add initialization

adnc/utils/initialization.py

@@ -0,0 +1,25 @@
+# Copyright 2018 Jörg Franke
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+import numpy as np
+import tensorflow as tf
+
+
+def unit_simplex_initialization(rng, batch_size, shape, dtype=tf.float32):
+    mat = []
+    for i in range(batch_size):
+        mat.append(rng.uniform(0, 1 / np.sum(shape), shape))
+    mat = np.array(mat)
+    return tf.constant(mat, dtype=dtype)

test/adnc/utils/test_initialization.py

@@ -0,0 +1,46 @@
+# Copyright 2018 Jörg Franke
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+import pytest
+import numpy as np
+import tensorflow as tf
+
+from adnc.utils.initialization import unit_simplex_initialization
+
+
+@pytest.fixture()
+def session():
+    with tf.Session() as sess:
+        yield sess
+    tf.reset_default_graph()
+
+
+@pytest.fixture()
+def np_rng():
+    seed = np.random.randint(1, 999)
+    return np.random.RandomState(seed)
+
+
+BATCH_SIZE = 4
+SHAPE = [2, 3]
+
+
+def test_unit_simplex_initialization(session, np_rng):
+    init_matrix = unit_simplex_initialization(np_rng, BATCH_SIZE, SHAPE, dtype=tf.float32)
+    np_init_matrix = init_matrix.eval()
+    for b in range(BATCH_SIZE):
+        tensor = np_init_matrix[b, :, :]
+        assert np.sum(tensor) <= 1
+        assert tensor.min() >= 0
+        assert tensor.max() <= 1

test/adnc/utils/test_normalization.py

@@ -19,18 +19,16 @@ import tensorflow as tf
 from adnc.utils.normalization import layer_norm
 
 
-@pytest.fixture()
-def np_rng():
-    seed = np.random.randint(1, 999)
-    return np.random.RandomState(seed)
-
-
 @pytest.fixture()
 def session():
     with tf.Session() as sess:
         yield sess
     tf.reset_default_graph()
 
+@pytest.fixture()
+def np_rng():
+    seed = np.random.randint(1, 999)
+    return np.random.RandomState(seed)
 
 def test_layer_norm(session, np_rng):
     np_weights = np_rng.normal(0, 1, [64, 128])

test/adnc/utils/test_tf_functions.py

@@ -18,12 +18,14 @@ import tensorflow as tf
 
 from adnc.utils.tf_functions import oneplus
 
+
 @pytest.fixture()
 def session():
     with tf.Session() as sess:
         yield sess
     tf.reset_default_graph()
 
+
 def test_oneplus(session):
     tf_x = tf.constant([1, 2, 3], dtype=tf.float32, )
     tf_x_oneplus = oneplus(tf_x)