What is the meaning of the word logits in TensorFlow ?
In TensorFlow, logits are the values to be used as input to softmax.
Hence that logits is a function in maths
The function σ−1(x) is called the logit in statistics, but this term is more rarely used in machine learning. σ−1(x) stands for the inverse function of logistic sigmoid function.
For TensorFlow, logits is commonly seen as the name of last layer
logits is the output of the neural network before going through the softmax activation function.
We can also use softmax as the activation function in the last layer of our design, for computation, we separate logits. So that it is more efficient to calculate softmax and cross-entropy loss together. Note that cross-entropy is a cost function cannot be used in forward propagation.