labels = le.fit_transform(data["names"])
num_classes = len(np.unique(labels))
labels = labels.reshape(-1, 1)
one_hot_encoder = OneHotEncoder(categorical_features=[0])
labels = one_hot_encoder.fit_transform(labels).toarray()
embeddings = np.array(data["embeddings"])
# Initialize Softmax training model arguments
BATCH_SIZE = 32
EPOCHS = 20
input_shape = embeddings.shape[1]
# Build sofmax classifier
softmax = SoftMax(input_shape=(input_shape, ), num_classes=num_classes)
model = softmax.build()
# Create KFold
cv = KFold(n_splits=5, random_state=42, shuffle=True)
history = {'acc': [], 'val_acc': [], 'loss': [], 'val_loss': []}
# Train
for train_idx, valid_idx in cv.split(embeddings):
X_train, X_val, y_train, y_val = embeddings[train_idx], embeddings[
valid_idx], labels[train_idx], labels[valid_idx]
his = model.fit(X_train,
y_train,
batch_size=BATCH_SIZE,
epochs=EPOCHS,
verbose=1,
validation_data=(X_val, y_val))
print(his.history['acc'])
X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = 0.20, random_state=123) X_train, X_test, y_train__, y_test__ = train_test_split(X,y__, test_size = 0.20, random_state=123) # softmax model = [] input_shape = X.shape[1] model_svm = pickle.loads(open(args.folder + "svm.pickle", "rb").read()) model_rf = pickle.loads(open(args.folder + "rf.pickle", "rb").read()) model_gb = pickle.loads(open(args.folder + "gb.pickle", "rb").read()) model_vt = pickle.loads(open(args.folder + "vt.pickle", "rb").read()) softmax = SoftMax(input_shape=(input_shape,), num_classes=num_classes) models = softmax.build() model_s1 = load_model(args.folder + "best_fold_1.hdf5") model_s2 = load_model(args.folder + "best_fold_2.hdf5") model_s3 = load_model(args.folder + "best_fold_3.hdf5") model_s4 = load_model(args.folder + "best_fold_4.hdf5") model_s5 = load_model(args.folder + "best_fold_5.hdf5") model.append(model_s1) model.append(model_s2) model.append(model_s3) model.append(model_s4) model.append(model_s5) model.append(model_svm) model.append(model_rf)
