解决Tensorflow ValueError: Found unexpected losses or metrics that do not correspond to any ... 问题

最新推荐文章于 2025-12-06 21:36:52 发布

原创最新推荐文章于 2025-12-06 21:36:52 发布 · 837 阅读

11 ·

CC 4.0 BY-SA版权

文章标签：

#tensorflow #人工智能 #python

深度学习专栏收录该内容

4 篇文章

订阅专栏

文章描述了在使用Tensorflow进行模型训练时遇到的ValueError，原因在于fit函数中的losses和metrics定义与模型输出层不匹配。解决方法是确保loss和metrics键对应于网络最后一层的名称。给出了编译模型和设置回调的代码示例。

部署运行你感兴趣的模型镜像

@[TOC](解决Tensorflow ValueError: Found unexpected losses or metrics that do not correspond to any Model output: dict_keys([‘xx’, ‘xx’]). Valid mode output names: [‘xx’, ‘xx’]. Received struct is: {‘xx’: ‘mse’, ‘xx’: ‘mse’}. 问题)

报错部分 fit函数

...
model = tf.keras.models.Model(inputs = [clean_input_layer, noise_input_layer], outputs = [clean_d, noise_d])
model.summary()

optimizer = tf.keras.optimizers.Adam(learning_rate = 0.001)
model.compile(
    optimizer = optimizer,
    loss = {'output1': 'mse', 'output2': 'mse'},
    metrics = {'output1': 'accuracy', 'output2': 'accuracy'})


es = tf.keras.callbacks.EarlyStopping(monitor='loss', mode='min', verbose=1, patience=5)
mc = tf.keras.callbacks.ModelCheckpoint('./best_model.h5', monitor='loss', mode='min', save_best_only=True)

history = model.fit(
    x = [clean, noise], y = [clean, noise], callbacks=[es, mc],
    epochs = 50, batch_size = 128, shuffle = True
)
...