def test_hybrid_discriminative_rbm_fit():
train = torchvision.datasets.KMNIST(
root="./data",
train=True,
download=True,
transform=torchvision.transforms.ToTensor(),
)
new_hybrid_discriminative_rbm = discriminative_rbm.HybridDiscriminativeRBM(
n_visible=784,
n_hidden=128,
n_classes=10,
learning_rate=0.1,
alpha=0.01,
momentum=0,
decay=0,
use_gpu=False,
)
loss, acc = new_hybrid_discriminative_rbm.fit(train,
batch_size=128,
epochs=1)
assert loss >= 0
assert acc >= 0
def test_hybrid_discriminative_rbm_class_sampling():
new_hybrid_discriminative_rbm = discriminative_rbm.HybridDiscriminativeRBM(
n_classes=10)
h = torch.ones(1, 128)
probs, states = new_hybrid_discriminative_rbm.class_sampling(h)
assert probs.size(1) == 10
assert states.size(1) == 10
def test_hybrid_discriminative_rbm_hidden_sampling():
new_hybrid_discriminative_rbm = discriminative_rbm.HybridDiscriminativeRBM(
n_classes=10)
v = torch.ones(1, 128)
y = torch.ones(128, 10)
probs, states = new_hybrid_discriminative_rbm.hidden_sampling(v, y)
assert probs.size(1) == 128
assert states.size(1) == 128
def test_hybrid_discriminative_rbm_alpha_setter():
new_hybrid_discriminative_rbm = discriminative_rbm.HybridDiscriminativeRBM(
)
try:
new_hybrid_discriminative_rbm.alpha = 'a'
except:
new_hybrid_discriminative_rbm.alpha = 0.01
assert new_hybrid_discriminative_rbm.alpha == 0.01
try:
new_hybrid_discriminative_rbm.alpha = -1
except:
new_hybrid_discriminative_rbm.alpha = 0.01
assert new_hybrid_discriminative_rbm.alpha == 0.01
def test_hybrid_discriminative_rbm_alpha():
new_hybrid_discriminative_rbm = discriminative_rbm.HybridDiscriminativeRBM(
)
assert new_hybrid_discriminative_rbm.alpha == 0.01