! [ -e /content ] && pip install -Uqq fastai # 在Colab上升级fastaiXResnet
from __future__ import annotations
from fastai.torch_basics import *
try: from torchvision.models.utils import load_state_dict_from_url
except ModuleNotFoundError: from torch.hub import load_state_dict_from_urlfrom nbdev.showdoc import *来自“杂技袋”论文的Resnet
def init_cnn(m):
if getattr(m, 'bias', None) is not None: nn.init.constant_(m.bias, 0)
if isinstance(m, (nn.Conv1d,nn.Conv2d,nn.Conv3d,nn.Linear)): nn.init.kaiming_normal_(m.weight)
for l in m.children(): init_cnn(l)class XResNet(nn.Sequential):
@delegates(ResBlock)
def __init__(self, block, expansion, layers, p=0.0, c_in=3, n_out=1000, stem_szs=(32,32,64),
widen=1.0, sa=False, act_cls=defaults.activation, ndim=2, ks=3, stride=2, **kwargs):
store_attr('block,expansion,act_cls,ndim,ks')
if ks % 2 == 0: raise Exception('kernel size has to be odd!')
stem_szs = [c_in, *stem_szs]
stem = [ConvLayer(stem_szs[i], stem_szs[i+1], ks=ks, stride=stride if i==0 else 1,
act_cls=act_cls, ndim=ndim)
for i in range(3)]
block_szs = [int(o*widen) for o in [64,128,256,512] +[256]*(len(layers)-4)]
block_szs = [64//expansion] + block_szs
blocks = self._make_blocks(layers, block_szs, sa, stride, **kwargs)
super().__init__(
*stem, MaxPool(ks=ks, stride=stride, padding=ks//2, ndim=ndim),
*blocks,
AdaptiveAvgPool(sz=1, ndim=ndim), Flatten(), nn.Dropout(p),
nn.Linear(block_szs[-1]*expansion, n_out),
)
init_cnn(self)
def _make_blocks(self, layers, block_szs, sa, stride, **kwargs):
return [self._make_layer(ni=block_szs[i], nf=block_szs[i+1], blocks=l,
stride=1 if i==0 else stride, sa=sa and i==len(layers)-4, **kwargs)
for i,l in enumerate(layers)]
def _make_layer(self, ni, nf, blocks, stride, sa, **kwargs):
return nn.Sequential(
*[self.block(self.expansion, ni if i==0 else nf, nf, stride=stride if i==0 else 1,
sa=sa and i==(blocks-1), act_cls=self.act_cls, ndim=self.ndim, ks=self.ks, **kwargs)
for i in range(blocks)])def _xresnet(pretrained, expansion, layers, **kwargs):
# TODO：预训练所有尺寸。当前情况下，若非使用xrn50，将会失败。
url = 'https://s3.amazonaws.com/fast-ai-modelzoo/xrn50_940.pth'
res = XResNet(ResBlock, expansion, layers, **kwargs)
if pretrained: res.load_state_dict(load_state_dict_from_url(url, map_location='cpu')['model'], strict=False)
return res
def xresnet18 (pretrained=False, **kwargs): return _xresnet(pretrained, 1, [2, 2, 2, 2], **kwargs)
def xresnet34 (pretrained=False, **kwargs): return _xresnet(pretrained, 1, [3, 4, 6, 3], **kwargs)
def xresnet50 (pretrained=False, **kwargs): return _xresnet(pretrained, 4, [3, 4, 6, 3], **kwargs)
def xresnet101(pretrained=False, **kwargs): return _xresnet(pretrained, 4, [3, 4, 23, 3], **kwargs)
def xresnet152(pretrained=False, **kwargs): return _xresnet(pretrained, 4, [3, 8, 36, 3], **kwargs)
def xresnet18_deep (pretrained=False, **kwargs): return _xresnet(pretrained, 1, [2,2,2,2,1,1], **kwargs)
def xresnet34_deep (pretrained=False, **kwargs): return _xresnet(pretrained, 1, [3,4,6,3,1,1], **kwargs)
def xresnet50_deep (pretrained=False, **kwargs): return _xresnet(pretrained, 4, [3,4,6,3,1,1], **kwargs)
def xresnet18_deeper(pretrained=False, **kwargs): return _xresnet(pretrained, 1, [2,2,1,1,1,1,1,1], **kwargs)
def xresnet34_deeper(pretrained=False, **kwargs): return _xresnet(pretrained, 1, [3,4,6,3,1,1,1,1], **kwargs)
def xresnet50_deeper(pretrained=False, **kwargs): return _xresnet(pretrained, 4, [3,4,6,3,1,1,1,1], **kwargs)se_kwargs1 = dict(groups=1 , reduction=16)
se_kwargs2 = dict(groups=32, reduction=16)
se_kwargs3 = dict(groups=32, reduction=0)
g0 = [2,2,2,2]
g1 = [3,4,6,3]
g2 = [3,4,23,3]
g3 = [3,8,36,3]def xse_resnet18(n_out=1000, pretrained=False, **kwargs): return XResNet(SEBlock, 1, g0, n_out=n_out, **se_kwargs1, **kwargs)
def xse_resnext18(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, g0, n_out=n_out, **se_kwargs2, **kwargs)
def xresnext18(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, g0, n_out=n_out, **se_kwargs3, **kwargs)
def xse_resnet34(n_out=1000, pretrained=False, **kwargs): return XResNet(SEBlock, 1, g1, n_out=n_out, **se_kwargs1, **kwargs)
def xse_resnext34(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, g1, n_out=n_out, **se_kwargs2, **kwargs)
def xresnext34(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, g1, n_out=n_out, **se_kwargs3, **kwargs)
def xse_resnet50(n_out=1000, pretrained=False, **kwargs): return XResNet(SEBlock, 4, g1, n_out=n_out, **se_kwargs1, **kwargs)
def xse_resnext50(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 4, g1, n_out=n_out, **se_kwargs2, **kwargs)
def xresnext50(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 4, g1, n_out=n_out, **se_kwargs3, **kwargs)
def xse_resnet101(n_out=1000, pretrained=False, **kwargs): return XResNet(SEBlock, 4, g2, n_out=n_out, **se_kwargs1, **kwargs)
def xse_resnext101(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 4, g2, n_out=n_out, **se_kwargs2, **kwargs)
def xresnext101(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 4, g2, n_out=n_out, **se_kwargs3, **kwargs)
def xse_resnet152(n_out=1000, pretrained=False, **kwargs): return XResNet(SEBlock, 4, g3, n_out=n_out, **se_kwargs1, **kwargs)
def xsenet154(n_out=1000, pretrained=False, **kwargs):
return XResNet(SEBlock, g3, groups=64, reduction=16, p=0.2, n_out=n_out)
def xse_resnext18_deep (n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, g0+[1,1], n_out=n_out, **se_kwargs2, **kwargs)
def xse_resnext34_deep (n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, g1+[1,1], n_out=n_out, **se_kwargs2, **kwargs)
def xse_resnext50_deep (n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 4, g1+[1,1], n_out=n_out, **se_kwargs2, **kwargs)
def xse_resnext18_deeper(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, [2,2,1,1,1,1,1,1], n_out=n_out, **se_kwargs2, **kwargs)
def xse_resnext34_deeper(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 1, [3,4,4,2,2,1,1,1], n_out=n_out, **se_kwargs2, **kwargs)
def xse_resnext50_deeper(n_out=1000, pretrained=False, **kwargs): return XResNet(SEResNeXtBlock, 4, [3,4,4,2,2,1,1,1], n_out=n_out, **se_kwargs2, **kwargs)tst = xse_resnext18()
x = torch.randn(64, 3, 128, 128)
y = tst(x)tst = xresnext18()
x = torch.randn(64, 3, 128, 128)
y = tst(x)tst = xse_resnet50()
x = torch.randn(8, 3, 64, 64)
y = tst(x)tst = xresnet18(ndim=1, c_in=1, ks=15)
x = torch.randn(64, 1, 128)
y = tst(x)tst = xresnext50(ndim=1, c_in=2, ks=31, stride=4)
x = torch.randn(8, 2, 128)
y = tst(x)tst = xresnet18(ndim=3, c_in=3, ks=3)
x = torch.randn(8, 3, 32, 32, 32)
y = tst(x)导出 -
from nbdev import *
nbdev_export()Converted 00_torch_core.ipynb.
Converted 01_layers.ipynb.
Converted 02_data.load.ipynb.
Converted 03_data.core.ipynb.
Converted 04_data.external.ipynb.
Converted 05_data.transforms.ipynb.
Converted 06_data.block.ipynb.
Converted 07_vision.core.ipynb.
Converted 08_vision.data.ipynb.
Converted 09_vision.augment.ipynb.
Converted 09b_vision.utils.ipynb.
Converted 09c_vision.widgets.ipynb.
Converted 10_tutorial.pets.ipynb.
Converted 11_vision.models.xresnet.ipynb.
Converted 12_optimizer.ipynb.
Converted 13_callback.core.ipynb.
Converted 13a_learner.ipynb.
Converted 13b_metrics.ipynb.
Converted 14_callback.schedule.ipynb.
Converted 14a_callback.data.ipynb.
Converted 15_callback.hook.ipynb.
Converted 15a_vision.models.unet.ipynb.
Converted 16_callback.progress.ipynb.
Converted 17_callback.tracker.ipynb.
Converted 18_callback.fp16.ipynb.
Converted 18a_callback.training.ipynb.
Converted 19_callback.mixup.ipynb.
Converted 20_interpret.ipynb.
Converted 20a_distributed.ipynb.
Converted 21_vision.learner.ipynb.
Converted 22_tutorial.imagenette.ipynb.
Converted 23_tutorial.vision.ipynb.
Converted 24_tutorial.siamese.ipynb.
Converted 24_vision.gan.ipynb.
Converted 30_text.core.ipynb.
Converted 31_text.data.ipynb.
Converted 32_text.models.awdlstm.ipynb.
Converted 33_text.models.core.ipynb.
Converted 34_callback.rnn.ipynb.
Converted 35_tutorial.wikitext.ipynb.
Converted 36_text.models.qrnn.ipynb.
Converted 37_text.learner.ipynb.
Converted 38_tutorial.text.ipynb.
Converted 39_tutorial.transformers.ipynb.
Converted 40_tabular.core.ipynb.
Converted 41_tabular.data.ipynb.
Converted 42_tabular.model.ipynb.
Converted 43_tabular.learner.ipynb.
Converted 44_tutorial.tabular.ipynb.
Converted 45_collab.ipynb.
Converted 46_tutorial.collab.ipynb.
Converted 50_tutorial.datablock.ipynb.
Converted 60_medical.imaging.ipynb.
Converted 61_tutorial.medical_imaging.ipynb.
Converted 65_medical.text.ipynb.
Converted 70_callback.wandb.ipynb.
Converted 71_callback.tensorboard.ipynb.
Converted 72_callback.neptune.ipynb.
Converted 73_callback.captum.ipynb.
Converted 74_callback.cutmix.ipynb.
Converted 97_test_utils.ipynb.
Converted 99_pytorch_doc.ipynb.
Converted dev-setup.ipynb.
Converted index.ipynb.
Converted quick_start.ipynb.
Converted tutorial.ipynb.