From 133acd0c4ca3652889330ee1e1278a4405f131c0 Mon Sep 17 00:00:00 2001 From: KKIEEK Date: Tue, 20 Dec 2022 13:19:42 +0900 Subject: [PATCH 1/3] Test Optuna --- configs/_base_/searcher/optuna.py | 1 + 1 file changed, 1 insertion(+) create mode 100644 configs/_base_/searcher/optuna.py diff --git a/configs/_base_/searcher/optuna.py b/configs/_base_/searcher/optuna.py new file mode 100644 index 00000000..8b712131 --- /dev/null +++ b/configs/_base_/searcher/optuna.py @@ -0,0 +1 @@ +searcher = dict(type='OptunaSearch') From fc8046c9a179ed784879ffe2f6f96f6369fb1f17 Mon Sep 17 00:00:00 2001 From: KKIEEK Date: Tue, 20 Dec 2022 15:09:10 +0900 Subject: [PATCH 2/3] Update README.md --- README.md | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 860616cd..6c7e23b9 100644 --- a/README.md +++ b/README.md @@ -19,9 +19,10 @@ SIATune is an open-source deep learning model hyperparameter tuning toolbox espe - **Support hyperparameter search algorithms** We provide hyperparameter search algorithms such as below; - - [x] [Nevergrad](https://github.com/facebookresearch/nevergrad) - - [x] [HyperOpt](https://github.com/hyperopt/hyperopt) - [x] [FLAML](https://github.com/microsoft/FLAML) + - [x] [HyperOpt](https://github.com/hyperopt/hyperopt) + - [x] [Nevergrad](https://github.com/facebookresearch/nevergrad) + - [x] [Optuna](https://github.com/optuna/optuna) - [ ] [Adaptive Experimentation (AX)](https://ax.dev/) - [ ] [Scikit-optimize](https://github.com/scikit-optimize/scikit-optimize) From 2619a93c1f57e12bcabf6aa90c977369bec8d8c1 Mon Sep 17 00:00:00 2001 From: KKIEEK Date: Tue, 20 Dec 2022 16:44:00 +0900 Subject: [PATCH 3/3] Add test code --- tests/test_hyper_optim/test_searchers.py | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/tests/test_hyper_optim/test_searchers.py b/tests/test_hyper_optim/test_searchers.py index 0eef59fe..ae5a5a79 100644 --- a/tests/test_hyper_optim/test_searchers.py +++ b/tests/test_hyper_optim/test_searchers.py @@ -89,3 +89,13 @@ def test_nevergrad(trainable, config): dict(type='NevergradSearch', optimizer='PSO', budget=1)), num_samples=2, config=config) + + +def test_optuna(trainable, config): + tune.run( + trainable, + metric='mean_loss', + mode='min', + search_alg=build_searcher(dict(type='OptunaSearch')), + num_samples=2, + config=config)