Skip to content

Introduce the independent sampling warning template#6188

Merged
nabenabe0928 merged 5 commits intooptuna:masterfrom
nabenabe0928:code-fix/bundle-independent-sampling-warning
Jul 3, 2025
Merged

Introduce the independent sampling warning template#6188
nabenabe0928 merged 5 commits intooptuna:masterfrom
nabenabe0928:code-fix/bundle-independent-sampling-warning

Conversation

@nabenabe0928
Copy link
Copy Markdown
Contributor

@nabenabe0928 nabenabe0928 commented Jul 1, 2025

Motivation

This PR refactors the warnings related to this issue;

Description of the changes

  • Bundle the warning message in a place
  • Use the template

@nabenabe0928 nabenabe0928 changed the title Code fix/bundle independent sampling warning Introduce the independent sampling warning template Jul 1, 2025
@nabenabe0928 nabenabe0928 added the code-fix Change that does not change the behavior, such as code refactoring. label Jul 1, 2025
@nabenabe0928 nabenabe0928 added this to the v4.5.0 milestone Jul 1, 2025
@nabenabe0928
Copy link
Copy Markdown
Contributor Author

nabenabe0928 commented Jul 1, 2025

@c-bata @fusawa-yugo @sawa3030

Could you review this PR?

@nabenabe0928 nabenabe0928 marked this pull request as ready for review July 1, 2025 05:01
@nabenabe0928
Copy link
Copy Markdown
Contributor Author

nabenabe0928 commented Jul 1, 2025 

import optuna


def objective(trial: optuna.Trial) -> float:
    x = trial.suggest_float("x", -5, 5)
    y = trial.suggest_float("y", -5, 5)
    if trial.number % 2 == 1:
        z = trial.suggest_float("z", -5, 5)
    return x**2 + y**2


samplers = [
    optuna.samplers.TPESampler(multivariate=True, n_startup_trials=1),
    optuna.samplers.QMCSampler(),
    optuna.samplers.CmaEsSampler(popsize=2),
]
for sampler in samplers:
    study = optuna.create_study(sampler=sampler)
    print(sampler.__class__.__name__)
    study.optimize(objective, n_trials=10)

Obrained log

[I 2025-07-01 07:06:55,608] A new study created in memory with name: no-name-edc9a93f-6045-4c4b-b118-91ea98ffced3
TPESampler
[I 2025-07-01 07:06:55,609] Trial 0 finished with value: 0.04633574737331082 and parameters: {'x': 0.12066026736478541, 'y': 0.17826061610117172}. Best is trial 0 with value: 0.04633574737331082.
[I 2025-07-01 07:06:55,613] Trial 1 finished with value: 0.4162103448759381 and parameters: {'x': 0.39543433154491797, 'y': -0.5097470297231382, 'z': -3.982494099322823}. Best is trial 0 with value: 0.04633574737331082.
[I 2025-07-01 07:06:55,615] Trial 2 finished with value: 23.69569333242624 and parameters: {'x': -1.849094064384325, 'y': 4.502948420033801}. Best is trial 0 with value: 0.04633574737331082.
[W 2025-07-01 07:06:55,618] The parameter `z` in Trial#3 is sampled independently using `RandomSampler` instead of `TPESampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space is not supported for `multivariate=True`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `TPESampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,619] Trial 3 finished with value: 23.87845285266918 and parameters: {'x': -4.069955587808934, 'y': -2.7044249603070916, 'z': 4.794344532233806}. Best is trial 0 with value: 0.04633574737331082.
[I 2025-07-01 07:06:55,621] Trial 4 finished with value: 23.39913568479115 and parameters: {'x': 4.83655428444309, 'y': -0.08293574878386697}. Best is trial 0 with value: 0.04633574737331082.
[W 2025-07-01 07:06:55,624] The parameter `z` in Trial#5 is sampled independently using `RandomSampler` instead of `TPESampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space is not supported for `multivariate=True`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `TPESampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,625] Trial 5 finished with value: 26.696393484156438 and parameters: {'x': 3.5391371083640357, 'y': 3.7644258542780586, 'z': 0.1809634932783732}. Best is trial 0 with value: 0.04633574737331082.
[I 2025-07-01 07:06:55,628] Trial 6 finished with value: 33.66496551926208 and parameters: {'x': 3.4037224309192418, 'y': -4.698897650781435}. Best is trial 0 with value: 0.04633574737331082.
[W 2025-07-01 07:06:55,631] The parameter `z` in Trial#7 is sampled independently using `RandomSampler` instead of `TPESampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space is not supported for `multivariate=True`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `TPESampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,632] Trial 7 finished with value: 11.716454641007923 and parameters: {'x': -0.4289044229559577, 'y': -3.3959528319717194, 'z': 0.05912067715973031}. Best is trial 0 with value: 0.04633574737331082.
[I 2025-07-01 07:06:55,635] Trial 8 finished with value: 17.993389573812188 and parameters: {'x': -3.999723171433253, 'y': 1.412658531178752}. Best is trial 0 with value: 0.04633574737331082.
[W 2025-07-01 07:06:55,637] The parameter `z` in Trial#9 is sampled independently using `RandomSampler` instead of `TPESampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space is not supported for `multivariate=True`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `TPESampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,638] Trial 9 finished with value: 1.8433202543326244 and parameters: {'x': -0.04430176046212424, 'y': 1.3569663254305837, 'z': 2.2038637100298333}. Best is trial 0 with value: 0.04633574737331082.
[I 2025-07-01 07:06:55,639] A new study created in memory with name: no-name-d698e1d8-cbe9-4a80-b39c-dc01d006913b
QMCSampler
[I 2025-07-01 07:06:55,639] Trial 0 finished with value: 9.34758610558263 and parameters: {'x': -2.962222565273378, 'y': 0.7568510945805906}. Best is trial 0 with value: 9.34758610558263.
[W 2025-07-01 07:06:55,858] The parameter `z` in Trial#1 is sampled independently using `RandomSampler` instead of `QMCSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `QMCSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `QMCSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,859] Trial 1 finished with value: 50.0 and parameters: {'x': -5.0, 'y': -5.0, 'z': 0.7861537536905647}. Best is trial 0 with value: 9.34758610558263.
[I 2025-07-01 07:06:55,859] Trial 2 finished with value: 0.0 and parameters: {'x': 0.0, 'y': 0.0}. Best is trial 2 with value: 0.0.
[W 2025-07-01 07:06:55,859] The parameter `z` in Trial#3 is sampled independently using `RandomSampler` instead of `QMCSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `QMCSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `QMCSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,859] Trial 3 finished with value: 12.5 and parameters: {'x': 2.5, 'y': -2.5, 'z': -1.3242496257859058}. Best is trial 2 with value: 0.0.
[I 2025-07-01 07:06:55,860] Trial 4 finished with value: 12.5 and parameters: {'x': -2.5, 'y': 2.5}. Best is trial 2 with value: 0.0.
[W 2025-07-01 07:06:55,860] The parameter `z` in Trial#5 is sampled independently using `RandomSampler` instead of `QMCSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `QMCSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `QMCSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,860] Trial 5 finished with value: 3.125 and parameters: {'x': -1.25, 'y': -1.25, 'z': -2.7891677492803337}. Best is trial 2 with value: 0.0.
[I 2025-07-01 07:06:55,860] Trial 6 finished with value: 28.125 and parameters: {'x': 3.75, 'y': 3.75}. Best is trial 2 with value: 0.0.
[W 2025-07-01 07:06:55,860] The parameter `z` in Trial#7 is sampled independently using `RandomSampler` instead of `QMCSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `QMCSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `QMCSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,860] Trial 7 finished with value: 15.625 and parameters: {'x': 1.25, 'y': -3.75, 'z': 0.16477497373272865}. Best is trial 2 with value: 0.0.
[I 2025-07-01 07:06:55,860] Trial 8 finished with value: 15.625 and parameters: {'x': -3.75, 'y': 1.25}. Best is trial 2 with value: 0.0.
[W 2025-07-01 07:06:55,861] The parameter `z` in Trial#9 is sampled independently using `RandomSampler` instead of `QMCSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `QMCSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `QMCSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,861] Trial 9 finished with value: 13.28125 and parameters: {'x': -3.125, 'y': -1.875, 'z': 1.5290615701757346}. Best is trial 2 with value: 0.0.
[I 2025-07-01 07:06:55,861] A new study created in memory with name: no-name-2a16056d-fddc-4d3d-a64d-8854ced1a288
CmaEsSampler
[I 2025-07-01 07:06:55,861] Trial 0 finished with value: 2.7067466219200575 and parameters: {'x': -1.6198644383165108, 'y': -0.2877252567947277}. Best is trial 0 with value: 2.7067466219200575.
[W 2025-07-01 07:06:55,863] The parameter `z` in Trial#1 is sampled independently using `RandomSampler` instead of `CmaEsSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `CmaEsSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `CmaEsSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,863] Trial 1 finished with value: 7.582984484434924 and parameters: {'x': -2.1189011307632457, 'y': -1.7587616332195681, 'z': 2.8452300535487325}. Best is trial 0 with value: 2.7067466219200575.
[I 2025-07-01 07:06:55,864] Trial 2 finished with value: 1.2071233687675391 and parameters: {'x': 1.0949347287065372, 'y': 0.09078165365249191}. Best is trial 2 with value: 1.2071233687675391.
[W 2025-07-01 07:06:55,864] The parameter `z` in Trial#3 is sampled independently using `RandomSampler` instead of `CmaEsSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `CmaEsSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `CmaEsSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,864] Trial 3 finished with value: 2.906472485689223 and parameters: {'x': 1.7008223187059857, 'y': -0.11694411434876706, 'z': -2.4205343506544663}. Best is trial 2 with value: 1.2071233687675391.
[I 2025-07-01 07:06:55,865] Trial 4 finished with value: 0.011137087297300854 and parameters: {'x': 0.10549324170031227, 'y': 0.0028745874243707448}. Best is trial 4 with value: 0.011137087297300854.
[W 2025-07-01 07:06:55,866] The parameter `z` in Trial#5 is sampled independently using `RandomSampler` instead of `CmaEsSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `CmaEsSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `CmaEsSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,866] Trial 5 finished with value: 1.2644498487781717 and parameters: {'x': 0.9154506492620804, 'y': 0.652993076183666, 'z': -2.4852580524551793}. Best is trial 4 with value: 0.011137087297300854.
[I 2025-07-01 07:06:55,866] Trial 6 finished with value: 3.205089277944816 and parameters: {'x': 0.30710046913622246, 'y': -1.7637399410914094}. Best is trial 4 with value: 0.011137087297300854.
[W 2025-07-01 07:06:55,867] The parameter `z` in Trial#7 is sampled independently using `RandomSampler` instead of `CmaEsSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `CmaEsSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `CmaEsSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,867] Trial 7 finished with value: 5.105197431828331 and parameters: {'x': 1.374187557468539, 'y': 1.7935456472381688, 'z': -3.073449792285553}. Best is trial 4 with value: 0.011137087297300854.
[I 2025-07-01 07:06:55,868] Trial 8 finished with value: 0.019021640558474943 and parameters: {'x': -0.03611262490849931, 'y': -0.13310717065843214}. Best is trial 4 with value: 0.011137087297300854.
[W 2025-07-01 07:06:55,868] The parameter `z` in Trial#9 is sampled independently using `RandomSampler` instead of `CmaEsSampler`, potentially degrading the optimization performance. This fallback happend because dynamic search space and `CategoricalDistribution` are not supported by `CmaEsSampler`. You can suppress this warning by setting `warn_independent_sampling` to `False` in the constructor of `CmaEsSampler` if this independent sampling is intended behavior.
[I 2025-07-01 07:06:55,868] Trial 9 finished with value: 0.8246624426031802 and parameters: {'x': 0.8775346454097672, 'y': -0.23365656145020708, 'z': 1.6923347958401145}. Best is trial 4 with value: 0.011137087297300854.

@nabenabe0928 nabenabe0928 mentioned this pull request Jul 1, 2025
Merged
c-bata
c-bata reviewed Jul 2, 2025
View reviewed changes
Copy link
Copy Markdown
Member

@c-bata c-bata left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

sawa3030
sawa3030 approved these changes Jul 3, 2025
View reviewed changes
Copy link
Copy Markdown
Collaborator

@sawa3030 sawa3030 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@nabenabe0928 nabenabe0928 merged commit 99af8a0 into optuna:master Jul 3, 2025
14 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@c-bata c-bata c-bata approved these changes

@sawa3030 sawa3030 sawa3030 approved these changes

Assignees

@c-bata c-bata

@sawa3030 sawa3030

Labels

code-fix Change that does not change the behavior, such as code refactoring.

Projects

None yet

Milestone

v4.5.0

Development

Successfully merging this pull request may close these issues.

3 participants

@nabenabe0928 @c-bata @sawa3030