问题描述
在我的Tensorflow神经网络开始训练之前,会打印出以下警告:
现在,根据错误消息,我可以通过将后备选项设置为'float64'
来使此错误消息静音。但是,我想深入了解这个问题并手动设置正确的 dtypes
。
完整代码:
从tensorflow.keras.layers导入tf
tensorflow从tensorflow.keras导入密集,连接
来自sklearn.datasets的模型
导入load_iris
虹膜,目标= load_iris(return_X_y = True)
X =虹膜[:,:3]
y =虹膜[: ,3]
ds = tf.data.Dataset.from_tensor_slices((X,y))。shuffle(25).batch(8)
class MyModel(Model) :
def __init __(self):
super(MyModel,self).__ init __()
self.d0 = Dense(16,activation ='relu')
self.d1 =密集(32,激活='relu')
self.d2 =密集(1,激活='线性')
def调用(self,x):
x = self.d0(x)
x = self.d1(x)
x = self.d2(x)
return x
model = MyModel()
loss_object = tf.keras.losses.MeanSquaredError()
优化程序= tf。 keras.optimizers.Adam(learning_rate = 5e-4)
损失= tf.keras.metrics.Mean(name ='loss')
错误= tf.keras.metrics.MeanSquaredError( )
@ tf.function
def train_step(输入,目标):以tf.GradientTape()作为磁带的
:
预测=模型(输入)
run_loss =损失对象(目标,预测)
梯度= tape.gradient(run_loss,model.trainable_variables)
Optimizer.apply_gradients(zip(gradients,model.trainable_variables))
损失(run_loss )
错误(预测,目标)
范围内的时期(10):
表示数据,ds中的标签:
train_step(data,标签)
template ='Epoch {:> 2},损失:{:> 7.4f},MSE:{:> 6.2f}'
print(template.format(epoch + 1 ,
loss.result(),
error.result()* 100))
#重置下一个时期的指标
loss.reset_states()
error .reset_states()
tl; dr 为避免这种情况,请将输入内容转换为 float32
X = tf.cast(iris [:,:3],tf.float32)
y = tf.cast(iris [:, 3] ,tf.float32)
或带有 numpy
:
X = np.array(iris [:,:3],dtype = np.float32)
y = np.array(iris [:, 3],dtype = np.float32)
说明
默认情况下,Tensorflow使用 floatx
,默认为 float32
,这是深度学习的标准。您可以验证这一点:
导入张量流为tf
tf.keras.backend.floatx()
Out [3]: float32
您提供的输入(虹膜数据集)的dtype为 float64
,因此Tensorflow的默认权重dtype不匹配,并且输入。 Tensorflow不喜欢这样,因为强制转换(更改dtype)的成本很高。在处理不同dtypes的张量时,Tensorflow通常会引发错误(例如,比较 float32
logits和 float64
标签)。 / p>
新行为它正在谈论:
是它将自动将输入dtype强制转换为 float32
。 Tensorflow 1.X在这种情况下可能引发了异常,尽管我不能说我曾经使用过它。
Before my Tensorflow neural network starts training, the following warning prints out:
Now, based on the error message, I am able to silence this error message by setting the backed to 'float64'
. But, I would like to get to the bottom of this and set the right dtypes
manually.
Full code:
import tensorflow as tf
from tensorflow.keras.layers import Dense, Concatenate
from tensorflow.keras import Model
from sklearn.datasets import load_iris
iris, target = load_iris(return_X_y=True)
X = iris[:, :3]
y = iris[:, 3]
ds = tf.data.Dataset.from_tensor_slices((X, y)).shuffle(25).batch(8)
class MyModel(Model):
def __init__(self):
super(MyModel, self).__init__()
self.d0 = Dense(16, activation='relu')
self.d1 = Dense(32, activation='relu')
self.d2 = Dense(1, activation='linear')
def call(self, x):
x = self.d0(x)
x = self.d1(x)
x = self.d2(x)
return x
model = MyModel()
loss_object = tf.keras.losses.MeanSquaredError()
optimizer = tf.keras.optimizers.Adam(learning_rate=5e-4)
loss = tf.keras.metrics.Mean(name='loss')
error = tf.keras.metrics.MeanSquaredError()
@tf.function
def train_step(inputs, targets):
with tf.GradientTape() as tape:
predictions = model(inputs)
run_loss = loss_object(targets, predictions)
gradients = tape.gradient(run_loss, model.trainable_variables)
optimizer.apply_gradients(zip(gradients, model.trainable_variables))
loss(run_loss)
error(predictions, targets)
for epoch in range(10):
for data, labels in ds:
train_step(data, labels)
template = 'Epoch {:>2}, Loss: {:>7.4f}, MSE: {:>6.2f}'
print(template.format(epoch+1,
loss.result(),
error.result()*100))
# Reset the metrics for the next epoch
loss.reset_states()
error.reset_states()
tl;dr to avoid this, cast your input to float32
X = tf.cast(iris[:, :3], tf.float32)
y = tf.cast(iris[:, 3], tf.float32)
or with numpy
:
X = np.array(iris[:, :3], dtype=np.float32)
y = np.array(iris[:, 3], dtype=np.float32)
Explanation
By default, Tensorflow uses floatx
, which defaults to float32
, which is standard for deep learning. You can verify this:
import tensorflow as tf
tf.keras.backend.floatx()
Out[3]: 'float32'
The input you provided (the Iris dataset), is of dtype float64
, so there is a mismatch between Tensorflow's default dtype for weights, and the input. Tensorflow doesn't like that, because casting (changing the dtype) is costly. Tensorflow will generally throw an error when manipulating tensors of different dtypes (e.g., comparing float32
logits and float64
labels).
The "new behavior" it's talking about:
Is that it will automatically cast the input dtype to float32
. Tensorflow 1.X probably threw an exception in this situation, although I can't say I've ever used it.
这篇关于警告:tensorflow:层my_model正在将输入张量从dtype float64强制转换为float32层的dtype,这是TensorFlow 2中的新行为的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持!