Image Classification
Transformers
PyTorch
Inference Endpoints
Francesco commited on
Commit
a7ec2ee
1 Parent(s): 5830c87

commit files to HF hub

Browse files
Files changed (1) hide show
  1. README.md +49 -0
README.md ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #resnet18
2
+ Implementation of ResNet proposed in [Deep Residual Learning for Image
3
+ Recognition](https://arxiv.org/abs/1512.03385)
4
+
5
+ ``` python
6
+ ResNet.resnet18()
7
+ ResNet.resnet26()
8
+ ResNet.resnet34()
9
+ ResNet.resnet50()
10
+ ResNet.resnet101()
11
+ ResNet.resnet152()
12
+ ResNet.resnet200()
13
+
14
+ Variants (d) proposed in `Bag of Tricks for Image Classification with Convolutional Neural Networks <https://arxiv.org/pdf/1812.01187.pdf`_
15
+
16
+ ResNet.resnet26d()
17
+ ResNet.resnet34d()
18
+ ResNet.resnet50d()
19
+ # You can construct your own one by chaning `stem` and `block`
20
+ resnet101d = ResNet.resnet101(stem=ResNetStemC, block=partial(ResNetBottleneckBlock, shortcut=ResNetShorcutD))
21
+ ```
22
+
23
+ Examples:
24
+
25
+ ``` python
26
+ # change activation
27
+ ResNet.resnet18(activation = nn.SELU)
28
+ # change number of classes (default is 1000 )
29
+ ResNet.resnet18(n_classes=100)
30
+ # pass a different block
31
+ ResNet.resnet18(block=SENetBasicBlock)
32
+ # change the steam
33
+ model = ResNet.resnet18(stem=ResNetStemC)
34
+ change shortcut
35
+ model = ResNet.resnet18(block=partial(ResNetBasicBlock, shortcut=ResNetShorcutD))
36
+ # store each feature
37
+ x = torch.rand((1, 3, 224, 224))
38
+ # get features
39
+ model = ResNet.resnet18()
40
+ # first call .features, this will activate the forward hooks and tells the model you'll like to get the features
41
+ model.encoder.features
42
+ model(torch.randn((1,3,224,224)))
43
+ # get the features from the encoder
44
+ features = model.encoder.features
45
+ print([x.shape for x in features])
46
+ #[torch.Size([1, 64, 112, 112]), torch.Size([1, 64, 56, 56]), torch.Size([1, 128, 28, 28]), torch.Size([1, 256, 14, 14])]
47
+ ```
48
+
49
+