- torch.nn.functional.gumbel_softmax(logits, tau=1, hard=False, eps=1e-10, dim=- 1)[source]¶
Samples from the Gumbel-Softmax distribution (Link 1 Link 2) and optionally discretizes.
logits (Tensor) – […, num_features] unnormalized log probabilities
tau (float) – non-negative scalar temperature
hard (bool) – if
True, the returned samples will be discretized as one-hot vectors, but will be differentiated as if it is the soft sample in autograd
dim (int) – A dimension along which softmax will be computed. Default: -1.
Sampled tensor of same shape as logits from the Gumbel-Softmax distribution. If
hard=True, the returned samples will be one-hot, otherwise they will be probability distributions that sum to 1 across dim.
- Return type:
This function is here for legacy reasons, may be removed from nn.Functional in the future.
The main trick for hard is to do y_hard - y_soft.detach() + y_soft
It achieves two things: - makes the output value exactly one-hot (since we add then subtract y_soft value) - makes the gradient equal to y_soft gradient (since we strip all other gradients)
>>> logits = torch.randn(20, 32) >>> # Sample soft categorical using reparametrization trick: >>> F.gumbel_softmax(logits, tau=1, hard=False) >>> # Sample hard categorical using "Straight-through" trick: >>> F.gumbel_softmax(logits, tau=1, hard=True)