在Keras(Tensorflow后端)中使用binary_crossentropy损失 [英] Using binary_crossentropy loss in Keras (Tensorflow backend)

问题描述

在Keras文档中的培训示例中，

In the training example in Keras documentation,

https://keras.io/getting-started/sequential-model -guide/#training

binary_crossentropy ，并在网络的最后一层添加 Sigmoid 激活，但是是否有必要在最后一层添加 Sigmoid ?正如我在源代码中找到的:

binary_crossentropy is used and sigmoid activation is added in the network's last layer, but is it necessary that add sigmoid in the last layer? As I found in the source code:

def binary_crossentropy(output, target, from_logits=False):
  """Binary crossentropy between an output tensor and a target tensor.
  Arguments:
      output: A tensor.
      target: A tensor with the same shape as `output`.
      from_logits: Whether `output` is expected to be a logits tensor.
          By default, we consider that `output`
          encodes a probability distribution.
  Returns:
      A tensor.
  """
  # Note: nn.softmax_cross_entropy_with_logits
  # expects logits, Keras expects probabilities.
  if not from_logits:
    # transform back to logits
    epsilon = _to_tensor(_EPSILON, output.dtype.base_dtype)
    output = clip_ops.clip_by_value(output, epsilon, 1 - epsilon)
    output = math_ops.log(output / (1 - output))
  return nn.sigmoid_cross_entropy_with_logits(labels=target, logits=output)

Keras在Tensorflow中调用 sigmoid_cross_entropy_with_logits ，但在 sigmoid_cross_entropy_with_logits 函数中，再次计算.

Keras invokes sigmoid_cross_entropy_with_logits in Tensorflow, but in sigmoid_cross_entropy_with_logits function, sigmoid(logits) is calculated again.

https://www.tensorflow.org/versions/master/api_docs/python/tf/nn/sigmoid_cross_entropy_with_logits

因此，我认为最后添加 Sigmoid 并不有意义，但是似乎我在网上发现的Keras中所有的二进制/多标签分类示例和教程都添加了 sigmoid .而且我不明白什么是

So I don't think it makes sense that add a sigmoid at last, but seemingly all the binary/multi-label classification examples and tutorials in Keras I found online added sigmoid at last. Besides I don't understand what is the meaning of

# Note: nn.softmax_cross_entropy_with_logits
# expects logits, Keras expects probabilities.

为什么Keras期望概率?它不使用 nn.softmax_cross_entropy_with_logits 函数吗?有道理吗?

Why Keras expects probabilities? Doesn't it use the nn.softmax_cross_entropy_with_logits function? Does it make sense?

谢谢.

推荐答案

您是对的，这就是正在发生的事情.我相信这是由于历史原因.

You're right, that's exactly what's happening. I believe this is due to historical reasons.

Keras是在张量流之前创建的，用作theano的包装.并且在theano中，必须手动计算Sigmax/softmax，然后应用交叉熵损失函数. Tensorflow可以在一个融合的操作中完成所有操作，但是社区已经采用了具有Sigmoid/softmax层的API.

Keras was created before tensorflow, as a wrapper around theano. And in theano, one has to compute sigmoid/softmax manually and then apply cross-entropy loss function. Tensorflow does everything in one fused op, but the API with sigmoid/softmax layer was already adopted by the community.

如果要避免不必要的logit<->概率转换，请用from_logits=True调用binary_crossentropy损失，并且不要添加S形层.

If you want to avoid unnecessary logit <-> probability conversions, call binary_crossentropy loss withfrom_logits=True and don't add the sigmoid layer.

这篇关于在Keras(Tensorflow后端)中使用binary_crossentropy损失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IT屋！