inference_rect.model_param_init()
inference_sphere1.model_param_init()
inference_sphere2.model_param_init()
params_rect = inference_rect.learning(n_restarts=10)
params_sphere1 = inference_sphere1.learning(n_restarts=10)
inference_sphere2.matrix_update(model_sphere1.param_to_vec())
if ndim == 2:
x1_grid, x2_grid = np.meshgrid(np.linspace(-1, 1, 50),
np.linspace(-1, 1, 50))
x_pred_points = Variable(
torch.from_numpy(
np.vstack([x1_grid.flatten(),
x2_grid.flatten()]).astype(np.float32)).t())
pred_mean_rect, pred_var_rect = inference_rect.predict(x_pred_points)
pred_std_rect = pred_var_rect**0.5
acq_rect = acquisition(x_pred_points,
inference_rect,
params_rect,
reference=reference)
pred_mean_sphere1, pred_var_sphere1 = inference_sphere1.predict(
x_pred_points)
pred_mean_sphere2, pred_var_sphere2 = inference_sphere2.predict(
x_pred_points)
pred_std_sphere1 = pred_var_sphere1**0.5
pred_std_sphere2 = pred_var_sphere2**0.5
acq_sphere1 = acquisition(x_pred_points,
inference_sphere1,
params_sphere1,
inference_normal = Inference((x_input, output), model_normal)
inference_shadow = ShadowInference((x_input, output), model_shadow)
inference_normal.init_parameters()
inference_shadow.init_parameters()
params_normal = inference_normal.learning(n_restarts=5)
inference_shadow.cholesky_update(model_normal.param_to_vec())
if ndim == 2:
x1_grid, x2_grid = np.meshgrid(np.linspace(-1, 1, 50),
np.linspace(-1, 1, 50))
x_pred_points = Variable(
torch.from_numpy(
np.vstack([x1_grid.flatten(),
x2_grid.flatten()]).astype(np.float32)).t())
pred_mean_normal, pred_var_normal = inference_normal.predict(
x_pred_points)
pred_std_normal = pred_var_normal**0.5
acq_normal = acquisition(x_pred_points,
deepcopy_inference(inference_normal,
params_normal),
reference=reference)
pred_mean_shadow, pred_var_shadow = inference_shadow.predict(
x_pred_points)
pred_std_shadow = pred_var_shadow**0.5
acq_shadow = acquisition(x_pred_points,
deepcopy_inference(inference_shadow,
params_normal),
reference=reference)
# ShadowInference unit test