Add explanations of seed to HEBO

optuna · Dec 17, 2024 · 685e4fa · 685e4fa
1 parent f5285e0
commit 685e4fa
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 2 deletions.
diff --git a/package/samplers/hebo/README.md b/package/samplers/hebo/README.md
@@ -37,7 +37,7 @@ pip install --upgrade pymoo
     HEBOSampler(search_space=search_space)
     ```
 
-  - `seed`: Seed for random number generator.
+  - `seed`: A seed for the initialization of `HEBOSampler`. Please note that the Bayesian optimization part is not deterministic even if seed is fixed due to the backend implementation.
 
   - `constant_liar`: If `True`, penalize running trials to avoid suggesting parameter configurations nearby. Default is `False`.
 

diff --git a/package/samplers/hebo/sampler.py b/package/samplers/hebo/sampler.py
@@ -37,7 +37,9 @@ class HEBOSampler(optunahub.samplers.SimpleBaseSampler):
             By specifying search_space, the sampling speed at each iteration becomes slightly quicker, but this argument is not necessary to run this sampler. Default is :obj:`None`.
 
         seed:
-            A seed for ``HEBOSampler``. Default is :obj:`None`.
+            A seed for the initialization of ``HEBOSampler``. Default is :obj:`None`.
+            Please note that the Bayesian optimization part is not deterministic even if seed is
+            fixed due to the backend implementation.
 
         constant_liar:
             If :obj:`True`, penalize running trials to avoid suggesting parameter configurations