def test_mlp(X, Y):
dbn_network: UnsupervisedDBN = UnsupervisedDBN.load(DBN_NETWORK_FILENAME)
mlp_network: MLPClassifier = joblib.load(MLP_NETWORK_FILENAME)
first_positions = dbn_network.transform(X[:, 0])
second_positions = dbn_network.transform(X[:, 1])
X = np.concatenate((first_positions, second_positions), axis=1)
return mlp_network.score(X, Y)
def predict_mlp(first, second):
First = np.array(first)
Second = np.array(second)
dbn_network: UnsupervisedDBN = UnsupervisedDBN.load(DBN_NETWORK_FILENAME)
mlp_network: MLPClassifier = joblib.load(MLP_NETWORK_FILENAME)
first_position = dbn_network.transform(First)
second_position = dbn_network.transform(Second)
X = np.concatenate((first_position, second_position))
result = mlp_network.predict(X.reshape(1, -1))
return result.flatten().tolist()
def __init__(self, load=False):
if load:
self.bands = pickle.load(open("bands.pickle", "rb"))
self.dbn = UnsupervisedDBN.load("dbn.pickle")
self.prepca = pickle.load(open("prepca.pickle", "rb"))
self.pca = pickle.load(open("pca.pickle", "rb"))
else:
self.dbn = UnsupervisedDBN(hidden_layers_structure=[50, 50, 50],
batch_size=1024,
learning_rate_rbm=0.001,
n_epochs_rbm=5,
contrastive_divergence_iter=2,
activation_function='sigmoid')
self.bands = [
{"start": 0, "end": 150, "pca_components":30, "energy_factor": 5},
{"start": 120, "end":300, "pca_components":70, "energy_factor": 1},
{"start": 250, "end":513, "pca_components":60, "energy_factor": 0.5}
]
self.prepca = [PCA(n_components=band["pca_components"]) for band in self.bands]
self.pca = PCA(n_components=16)
def train_mlp(X, Y):
dbn_network = UnsupervisedDBN.load(DBN_NETWORK_FILENAME)
first_positions = dbn_network.transform(X[:, 0])
second_positions = dbn_network.transform(X[:, 1])
X = np.concatenate((first_positions, second_positions), axis=1)
print(f'Shape after Concatenation = {X.shape}')
mlp_network = MLPClassifier(hidden_layer_sizes=(90, 45, 20, 2),
learning_rate='constant',
learning_rate_init=1e-2,
max_iter=1000)
mlp_network.fit(X, Y)
joblib.dump(mlp_network, MLP_NETWORK_FILENAME)
def train_dbn(X):
dbn_network = UnsupervisedDBN(
hidden_layers_structure=[337, 260, 170, 85, 45],
activation_function='relu',
learning_rate_rbm=5 * 1e-3,
n_epochs_rbm=200)
dbn_network.fit(X)
dbn_network.save(DBN_NETWORK_FILENAME)
class DBNFeatureExtractor(object):
def __init__(self, load=False):
if load:
self.bands = pickle.load(open("bands.pickle", "rb"))
self.dbn = UnsupervisedDBN.load("dbn.pickle")
self.prepca = pickle.load(open("prepca.pickle", "rb"))
self.pca = pickle.load(open("pca.pickle", "rb"))
else:
self.dbn = UnsupervisedDBN(hidden_layers_structure=[50, 50, 50],
batch_size=1024,
learning_rate_rbm=0.001,
n_epochs_rbm=5,
contrastive_divergence_iter=2,
activation_function='sigmoid')
self.bands = [
{"start": 0, "end": 150, "pca_components":30, "energy_factor": 5},
{"start": 120, "end":300, "pca_components":70, "energy_factor": 1},
{"start": 250, "end":513, "pca_components":60, "energy_factor": 0.5}
]
self.prepca = [PCA(n_components=band["pca_components"]) for band in self.bands]
self.pca = PCA(n_components=16)
def save(self):
self.dbn.save("dbn.pickle")
pickle.dump(self.bands, open("bands.pickle", "wb"))
pickle.dump(self.prepca, open("prepca.pickle", "wb"))
pickle.dump(self.pca, open("pca.pickle", "wb"))
def normalize_energy(self, banded_frames):
bands_energy = [band["energy_factor"] * np.sum(np.square(banded_frames[bandIndex]))/(band["end"] - band["start"])
for bandIndex, band in enumerate(self.bands)]
return [bf/math.sqrt(be) for bf,be in zip(banded_frames, bands_energy)]
def train(self, frames):
banded_frames = [self.prepca[bandIndex].fit_transform(frames[:,band["start"]:band["end"]])
for bandIndex, band in enumerate(self.bands)]
banded_frames = self.normalize_energy(banded_frames)
frames = np.concatenate(banded_frames, axis=1)
print(frames.shape)
self.dbn.fit(frames)
raw_features = self.dbn.transform(frames)
self.pca.fit(raw_features)
def extractFeatures(self, frames):
banded_frames = [self.prepca[bandIndex].transform(frames[:,band["start"]:band["end"]])
for bandIndex, band in enumerate(self.bands)]
banded_frames = self.normalize_energy(banded_frames)
frames = np.concatenate(banded_frames, axis=1)
raw_features = self.dbn.transform(frames)
return self.pca.transform(raw_features)
if __name__ == '__main__':
# Load the dataset
(X_train, Y_train), (_, _) = mnist.load_data()
X_train, Y_train = shuffle(X_train, Y_train, random_state=1000)
width = X_train.shape[1]
height = X_train.shape[2]
X = X_train[0:nb_samples].reshape(
(nb_samples, width * height)).astype(np.float32) / 255.0
Y = Y_train[0:nb_samples]
# Train the unsupervised DBN
unsupervised_dbn = UnsupervisedDBN(hidden_layers_structure=[512, 256, 64],
learning_rate_rbm=0.05,
n_epochs_rbm=100,
batch_size=64,
activation_function='sigmoid')
X_dbn = unsupervised_dbn.fit_transform(X)
# Perform t-SNE
tsne = TSNE(n_components=2, perplexity=20, random_state=1000)
X_tsne = tsne.fit_transform(X_dbn)
# Show the result
fig, ax = plt.subplots(figsize=(18, 10))
colors = [cm.tab10(i) for i in Y]
for i in range(nb_samples):