tf.multinomial(logits, num_samples, seed=None, name=None)
See the guide: Constants, Sequences, and Random Values > Random Tensors
Draws samples from a multinomial distribution.
# samples has shape [1, 5], where each value is either 0 or 1 with equal # probability. samples = tf.multinomial(tf.log([[10., 10.]]), 5)
logits: 2-D Tensor with shape
[batch_size, num_classes]. Each slice
[i, :]represents the unnormalized log probabilities for all classes.
num_samples: 0-D. Number of independent samples to draw for each row slice.
seed: A Python integer. Used to create a random seed for the distribution. See
name: Optional name for the operation.
The drawn samples of shape