yphacker / ai_yanxishe Goto Github PK

def main(op): if op == 'train': train_df = pd.read_csv('../data/train.csv') print(train_df['Sports'].value_counts()) train_df['filename'] = train_df['filename'].apply(lambda x: '../data/train/{0}'.format(x)) if mode == 1: n_splits = 5 x = train_df['filename'].values y = train_df['label'].values skf = StratifiedKFold(n_splits=n_splits, random_state=0, shuffle=True) for fold_idx, (train_idx, val_idx) in enumerate(skf.split(x, y)): train(train_df.iloc[train_idx], train_df.iloc[val_idx], fold_idx)
在main.py的第162行，在读取y值时，为什么直接y = train_df['label'].values呢，数据集的csv是有三个属性的呀，这个该如何修改进行多标签分类呢
我试过给label改成Sports单属性，但会报错输入中含有NaN