smhh24's picture
Upload 90 files
560b597 verified
raw
history blame contribute delete
722 Bytes
import torch
import torch.nn as nn
def drop_path(x: torch.Tensor, drop_prob: float = 0.0, training: bool = False):
if drop_prob == 0.0 or not training:
return x
keep_prob = 1 - drop_prob
shape = (x.shape[0],) + (1,) * (
x.ndim - 1
) # work with diff dim tensors, not just 2D ConvNets
random_tensor = x.new_empty(shape).bernoulli_(keep_prob)
if keep_prob > 0.0:
random_tensor.div_(keep_prob)
output = x * random_tensor
return output
class DropPath(nn.Module):
def __init__(self, drop_prob=None):
super(DropPath, self).__init__()
self.drop_prob = drop_prob
def forward(self, x):
return drop_path(x, self.drop_prob, self.training)