Skip to content

Instantly share code, notes, and snippets.

@thomwolf
Last active January 25, 2021 00:51
Show Gist options
  • Star 4 You must be signed in to star a gist
  • Fork 1 You must be signed in to fork a gist
  • Save thomwolf/dec72992ea6817290273d42f6b95c04c to your computer and use it in GitHub Desktop.
Save thomwolf/dec72992ea6817290273d42f6b95c04c to your computer and use it in GitHub Desktop.
A pyTorch attention layer for torchMoji model
class Attention(Module):
"""
Computes a weighted average of channels across timesteps (1 parameter pr. channel).
"""
def __init__(self, attention_size, return_attention=False):
""" Initialize the attention layer
# Arguments:
attention_size: Size of the attention vector.
return_attention: If true, output will include the weight for each input token
used for the prediction
"""
super(Attention, self).__init__()
self.return_attention = return_attention
self.attention_size = attention_size
self.attention_vector = Parameter(torch.FloatTensor(attention_size))
def __repr__(self):
s = '{name}({attention_size}, return attention={return_attention})'
return s.format(name=self.__class__.__name__, **self.__dict__)
def forward(self, inputs, input_lengths):
""" Forward pass.
# Arguments:
inputs (Torch.Variable): Tensor of input sequences
input_lengths (torch.LongTensor): Lengths of the sequences
# Return:
Tuple with (representations and attentions if self.return_attention else None).
"""
logits = inputs.matmul(self.attention_vector)
unnorm_ai = (logits - logits.max()).exp()
# Compute a mask for the attention on the padded sequences
# See e.g. https://discuss.pytorch.org/t/self-attention-on-words-and-masking/5671/5
max_len = unnorm_ai.size(1)
idxes = torch.arange(0, max_len, out=torch.LongTensor(max_len)).unsqueeze(0)
if torch.cuda.is_available():
idxes = idxes.cuda()
mask = Variable((idxes < input_lengths.unsqueeze(1)).float())
# apply mask and renormalize attention scores (weights)
masked_weights = unnorm_ai * mask
att_sums = masked_weights.sum(dim=1, keepdim=True) # sums per sequence
attentions = masked_weights.div(att_sums)
# apply attention weights
weighted = torch.mul(inputs, attentions.unsqueeze(-1).expand_as(inputs))
# get the final fixed vector representations of the sentences
representations = weighted.sum(dim=1)
return (representations, attentions if self.return_attention else None)
@BramVanroy
Copy link

BramVanroy commented Sep 4, 2019

Came here after reading the blog post. I assume that the last return has to be indented as well? The same issue arises in some other snippets in that post.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment