search

pyannote / pyannote-pipeline

Tunable pipelines
Other
26 stars 11 forks source link

Error when using a pruner in the Optimizer #52

Open arxaqapi opened 1 month ago

arxaqapi commented 1 month ago

The following error is encountered Trial.should_prune() takes 1 positional argument but 2 were given when running an Optimizer instance created using a pruner like seen below:

optimizer = Optimizer(pipeline, sampler="TPESampler", pruner="SuccessiveHalvingPruner")

Here is a MRE to display the error: MRE_pyannote_pipeline_optuna.ipynb

The optuna version used in the MRE is optuna==3.6.1

FrenchKrab commented 3 weeks ago

I think the changes in your repository fully fix this issue, you should submit a pull request ! EDIT: nevermind i'm blind