Ошибка при работе с массивами python

При попытке загрузки данных возникает ошибка: ValueError: could not broadcast input array from shape (2,) into shape (1,) Сама функция загрузки:

class data_loader():
    def __init__(self, frame_len, feature_name, N_classes):       
        self.feature_names = feature_name  
        self.N_Feature = 1 
        self.frame_length  = frame_len     
        self.hop_size = frame_len//2       
        self.N_classes  = 1        
        self.batch_size = 128          
        
        self.label_encoder = OneHotEncoder(sparse=False)
        
        self.add_cols = False # Flag 
        self.result_list = [] # List with pandas cols
        self.adding_lis = []  # List with pandas cols with added to result_list
        
        self.noise_flag = True # Flag 
        self.dsnr = 25         # SNR in dB between noise and data
        
        self.chop_flag = False # Flag
        self.chop_rate = 0.02  
def load_data(self, files, label):
        # load data
        for i in range(len(files)):
            # read file
            tmp_data = pd.read_csv(files[i])
        
            # add cols if add_cols=True
            if self.add_cols:
                tmp_data = self.add_columns(tmp_data, self.result_list, self.adding_list)
                        
            # calculate number of frames
            N_Blocks = 1 + (np.shape(tmp_data)[0]-self.frame_length)//self.hop_size

            # temporary data storage
            tmp_feature_mat = np.zeros((N_Blocks, self.frame_length, self.N_Feature))# dim [NBlocks, frame length, N feature]
    
            # temporay label storage
            tmp_label_vec = np.zeros((N_Blocks, self.N_classes)) # dim [N Blocks, N Classes]
            for j in range(N_Blocks):
                tmp_label_vec[j, :] = label[i, :] # dim [N Blocks, N Classes]
        
            # loop over feature
            for idf, feat in enumerate(self.feature_names):
                # create frame matrix out of time series
                frame_matrix = self.framing(tmp_data[feat].to_numpy())  # dim [NBlocks, frame length]
                #frame_matrix = self.z_score_normalization(frame_matrix)
                tmp_feature_mat[:, :, idf] =  frame_matrix # dim [NBlocks, frame length, Nfeature]
    
            # create feature and lable matrix with all data
            if i == 0: 
                feature_matrix = tmp_feature_mat # dim [NBlocks, frame length, Nfeature]
                label_matrix = tmp_label_vec     # dim [N Blocks, NClasses]
            else:
                feature_matrix = np.append(feature_matrix, tmp_feature_mat, axis=0) # dim [NBlocks, frame length, Nfeature]
                label_matrix = np.append(label_matrix, tmp_label_vec, axis=0)       # dim [NBlocks, NClasses]
        
        return feature_matrix, label_matrix

А ошибка возникает при выполнении этой части:

files_train, files_valid, y_train, y_valid = train_test_split(files[5:], label_frame, test_size=0.2, random_state=0)

loader1 = data_loader(frame_length, Feature, N_classes)
loader1.load_trainings_data(files_train, y_train)
loader1.load_validation_data(files_valid, y_valid)
loader1.batch_size = batch_size

loader2 = data_loader(frame_length, ['gravity.x'], N_classes)
loader2.add_cols = True
loader2.result_list = ['gravity.x']
loader2.adding_list = ['gravity.x']
loader2.load_trainings_data(files_train, y_train)
loader2.load_validation_data(files_valid, y_valid)
loader2.batch_size = batch_size