Keras code that always gives a 0

eval output.txt

[[32849     0]
 [ 7215     0]]

evaluate.py

from sklearn.metrics import confusion_matrix
from keras.preprocessing.text import Tokenizer
from keras.preprocessing.sequence import pad_sequences
from keras.models import load_model
from sklearn.model_selection import train_test_split
def load(file):
    X=[]
    Y=[]
    with open(file,encoding="utf-8") as f:
        for line in f.readlines():
            X.append(line.lower().strip("\n").split("|")[0])
            Y.append(int(line.lower().strip("\n").split("|")[1]))
    return X,Y
model=load_model('spaces.h5')
X,Y = load("formatted.csv")
T_2 = Tokenizer(filters='',char_level=True)
T_2.fit_on_texts(X)
X=T_2.texts_to_sequences(X)
X=pad_sequences(X, maxlen=1200)
y_pred = model.predict(X)
y_pred = (y_pred > 0.5)
cm = confusion_matrix(Y, y_pred)
print(cm)

train.py

from keras.models import Sequential
from keras.layers import Dense
from keras.preprocessing.text import Tokenizer
from keras.preprocessing.sequence import pad_sequences
from sklearn.model_selection import train_test_split
import numpy as np
from keras import callbacks
def load(file):
    X=[]
    Y=[]
    with open(file,encoding="utf-8") as f:
        for line in f.readlines():
            X.append(line.lower().strip("\n").split("|")[0])
            Y.append(int(line.lower().strip("\n").split("|")[1]))
    return X,Y
X,Y = load("formatted.csv")
T_2 = Tokenizer(filters='',char_level=True)
T_2.fit_on_texts(X)
X=T_2.texts_to_sequences(X)
X=pad_sequences(X, maxlen=1200)
x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, random_state=42)
model = Sequential()
model.add(Dense(256, input_dim=1200, activation='sigmoid'))
model.add(Dense(128, activation='sigmoid'))
model.add(Dense(26, activation='sigmoid'))
model.add(Dense(1, activation='relu'))

model.compile(loss="binary_crossentropy", optimizer="adam", metrics=['accuracy'])
earlystop=callbacks.EarlyStopping(monitor='val_loss',
                                  min_delta=0,
                                  patience=5,
                                  verbose=0, mode='auto')
callbacks_list=[earlystop]
model.fit(x_train, y_train, epochs=100, batch_size=20, validation_data=(x_test, y_test),callbacks=callbacks_list)
scores = model.evaluate(X, Y)
print("\n%s: %.2f%%" % (model.metrics_names[1], scores[1]*100))
model.save('model.h5')

import pickle
with open('tokenizer.pickle', 'wb') as handle:
    pickle.dump(T_2, handle, protocol=pickle.HIGHEST_PROTOCOL)