Source code for catalyst.contrib.layers.arcmargin
import torch
import torch.nn as nn
import torch.nn.functional as F
[docs]class ArcMarginProduct(nn.Module):
"""Implementation of Arc Margin Product.
Args:
in_features: size of each input sample.
out_features: size of each output sample.
Shape:
- Input: :math:`(batch, H_{in})` where
:math:`H_{in} = in\_features`.
- Output: :math:`(batch, H_{out})` where
:math:`H_{out} = out\_features`.
Example:
>>> layer = ArcMarginProduct(5, 10)
>>> loss_fn = nn.CrosEntropyLoss()
>>> embedding = torch.randn(3, 5, requires_grad=True)
>>> target = torch.empty(3, dtype=torch.long).random_(10)
>>> output = layer(embedding)
>>> loss = loss_fn(output, target)
>>> self.engine.backward(loss)
"""
def __init__(self, in_features: int, out_features: int): # noqa: D107
super(ArcMarginProduct, self).__init__()
self.in_features = in_features
self.out_features = out_features
self.weight = nn.Parameter(torch.Tensor(out_features, in_features))
nn.init.xavier_uniform_(self.weight)
def __repr__(self) -> str:
"""Object representation."""
rep = (
"ArcMarginProduct("
f"in_features={self.in_features},"
f"out_features={self.out_features}"
")"
)
return rep
[docs] def forward(self, input: torch.Tensor) -> torch.Tensor:
"""
Args:
input: input features,
expected shapes ``BxF`` where ``B``
is batch dimension and ``F`` is an
input feature dimension.
Returns:
tensor (logits) with shapes ``BxC``
where ``C`` is a number of classes
(out_features).
"""
cosine = F.linear(F.normalize(input), F.normalize(self.weight))
return cosine
__all__ = ["ArcMarginProduct"]