Animations with PyALAF

It is very simple to create animations for 1D models (with only one feature and one objective) in PyALAF. This is demonstrated below for a very simple function. The animations can be helpful for illustrating how different acquisition functions select data.

import numpy as np
from PyALAF.models import inv_sphere
import matplotlib.pyplot as plt

random_state = 420

model_sphere = inv_sphere(d=1, random_state=random_state)
grid = np.linspace(-5,5,100).reshape(-1,1)

y = model_sphere.evaluate(grid)

from PyALAF.animation import create_animation_continuous

from sklearn.gaussian_process import GaussianProcessRegressor as GPR
from sklearn.gaussian_process.kernels import RBF

kernel = RBF()
gpr = GPR(kernel)

#Bayesian Optimization
acquisition_function = 'ideal'

anim = create_animation_continuous(
                        model=model_sphere,
                        gpr=gpr,
                        acquisition_function=acquisition_function,
                        grid_simple=grid,
                        n_iterations=20,
                        opt_method='PSO',
                        n_observations=3,
                        noise_level=0,
                        alpha=1,
                        random_state=42,
                        legend=True,
                        plot_std=True
)

Disabled warnings
Disabled warnings

2026-01-30 14:43:08,810 - matplotlib.animation - INFO - Animation.save using <class 'matplotlib.animation.HTMLWriter'>

anim

Once Loop Reflect