From 744df7e6b518780fa1c12a38210d48063e25e02d Mon Sep 17 00:00:00 2001 From: Darkshades Date: Thu, 1 Jun 2023 22:49:15 +0100 Subject: [PATCH] readme edit --- readme.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/readme.md b/readme.md index 64476f6..24c9ecf 100644 --- a/readme.md +++ b/readme.md @@ -82,7 +82,7 @@ sampled_model = metatune.build_sampled_model(study.best_trial)
-**NOTE:** A thing to note is that for large datasets, hyperparameter search ought to be used on a subset of the data rather than the entire dataset. The reason for this being that the aim of hyperparameter tuning is not fit all of the data, but rather to sample the parameters that best model the data. After tuning, you can instantiate the sampled model with the corresponding sampled parameters and fine-tune the model on the large dataset, this way you avoid the unnecessary computation required to model the entirety of the available dataset. +**NOTE:** A thing to note is that for large datasets, model and hyperparameter search ought to be used on a subset of the data rather than the entire dataset. The reason for this being that the aim of hyperparameter tuning is not TO fit all of the data, but rather to sample the parameters that best model the data. After tuning, you can instantiate the sampled model with the corresponding sampled parameters and fine-tune the model on the large dataset, this way you avoid the unnecessary computation required to model the entirety of the available dataset.