def __init__(self,args): BaseTupleLoader.__init__(self,args) self.img_path = args['db_path'] + '/images/' lbls = self.data_df['label'] lbl2idx = np.sort(np.unique(lbls)) self.lbl2idx_dict = {k: v for v, k in enumerate(lbl2idx)} self.final_lbls = [self.lbl2idx_dict[x] for x in list(lbls.values)] self.num_classes = len(self.lbl2idx_dict.keys()) print(self.__class__.__name__, ' Data size ', self.data_df.shape[0], 'Num lbls', len(self.lbl2idx_dict.keys()))
def __init__(self,args=None): BaseTupleLoader.__init__(self,args) self.img_path = config.db_path + '/jpg/' lbls = self.data_df['label'] lbl2idx = np.sort(np.unique(lbls)) self.lbl2idx_dict = {k: v for v, k in enumerate(lbl2idx)} self.final_lbls = [self.lbl2idx_dict[x] for x in list(lbls.values)] self.num_classes = len(self.lbl2idx_dict.keys()) print('Data size ', self.data_df.shape[0], 'Num lbls', len(self.lbl2idx_dict.keys()))
def __init__(self, args): BaseTupleLoader.__init__(self) csv_file = args['csv_file'] self.data_df = pd.read_csv(config.db_path + csv_file) self.img_path = config.db_path + '/' lbls = self.data_df['label'] lbl2idx = np.sort(np.unique(lbls)) self.lbl2idx_dict = {k: v for v, k in enumerate(lbl2idx)} self.final_lbls = [self.lbl2idx_dict[x] for x in list(lbls.values)] self.num_classes = len(self.lbl2idx_dict.keys()) self.data_permutation = np.random.permutation(self.data_df.shape[0]) self.data_idx = 0 print('Data size ', self.data_df.shape[0], 'Num lbls', len(self.lbl2idx_dict.keys()))