类型错误：'Mul' Op 的输入'y'的类型 float32 与参数 'x' 的 int64 类型不匹配

Question

在这段代码之后，我在categoricalfocalloss中得到错误，我没有得到wherint64错误来了

def categorical_focal_loss(gamma=2., alpha=.25):
def categorical_focal_loss_fixed(y_true, y_pred):
y_pred /= K.sum(y_pred, axis=-1, keepdims=True)
epsilon = K.epsilon()
y_pred = K.clip(y_pred, epsilon, 1. - epsilon)
y_pred = tf.cast(y_pred, dtype= tf.float32)
cross_entropy = -y_true * K.log(y_pred)
loss = alpha * K.pow(1 - y_pred, gamma) * cross_entropy
return K.sum(loss, axis=1)
return categorical_focal_loss_fixed

模型描述在此代码中，在loss分类中使用了focal loss

with strategy.scope():
ef7 =tf.keras.Sequential()
ef7.add(enet)
ef7.add(tf.keras.layers.MaxPooling2D())
ef7.add(tf.keras.layers.Conv2D(4096,3,padding='same'))
ef7.add(tf.keras.layers.BatchNormalization())
ef7.add(tf.keras.layers.ReLU())
ef7.add(tf.keras.layers.GlobalAveragePooling2D())
ef7.add(tf.keras.layers.Dropout(0.35))
ef7.add(tf.keras.layers.Flatten())

ef7.add(tf.keras.layers.Dense(2048,activation='relu'))
ef7.add(tf.keras.layers.BatchNormalization())
ef7.add(tf.keras.layers.LeakyReLU())
ef7.add(tf.keras.layers.Dropout(0.35))

ef7.add(tf.keras.layers.Dense(1024,activation='relu'))
ef7.add(tf.keras.layers.BatchNormalization())
ef7.add(tf.keras.layers.LeakyReLU())
ef7.add(tf.keras.layers.Dropout(0.25))
ef7.add(tf.keras.layers.Dense(3,activation='softmax'))
ef7.compile(
optimizer=tf.optimizers.Adam(lr=0.0001),
loss=categorical_focal_loss(gamma=2., alpha=.25),
metrics=['categorical_accuracy',
tf.keras.metrics.Recall(),
tf.keras.metrics.Precision(),   
tf.keras.metrics.AUC(),
tfa.metrics.F1Score(num_classes=3, average="macro")
])

在模型中，我使用了分类焦点损失当我在训练数据集中运行这个时，我不知道如何将其转换为intoint64

h7=ef7.fit(
train_dataset,
steps_per_epoch=train_labels.shape[0] // BATCH_SIZE,
callbacks=[lr_callback],
epochs=EPOCHS)

得到的错误如下所示

Epoch 1/20

```Epoch 00001: LearningRateScheduler reducing learning rate to 1e-05.```
---------------------------------------------------------------------------
>TypeError                                 Traceback (most recent call last)
<ipython-input-133-d27eee469b2b> in <module>()
3     steps_per_epoch=train_labels.shape[0] // BATCH_SIZE,
4     callbacks=[lr_callback],
----> 5     epochs=EPOCHS)

9 frames
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/func_graph.py in wrapper(*args, **kwargs)
975           except Exception as e:  # pylint:disable=broad-except
976             if hasattr(e, "ag_error_metadata"):
--> 977               raise e.ag_error_metadata.to_exception(e)
978             else:
979               raise

TypeError: in user code:

>/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:805 train_function  *
>return step_function(self, iterator)
><ipython-input-68-de42355e464e>:7 categorical_focal_loss_fixed  *
cross_entropy = -y_true * K.log(y_pred)
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:1180 binary_op_wrapper
>raise e
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:1164 binary_op_wrapper
>return func(x, y, name=name)
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:1496 _mul_dispatch
>return multiply(x, y, name=name)
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/util/dispatch.py:201 wrapper
return target(*args, **kwargs)
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:518 multiply
>return gen_math_ops.mul(x, y, name)
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/gen_math_ops.py:6078 mul
>   "Mul", x=x, y=y, name=name)
>/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py:558 _apply_op_helper
>   inferred_from[input_arg.type_attr]))

>TypeError: Input 'y' of 'Mul' Op has type float32 that does not match type int64 of argument 'x'.```

Answer 1

错误指向这行代码:

cross_entropy = -y_true * K.log(y_pred)

是从tensorflow包中的math_ops.py中的乘法函数抛出的。深入研究该文件，我找到了参数需求的摘要。

Args:
x: A Tensor. Must be one of the following types: `bfloat16`,
`half`, `float32`, `float64`, `uint8`, `int8`, `uint16`,
`int16`, `int32`, `int64`, `complex64`, `complex128`.
y: A `Tensor`. Must have the same type as `x`.
name: A name for the operation (optional).
Returns:
A `Tensor`.  Has the same type as `x`.
Raises:
* InvalidArgumentError: When `x` and `y` have incompatible shapes or types

回头看错误

TypeError: Input 'y' of 'Mul' Op has type float32 that does not match type int64 of argument 'x'.```

表示-y_true为'x',K.log(y_pred)为'y'。要执行此操作，您必须将-y_true强制转换为float32或强制转换K.log(y_pred)转换为int64类型，或者只要它们匹配，就将它们都转换为任何其他类型。.