« Back to top page

Vizier Sampler

Vizier Sampler of OptunaHub

Abstract

This package provides a sampler using Google Vizier. It is a simple wrapper around Vizier’s Python client, enabling its optimization capabilities within the Optuna framework.

Class or Function Names

  • VizierSampler

Installation

This package currently supports Python 3.10–3.13.

pip install -r https://hub.optuna.org/samplers/vizier/requirements.txt

Example

import optuna
from optuna.visualization import plot_hypervolume_history
import optunahub


wfg = optunahub.load_module("benchmarks/wfg")
wfg4 = wfg.Problem(function_id=4, n_objectives=2, dimension=5, k=1)


vizier = optunahub.load_module("samplers/vizier")

study = optuna.create_study(
    study_name="Vizier",
    sampler=vizier.VizierSampler(search_space=wfg4.search_space),
    directions=wfg4.directions,
)
study.optimize(wfg4, n_trials=100)

plot_hypervolume_history(study, reference_point=[0.5, 0.5])

Benchmarks

Following images are the benchmark of Vizier and other samplers.

Rosenbrock Function Gallagher’s Gaussian 101-me Peaks Function HPOLib WFG 4 WFG 8

Others

This sampler is a wrapper for Google Vizier. See the official Google Vizier documentation for more details.

Package
samplers/vizier
Author
gen740
License
MIT License
Verified Optuna version
  • 4.3.0
Dependencies (.txt)
  • google-vizier[jax]==0.1.24
  • jax==0.4.34
  • jaxlib==0.4.34
Last update
2026-08-17
Discussions & Issues
Create a discussion
Create a bug report