splitingdata.py

rating = []
df = pd.read_csv('..')
rating = df['target'].values
del df
'''
These values are determined by len of your dataset. If you have 25k as length then you can use 20k for training and 5k for testing.
'''
train_arrays = np.ones((20000, 250))
train_labels = np.ones(20000,dtype='int')
test_arrays = np.zeros((5000, 250))
test_labels = np.zeros(5000,dtype='int')
for i in range(20000):
    train_arrays[i] = model[i]
    train_labels[i] = rating[i]   
    
for i in range(5000):
    test_arrays[i] = model[20000+i]
    test_labels[i] = rating[20000+i]