Use local RandomState instead of seeding the global RNG (keras-team#1…

…2259)

* Use local RandomState instead of seeding the global RNG

* Create a unit test module for datasets and move tests there

* Move initializer test to the proper file

keras/datasets/boston_housing.py

@@ -30,9 +30,9 @@ def load_data(path='boston_housing.npz', test_split=0.2, seed=113):
         x = f['x']
         y = f['y']
 
-    np.random.seed(seed)
+    rng = np.random.RandomState(seed)
     indices = np.arange(len(x))
-    np.random.shuffle(indices)
+    rng.shuffle(indices)
     x = x[indices]
     y = y[indices]
 

keras/datasets/imdb.py

@@ -59,14 +59,14 @@ def load_data(path='imdb.npz', num_words=None, skip_top=0,
         x_train, labels_train = f['x_train'], f['y_train']
         x_test, labels_test = f['x_test'], f['y_test']
 
-    np.random.seed(seed)
+    rng = np.random.RandomState(seed)
     indices = np.arange(len(x_train))
-    np.random.shuffle(indices)
+    rng.shuffle(indices)
     x_train = x_train[indices]
     labels_train = labels_train[indices]
 
     indices = np.arange(len(x_test))
-    np.random.shuffle(indices)
+    rng.shuffle(indices)
     x_test = x_test[indices]
     labels_test = labels_test[indices]
 

keras/datasets/reuters.py

@@ -56,9 +56,9 @@ def load_data(path='reuters.npz', num_words=None, skip_top=0,
     with np.load(path) as f:
         xs, labels = f['x'], f['y']
 
-    np.random.seed(seed)
+    rng = np.random.RandomState(seed)
     indices = np.arange(len(xs))
-    np.random.shuffle(indices)
+    rng.shuffle(indices)
     xs = xs[indices]
     labels = labels[indices]
 

keras/initializers.py

@@ -248,9 +248,10 @@ def __call__(self, shape, dtype=None):
             num_rows *= dim
         num_cols = shape[-1]
         flat_shape = (num_rows, num_cols)
+        rng = np.random
         if self.seed is not None:
-            np.random.seed(self.seed)
-        a = np.random.normal(0.0, 1.0, flat_shape)
+            rng = np.random.RandomState(self.seed)
+        a = rng.normal(0.0, 1.0, flat_shape)
         u, _, v = np.linalg.svd(a, full_matrices=False)
         # Pick the one with the correct shape.
         q = u if u.shape == flat_shape else v

tests/keras/datasets/datasets_test.py

@@ -0,0 +1,90 @@
+import tempfile
+
+import numpy as np
+import pytest
+
+from keras.datasets import boston_housing
+from keras.datasets import imdb
+from keras.datasets import reuters
+
+
+@pytest.fixture
+def fake_downloaded_boston_path(monkeypatch):
+    num_rows = 100
+    num_cols = 10
+    rng = np.random.RandomState(123)
+
+    x = rng.randint(1, 100, size=(num_rows, num_cols))
+    y = rng.normal(loc=100, scale=15, size=num_rows)
+
+    with tempfile.NamedTemporaryFile('wb', delete=True) as f:
+        np.savez(f, x=x, y=y)
+        monkeypatch.setattr(boston_housing, 'get_file',
+                            lambda *args, **kwargs: f.name)
+        yield f.name
+
+
+@pytest.fixture
+def fake_downloaded_imdb_path(monkeypatch):
+    train_rows = 100
+    test_rows = 20
+    seq_length = 10
+    rng = np.random.RandomState(123)
+
+    x_train = rng.randint(1, 100, size=(train_rows, seq_length))
+    y_train = rng.binomial(n=1, p=0.5, size=train_rows)
+    x_test = rng.randint(1, 100, size=(test_rows, seq_length))
+    y_test = rng.binomial(n=1, p=0.5, size=test_rows)
+
+    with tempfile.NamedTemporaryFile('wb', delete=True) as f:
+        np.savez(f, x_train=x_train, y_train=y_train, x_test=x_test, y_test=y_test)
+        monkeypatch.setattr(imdb, 'get_file', lambda *args, **kwargs: f.name)
+        yield f.name
+
+
+@pytest.fixture
+def fake_downloaded_reuters_path(monkeypatch):
+    num_rows = 100
+    seq_length = 10
+    rng = np.random.RandomState(123)
+
+    x = rng.randint(1, 100, size=(num_rows, seq_length))
+    y = rng.binomial(n=1, p=0.5, size=num_rows)
+
+    with tempfile.NamedTemporaryFile('wb', delete=True) as f:
+        np.savez(f, x=x, y=y)
+        monkeypatch.setattr(reuters, 'get_file', lambda *args, **kwargs: f.name)
+        yield f.name
+
+
+def test_boston_load_does_not_affect_global_rng(fake_downloaded_boston_path):
+    np.random.seed(1337)
+    before = np.random.randint(0, 100, size=10)
+
+    np.random.seed(1337)
+    boston_housing.load_data(path=fake_downloaded_boston_path, seed=9876)
+    after = np.random.randint(0, 100, size=10)
+
+    assert np.array_equal(before, after)
+
+
+def test_imdb_load_does_not_affect_global_rng(fake_downloaded_imdb_path):
+    np.random.seed(1337)
+    before = np.random.randint(0, 100, size=10)
+
+    np.random.seed(1337)
+    imdb.load_data(path=fake_downloaded_imdb_path, seed=9876)
+    after = np.random.randint(0, 100, size=10)
+
+    assert np.array_equal(before, after)
+
+
+def test_reuters_load_does_not_affect_global_rng(fake_downloaded_reuters_path):
+    np.random.seed(1337)
+    before = np.random.randint(0, 100, size=10)
+
+    np.random.seed(1337)
+    reuters.load_data(path=fake_downloaded_reuters_path, seed=9876)
+    after = np.random.randint(0, 100, size=10)
+
+    assert np.array_equal(before, after)

tests/keras/initializers_test.py

@@ -104,6 +104,18 @@ def test_orthogonal(tensor_shape):
             target_mean=0.)
 
 
+def test_orthogonal_init_does_not_affect_global_rng():
+    np.random.seed(1337)
+    before = np.random.randint(0, 100, size=10)
+
+    np.random.seed(1337)
+    init = initializers.orthogonal(seed=9876)
+    init(shape=(10, 5))
+    after = np.random.randint(0, 100, size=10)
+
+    assert np.array_equal(before, after)
+
+
 @pytest.mark.parametrize('tensor_shape',
                          [(100, 100), (10, 20), (30, 80), (1, 2, 3, 4)],
                          ids=['FC', 'RNN', 'RNN_INVALID', 'CONV'])