Finishing up DBN tests.

tests/learnergy/models/test_dbn.py

@@ -0,0 +1,232 @@
+import pytest
+import torch
+import torchvision
+
+from learnergy.models import dbn
+
+
+def test_dbn_n_visible():
+    new_dbn = dbn.DBN()
+
+    assert new_dbn.n_visible == 128
+
+
+def test_dbn_n_visible_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.n_visible = 'a'
+    except:
+        new_dbn.n_visible = 1
+
+    assert new_dbn.n_visible == 1
+
+    try:
+        new_dbn.n_visible = 0
+    except:
+        new_dbn.n_visible = 1
+
+    assert new_dbn.n_visible == 1
+
+
+def test_dbn_n_hidden():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.n_hidden) == 1
+
+
+def test_dbn_n_hidden_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.n_hidden = 'a'
+    except:
+        new_dbn.n_hidden = [128]
+
+    assert len(new_dbn.n_hidden) == 1
+
+
+def test_dbn_n_layers():
+    new_dbn = dbn.DBN()
+
+    assert new_dbn.n_layers == 1
+
+
+def test_dbn_n_layers_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.n_layers = 0
+    except:
+        new_dbn.n_layers = 1
+
+    assert new_dbn.n_layers == 1
+
+    try:
+        new_dbn.n_layers = 'a'
+    except:
+        new_dbn.n_layers = 1
+
+    assert new_dbn.n_layers == 1
+
+
+def test_dbn_steps():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.steps) == 1
+
+
+def test_dbn_steps_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.steps = 'a'
+    except:
+        new_dbn.steps = [1]
+
+    assert len(new_dbn.steps) == 1
+
+    try:
+        new_dbn.steps = [1, 1]
+    except:
+        new_dbn.steps = [1]
+
+    assert len(new_dbn.steps) == 1
+
+
+def test_dbn_lr():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.lr) == 1
+
+
+def test_dbn_lr_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.lr = 'a'
+    except:
+        new_dbn.lr = [0.1]
+
+    assert len(new_dbn.lr) == 1
+
+    try:
+        new_dbn.lr = [0.1, 0.1]
+    except:
+        new_dbn.lr = [0.1]
+
+    assert len(new_dbn.lr) == 1
+
+
+def test_dbn_momentum():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.momentum) == 1
+
+
+def test_dbn_momentum_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.momentum = 'a'
+    except:
+        new_dbn.momentum = [0]
+
+    assert len(new_dbn.momentum) == 1
+
+    try:
+        new_dbn.momentum = [0, 0]
+    except:
+        new_dbn.momentum = [0]
+
+    assert len(new_dbn.momentum) == 1
+
+
+def test_dbn_decay():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.decay) == 1
+
+
+def test_dbn_decay_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.decay = 'a'
+    except:
+        new_dbn.decay = [0]
+
+    assert len(new_dbn.decay) == 1
+
+    try:
+        new_dbn.decay = [0, 0]
+    except:
+        new_dbn.decay = [0]
+
+    assert len(new_dbn.decay) == 1
+
+
+def test_dbn_T():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.T) == 1
+
+
+def test_dbn_T_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.T = 'a'
+    except:
+        new_dbn.T = [0]
+
+    assert len(new_dbn.T) == 1
+
+    try:
+        new_dbn.T = [0, 0]
+    except:
+        new_dbn.T = [0]
+
+    assert len(new_dbn.T) == 1
+
+
+def test_dbn_models():
+    new_dbn = dbn.DBN()
+
+    assert len(new_dbn.models) == 1
+
+
+def test_dbn_models_setter():
+    new_dbn = dbn.DBN()
+
+    try:
+        new_dbn.models = 'a'
+    except:
+        new_dbn.models = []
+
+    assert len(new_dbn.models) == 0
+
+
+def test_dbn_fit():
+    train = torchvision.datasets.MNIST(
+        root='./data', train=True, download=True, transform=torchvision.transforms.ToTensor())
+
+    new_dbn = dbn.DBN(n_visible=784, n_hidden=[128, 128], steps=[1, 1],
+                      learning_rate=[0.1, 0.1], momentum=[0, 0], decay=[0, 0], temperature=[1, 1], use_gpu=False)
+
+    e, pl = new_dbn.fit(train, batch_size=128, epochs=[1, 1])
+
+    assert len(e) == 2
+    assert len(pl) == 2
+
+def test_dbn_reconstruct():
+    test = torchvision.datasets.MNIST(
+        root='./data', train=False, download=True, transform=torchvision.transforms.ToTensor())
+
+    new_dbn = dbn.DBN(n_visible=784, n_hidden=[128, 128], steps=[1, 1],
+                      learning_rate=[0.1, 0.1], momentum=[0, 0], decay=[0, 0], temperature=[1, 1], use_gpu=False)
+
+    e, v = new_dbn.reconstruct(test)
+
+    assert e >= 0
+    assert v.size(1) == 784