我编写了一个自定义生成器来加载我的数据集并将其馈送到 fit_generator 方法中。但我遇到了一个错误。question_train、img_lis_train 和 answers_train 都是字符串列表。我希望 mygen 能返回每批32个样本，格式为：

[images,questions] , answers

这是代码：

def mygen(questions_train,img_lis_train,answers_train):    start = 0      data_size = len(questions_train)    batch_size = 32            while True:                  if( start+batch_size <= data_size ):            batch_ques = questions_train[ start : start+batch_size ]             batch_ans = answers_train[ start : start+batch_size ]             batch_img_names = img_lis_train[ start : start+batch_size ]         elif(start < data_size):            batch_ques = questions_train[ start : ]             batch_ans = answers_train[ start : ]             batch_img_names = img_lis_train[ start : start+batch_size ]         else:            start = 0            continue               batch_img = []        for img_name in batch_img_names:            img = load_img('./dataset/images/' + str(img_name) + '.png' , target_size = (224,224))            img = img_to_array(img)               batch_img.append( preprocess_input(img) )            start += 32        print('start = ' + str(start))        yield [batch_img, batch_ques] ,batch_ansfc_model.fit_generator(mygen, steps_per_epoch=3376, epochs = 10)

这是我得到的错误：

  File "mycode.py", line 210, in <module>    fc_model.fit_generator(mygen, steps_per_epoch=3376, epochs = 10)  File "/opt/apps/Python-3.5.1/lib/python3.5/site-packages/keras/legacy/interfaces.py", line 87, in wrapper    return func(*args, **kwargs)  File "/opt/apps/Python-3.5.1/lib/python3.5/site-packages/keras/models.py", line 1223, in fit_generator    initial_epoch=initial_epoch)  File "/opt/apps/Python-3.5.1/lib/python3.5/site-packages/keras/legacy/interfaces.py", line 87, in wrapper    return func(*args, **kwargs)  File "/opt/apps/Python-3.5.1/lib/python3.5/site-packages/keras/engine/training.py", line 2083, in fit_generator    generator_output = next(output_generator)StopIterationException ignored in: <bound method BaseSession.__del__ of <tensorflow.python.client.session.Session object at 0x7f936527fcc0>>Traceback (most recent call last):  File "/opt/apps/Python-3.5.1/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 712, in __del__  File "/opt/apps/Python-3.5.1/lib/python3.5/site-packages/tensorflow/python/framework/c_api_util.py", line 31, in __init__TypeError: 'NoneType' object is not callable

回答：

你需要通过调用生成器并传递你的数据来构建生成器，否则它如何生成批次？这是正确的方法：

fc_model.fit_generator(mygen(q_train, i_train, a_train), ...)