1

有没有重新训练 SavedModel 的例子？在许多地方，他们声称这是可能的，而不是使用检查点，但没有提供示例。当我尝试执行时，模型的变量保持不变：

...
model_save_path = "test.pb"

with tf.Session(graph=tf.Graph()) as net:
        ...
        for e in range(epochs):
            # Train the model
            ...
            builder = saved_model.builder.SavedModelBuilder(model_save_path)    
            signature = predict_signature_def(inputs={'myInput': X, 'errorInput': Y},
                                                  outputs={'myOutput': out, 'errorOutput': mse})
            builder.add_meta_graph_and_variables(sess=net,
                                 tags=[tag_constants.TRAINING],
                                 signature_def_map={'predict': signature})
            builder.save()

            print(error)

上面的代码训练模型，将每个交互存储在模型中并打印相关的错误。该代码有一个错误正在改善的输出：

2773.6885

291.35968

263.40912

255.27612

当我们再次加载它并尝试训练它时，错误保持不变：

...
# Load the model
model_save_path = "test.pb"
loaded = tf.saved_model.load(net, ["train"], model_save_path)
graph = tf.get_default_graph()
...    

with tf.Session(graph=tf.Graph()) as net:
        ...
        for e in range(epochs):
            # Train the model
            ...

            builder = saved_model.builder.SavedModelBuilder(model_save_path)    
            signature = predict_signature_def(inputs={'myInput': X, 'errorInput': Y},
                                                  outputs={'myOutput': out, 'errorOutput': mse})
            builder.add_meta_graph_and_variables(sess=net,
                                 tags=[tag_constants.TRAINING],
                                 signature_def_map={'predict': signature})
            builder.save()

            print(error)

输出始终是初始训练的错误：

255.27612

255.27612

255.27612

255.27612

4

0 回答 0