Getting configuration not found when doing .fit()

[evsumantad]

[mfeurer]

Thanks a lot for reporting this issue. Unfortunately, this is not sufficient for us to detect the underlying problem. Could you please provide the information asked for in the issue template?

[makslevental]

I'm having the same issue. Can you tell me where this error comes from in the source so I can try to debug it?

Here's the entire log https://gist.github.com/makslevental/67c72fd634fc4d87ea5c73fb019c85bb

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs for the next 7 days. Thank you for your contributions.

[franchuterivera]

Hello, by looking at the error msgs:

[DEBUG] [2021-04-14 13:10:54,986:Client-AutoMLSMBO(1)::b355d252-9d4c-11eb-9785-08f1ea94010a] Error reading configurations: Trying to set illegal value 'gradient_boosting' (type '<class 'str'>') for hyperparameter 'classifier:__choice__, Type: Categorical, Choices: {bernoulli_nb, decision_tree, extra_trees, gaussian_nb, k_nearest_neighbors, lda, liblinear_svc, mlp, multinomial_nb, passive_aggressive, qda, random_forest}, Default: random_forest' (default-value has type '<class 'str'>').

I was wondering with what arguments you run auto-sklearn? I have also seen this problem if one runs with a version of the development branch that did not have metalearning complete. Can you please try using the latest auto-sklearn version via pip install auto-sklearn==0.12.6?

[mfeurer]

Hey, This looks like gradient boosting is not part of the search space. This is really confusing me and we should

• print the search space used in the log file
• print the auto-sklearn version (as we print all other versions) @makslevental could you please provide a full working example?

[SrMarkLau]

Hey, This looks like gradient boosting is not part of the search space. This is really confusing me and we should

• print the search space used in the log file
• print the auto-sklearn version (as we print all other versions) @makslevental could you please provide a full working example?

models = ["gradient_boosting","decision_tree","random_forest", "k_nearest_neighbors", "libsvm_svc"]

        automl = autosklearn.classification.AutoSklearnClassifier(
          include_estimators=[model],
          exclude_estimators=None,
          resampling_strategy='cv',
          resampling_strategy_arguments={'folds': num_folds},
          ensemble_size=1,
          seed=seed,
          n_jobs=4
		    )

[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 536 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 32 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 393 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 476 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 440 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 81 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 176 not found
[WARNING] [2021-08-11 09:08:48,772:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 9 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 566 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 407 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 663 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 184 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 257 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 687 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 560 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 162 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 329 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 601 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 399 not found
[WARNING] [2021-08-11 09:08:48,773:Client-AutoMLSMBO(42)::e1386b70-fa9c-11eb-8550-00155d5a7d4a] Configuration 358 not found

[Bomarjee]

Dear All,

Hmm i'm also having this same error when doing a .fit()

print(autosklearn.version) : 0.13.0 Ubuntu 20.14

import autosklearn.classification automl = autosklearn.classification.AutoSklearnClassifier( include_estimators=["random_forest","decision_tree"], exclude_estimators=None, include_preprocessors=["no_preprocessing"], exclude_preprocessors=None)

automl.fit(test_x, test_Y)

[WARNING] [2021-08-17 23:18:00,953:Client-AutoMLSMBO(1)::8b2a91c1-ff5d-11eb-b08f-2b6c8edb2cd4] Configuration 237 not found [WARNING] [2021-08-17 23:18:00,953:Client-AutoMLSMBO(1)::8b2a91c1-ff5d-11eb-b08f-2b6c8edb2cd4] Configuration 99 not found [WARNING] [2021-08-17 23:18:00,953:Client-AutoMLSMBO(1)::8b2a91c1-ff5d-11eb-b08f-2b6c8edb2cd4] Configuration 88 not found [WARNING] [2021-08-17 23:18:00,953:Client-AutoMLSMBO(1)::8b2a91c1-ff5d-11eb-b08f-2b6c8edb2cd4] Configuration 485 not found [WARNING] [2021-08-17 23:18:00,953:Client-AutoMLSMBO(1)::8b2a91c1-ff5d-11eb-b08f-2b6c8edb2cd4] Configuration 608 not found

[Rattko]

Hey, I've come across this issue a couple of times when restricting the search space using include and exclude (formerly include_estimators, exclude_estimators, etc.) parameters, i.e.

// Imports and data loading ...

automl = AutoSklearnClassifier(
    time_left_for_this_task=30,
    include={
        'classifier': ['random_forest']
    }
).fit(X, y)

A part of the output looks like this

[WARNING] [2021-09-24 11:33:19,521:Client-AutoMLSMBO(1)::73b28d22-1d1a-11ec-8dc7-acde48001122] Configuration 176 not found
[WARNING] [2021-09-24 11:33:19,521:Client-AutoMLSMBO(1)::73b28d22-1d1a-11ec-8dc7-acde48001122] Configuration 162 not found
[WARNING] [2021-09-24 11:33:19,521:Client-AutoMLSMBO(1)::73b28d22-1d1a-11ec-8dc7-acde48001122] Configuration 601 not found
[WARNING] [2021-09-24 11:33:19,521:Client-AutoMLSMBO(1)::73b28d22-1d1a-11ec-8dc7-acde48001122] Configuration 332 not found
[WARNING] [2021-09-24 11:33:19,521:Client-AutoMLSMBO(1)::73b28d22-1d1a-11ec-8dc7-acde48001122] Configuration 492 not found

I've managed to get rid of the warnings by specifying the initial_configurations_via_metalearning parameter. The following code doesn't produce any warnings.

// Imports and data loading ...

automl = AutoSklearnClassifier(
    time_left_for_this_task=30,
    include={
        'classifier': ['random_forest']
    },
    initial_configurations_via_metalearning=0
).fit(X, y)

[eddiebergman]

Note to self and @mfeurer, for debugging this we should extend the warning to include what particular configurations these are.

[asmgx]

I have same error but using autosklearn.experimental.askl2.AutoSklearn2Classifier

ValueError: Trying to set illegal value 'gradient_boosting' (type '<class 'str'>') for hyperparameter 'classifier:choice, Type: Categorical, Choices: {extra_trees, mlp, passive_aggressive, random_forest}, Default: random_forest' (default-value has type '<class 'str'>').

[stupidoge]

this issue not be solved, I still met the same. @eddiebergman Sir, could you help us?