Shortcuts

mmdet.models.utils.brick_wrappers 源代码

# Copyright (c) OpenMMLab. All rights reserved.
import torch
import torch.nn as nn
import torch.nn.functional as F
from mmcv.cnn.bricks.wrappers import NewEmptyTensorOp, obsolete_torch_version

if torch.__version__ == 'parrots':
    TORCH_VERSION = torch.__version__
else:
    # torch.__version__ could be 1.3.1+cu92, we only need the first two
    # for comparison
    TORCH_VERSION = tuple(int(x) for x in torch.__version__.split('.')[:2])


[文档]def adaptive_avg_pool2d(input, output_size): """Handle empty batch dimension to adaptive_avg_pool2d. Args: input (tensor): 4D tensor. output_size (int, tuple[int,int]): the target output size. """ if input.numel() == 0 and obsolete_torch_version(TORCH_VERSION, (1, 9)): if isinstance(output_size, int): output_size = [output_size, output_size] output_size = [*input.shape[:2], *output_size] empty = NewEmptyTensorOp.apply(input, output_size) return empty else: return F.adaptive_avg_pool2d(input, output_size)
[文档]class AdaptiveAvgPool2d(nn.AdaptiveAvgPool2d): """Handle empty batch dimension to AdaptiveAvgPool2d."""
[文档] def forward(self, x): # PyTorch 1.9 does not support empty tensor inference yet if x.numel() == 0 and obsolete_torch_version(TORCH_VERSION, (1, 9)): output_size = self.output_size if isinstance(output_size, int): output_size = [output_size, output_size] else: output_size = [ v if v is not None else d for v, d in zip(output_size, x.size()[-2:]) ] output_size = [*x.shape[:2], *output_size] empty = NewEmptyTensorOp.apply(x, output_size) return empty return super().forward(x)
Read the Docs v: v2.25.1
Versions
latest
stable
v2.25.1
v2.25.0
v2.24.1
v2.24.0
v2.23.0
v2.22.0
v2.21.0
v2.20.0
v2.19.1
v2.19.0
v2.18.1
v2.18.0
v2.17.0
v2.16.0
v2.15.1
v2.15.0
v2.14.0
Downloads
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.