KERAS CUSTOM LOSS 中的错误“TypeError：传递给参数 'reduction_indices' 的值的 DataType float32 不在允许值列表中：int32，int64”答案

【问题标题】：ERROR IN KERAS CUSTOM LOSS "TypeError: Value passed to parameter 'reduction_indices' has DataType float32 not in list of allowed values: int32, int64"KERAS CUSTOM LOSS 中的错误“TypeError：传递给参数 'reduction_indices' 的值的 DataType float32 不在允许值列表中：int32，int64”
【发布时间】：2020-10-30 23:16:14
【问题描述】：

我在 Keras 中为函数定义了一个自定义损失：

(y - yhat)^2 + (y * yhat).

def customLoss(y_true, y_pred, sample_weight=None):
    y_true = K.cast(y_true, 'float32')
    y_pred = K.cast(y_pred, 'float32')
    loss = K.square(y_true - y_pred) + K.prod(y_true, y_pred)
    loss = loss * K.cast(sample_weights, 'float32')
    return loss

当我运行 model.fit 时，它在 TypeError 上失败：

earlystopping = tf.keras.callbacks.EarlyStopping(monitor='val_loss', 
                                            mode='min', verbose=1, patience=20)
history = model.fit(Xtrain, ytrain_raw, 
                    validation_data=(Xval, yval_raw), batch_size=128,
                    epochs=500, verbose=1, callbacks=[earlystopping],
                    sample_weight=sample_weights)

错误：

TypeError: in user code:

    /usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:571 train_function  *
        outputs = self.distribute_strategy.run(
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:951 run  **
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:2290 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:2649 _call_for_each_replica
        return fn(*args, **kwargs)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:533 train_step  **
        y, y_pred, sample_weight, regularization_losses=self.losses)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/compile_utils.py:205 __call__
        loss_value = loss_obj(y_t, y_p, sample_weight=sw)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/losses.py:143 __call__
        losses = self.call(y_true, y_pred)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/losses.py:246 call
        return self.fn(y_true, y_pred, **self._fn_kwargs)
    <ipython-input-477-99f75f332877>:4 customLoss
        loss = K.square(y_true - y_pred) + K.prod(y_true, y_pred)
    /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1716 prod
        return tf.reduce_prod(x, axis, keepdims)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/util/dispatch.py:180 wrapper
        return target(*args, **kwargs)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:2196 reduce_prod
        name=name))
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/gen_math_ops.py:6642 prod
        name=name)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py:578 _apply_op_helper
        param_name=input_name)
    /usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py:61 _SatisfiesTypeConstraint
        ", ".join(dtypes.as_dtype(x).name for x in allowed_list)))

    TypeError: Value passed to parameter 'reduction_indices' has DataType float32 not in list of allowed values: int32, int64

但是，如果我删除了K.prod(y_true, y_pred) 部分，代码将毫无问题地运行。

def customLoss(y_true, y_pred, sample_weight=None):
    y_true = K.cast(y_true, 'float32')
    y_pred = K.cast(y_pred, 'float32')
    loss = K.square(y_true - y_pred) #+ K.prod(y_true, y_pred)
    loss = loss * K.cast(sample_weights, 'float32')
    return loss

有什么问题？？？

【问题讨论】：

K.prod 将张量中的值与指定的轴相乘...这是您想要的吗？考虑到您的损失，一个简单的产品似乎是合适的
天哪！这么愚蠢的错误！成功了，谢谢！

标签： python tensorflow machine-learning keras typeerror

【解决方案1】：

我相信错误是由您调用K.prod() 中的第二个参数引起的。此函数采用单个张量 x，但您指定了两个张量 y_true 和 y_pred。

错误本身的出现是因为K.prod() 的第二个参数引用了一个轴，该轴必须是整数，而不是浮点数。

听起来您可能想使用tf.keras.layers.multiply() 或tf.keras.layers.dot()。

【讨论】：