 0, 1) for ii in range(25)])**2},
loss_target=0)
algo = TreeParzenEstimator(bandit,
prior_weight=.5,
n_startup_jobs=0,
n_EI_candidates=1,
gamma=0.15)
trials = Trials()
experiment = Experiment(trials, algo, async=False)
experiment.max_queue_len = 1
experiment.run(40)
print list(sorted(trials.losses()))
idxs, vals = miscs_to_idxs_vals(trials.miscs)
if 1:
import hyperopt.plotting
hyperopt.plotting.main_plot_vars(trials, bandit, do_show=1)
else:
import matplotlib.pyplot as plt
begin = [v[:10] for k, v in vals.items()]
end = [v[-10:] for k, v in vals.items()]
plt.subplot(2, 1, 1)
plt.title('before')
plt.hist(np.asarray(begin).flatten())
plt.subplot(2, 1, 2)
plt.title('after')
plt.hist(np.asarray(end).flatten())
plt.show()
def ap_categorical_sampler(obs, prior_weight, upper,
size=(), rng=None, LF=DEFAULT_LF):
weights = scope.linear_forgetting_weights(scope.len(obs), LF=LF)
counts = scope.bincount(obs, minlength=upper, weights=weights)
# -- add in some prior pseudocounts
pseudocounts = counts + prior_weight
return scope.categorical(pseudocounts / scope.sum(pseudocounts),
upper=upper, size=size, rng=rng)
def ap_categorical_sampler(obs, prior_weight, upper, size=(), rng=None):
counts = scope.bincount(obs, minlength=upper)
# -- add in some prior pseudocounts
pseudocounts = counts + prior_weight
return scope.categorical(pseudocounts / scope.sum(pseudocounts),
size=size, rng=rng)
