Add test case for ImmutableContainer

tests/test_ray/test_searchers.py

@@ -4,6 +4,7 @@
 from mmtune.ray.searchers import (SEARCHERS, AxSearch, BlendSearch, CFOSearch,
                                   HyperOptSearch, NevergradSearch,
                                   TrustRegionSearcher, build_searcher)
+from mmtune.utils.container import ImmutableContainer
 
 
 def test_build_searcher():
@@ -21,6 +22,11 @@ def config():
         steps=10, width=tune.uniform(0, 20), height=tune.uniform(-100, 100))
 
 
+@pytest.fixture
+def config_im(config):
+    return ImmutableContainer(config)
+
+
 @pytest.fixture
 def trainable():
 
@@ -44,6 +50,16 @@ def test_ax(trainable, config):
         config=config)
 
 
+def test_ax_im(trainable, config_im):
+    tune.run(
+        trainable,
+        metric='mean_loss',
+        mode='min',
+        search_alg=AxSearch(),
+        num_samples=2,
+        config=config_im)
+
+
 def test_blend(trainable, config):
     tune.run(
         trainable,
@@ -54,6 +70,16 @@ def test_blend(trainable, config):
         config=config)
 
 
+def test_blend_im(trainable, config_im):
+    tune.run(
+        trainable,
+        metric='mean_loss',
+        mode='min',
+        search_alg=BlendSearch(),
+        num_samples=2,
+        config=config_im)
+
+
 def test_cfo(trainable, config):
     tune.run(
         trainable,
@@ -64,6 +90,16 @@ def test_cfo(trainable, config):
         config=config)
 
 
+def test_cfo_im(trainable, config_im):
+    tune.run(
+        trainable,
+        metric='mean_loss',
+        mode='min',
+        search_alg=CFOSearch(),
+        num_samples=2,
+        config=config_im)
+
+
 def test_hyperopt(trainable, config):
     tune.run(
         trainable,
@@ -74,6 +110,16 @@ def test_hyperopt(trainable, config):
         config=config)
 
 
+def test_hyperopt_im(trainable, config_im):
+    tune.run(
+        trainable,
+        metric='mean_loss',
+        mode='min',
+        search_alg=HyperOptSearch(),
+        num_samples=2,
+        config=config_im)
+
+
 def test_nevergrad(trainable, config):
     tune.run(
         trainable,
@@ -84,6 +130,16 @@ def test_nevergrad(trainable, config):
         config=config)
 
 
+def test_nevergrad_im(trainable, config_im):
+    tune.run(
+        trainable,
+        metric='mean_loss',
+        mode='min',
+        search_alg=NevergradSearch(optimizer='PSO', budget=2),
+        num_samples=2,
+        config=config_im)
+
+
 def test_trust_region(trainable, config):
     tune.run(
         trainable,
@@ -92,3 +148,13 @@ def test_trust_region(trainable, config):
         search_alg=TrustRegionSearcher(),
         num_samples=2,
         config=config)
+
+
+def test_trust_region_im(trainable, config_im):
+    tune.run(
+        trainable,
+        metric='mean_loss',
+        mode='min',
+        search_alg=TrustRegionSearcher(),
+        num_samples=2,
+        config=config_im)