r"""Wide ResNet-50-2 model from `"Wide Residual Networks" `_ The model is the same as ResNet except for the bottleneck number of channels which is twice larger in every block. The number of channels in outer 1x1 convolutions is the same, e.g. las
(pretrained=False, progress=True, **kwargs)
| 379 | |
| 380 | |
| 381 | def wide_resnet50_2(pretrained=False, progress=True, **kwargs): |
| 382 | r"""Wide ResNet-50-2 model from |
| 383 | `"Wide Residual Networks" <https://arxiv.org/pdf/1605.07146.pdf>`_ |
| 384 | The model is the same as ResNet except for the bottleneck number of channels |
| 385 | which is twice larger in every block. The number of channels in outer 1x1 |
| 386 | convolutions is the same, e.g. last block in ResNet-50 has 2048-512-2048 |
| 387 | channels, and in Wide ResNet-50-2 has 2048-1024-2048. |
| 388 | Args: |
| 389 | pretrained (bool): If True, returns a model pre-trained on ImageNet |
| 390 | progress (bool): If True, displays a progress bar of the download to stderr |
| 391 | """ |
| 392 | kwargs['width_per_group'] = 64 * 2 |
| 393 | return _resnet('wide_resnet50_2', Bottleneck, [3, 4, 6, 3], |
| 394 | pretrained, progress, **kwargs) |
| 395 | |
| 396 | |
| 397 | def wide_resnet101_2(pretrained=False, progress=True, **kwargs): |