din模型训练成功并且能保存pb模型但是无法加载训练好的pb模型

[hjl150505]

Please refer to the FAQ in doc and search for the related issues before you ask the question.

Describe the question(问题描述) 运行： if name == "main": x, y, feature_columns, behavior_feature_list = get_xy_fd() model = DIN(feature_columns, behavior_feature_list)

model = BST(feature_columns, behavior_feature_list,att_head_num=4)

model.compile('adam', 'binary_crossentropy',
              metrics=['binary_crossentropy'])
history = model.fit(x, y, verbose=1, epochs=10, validation_split=0.5)  # 训练成功
model.save("./mydinModel")  # 可以正常保存
reloadModel = models.load_model("mydinModel")  # 此处报错

报错信息： `ValueError: Could not find matching function to call loaded from the SavedModel. Got: Positional arguments (2 total):

• Tensor("x:0", shape=(None, 1, 8), dtype=float32)
• None Keyword arguments: {'training': True}

Expected these arguments to match one of the following 8 option(s):

Option 1: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 8), dtype=tf.float32, name='x')
• TensorSpec(shape=(None, 1), dtype=tf.bool, name='mask') Keyword arguments: {'training': True}

Option 2: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 4), dtype=tf.float32, name='x')
• TensorSpec(shape=(None, 1), dtype=tf.bool, name='mask') Keyword arguments: {'training': False}

Option 3: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 10), dtype=tf.float32, name='x')
• None Keyword arguments: {'training': True}

Option 4: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 4), dtype=tf.float32, name='x')
• TensorSpec(shape=(None, 1), dtype=tf.bool, name='mask') Keyword arguments: {'training': True}

Option 5: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 4), dtype=tf.float32, name='x')
• None Keyword arguments: {'training': False}

Option 6: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 8), dtype=tf.float32, name='x')
• TensorSpec(shape=(None, 1), dtype=tf.bool, name='mask') Keyword arguments: {'training': False}

Option 7: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 4), dtype=tf.float32, name='x')
• None Keyword arguments: {'training': True}

Option 8: Positional arguments (2 total):

• TensorSpec(shape=(None, 1, 10), dtype=tf.float32, name='x')
• None Keyword arguments: {'training': False}

Process finished with exit code 1 `

Additional context Add any other context about the problem here.

Operating environment(运行环境):

• windows10 操作系统
• python version [3.6]
• tensorflow version [2.1.0]
• deepctr version [0.9.1]

[shuDaoNan9]

其他模型也有类似情况，PB格式模型用tensorflow.python.keras.models.load_model加载的时候就报这个，用Java等语言加载则正常；如果还是用Python处理的话H5格式模型我这边加载正常。

[chentao169]

https://github.com/shenweichen/DeepCTR/blob/master/docs/source/FAQ.md pass in custom_objects when loading model load_model(model_path, custom_objects)