Views
No views yet
timm by Ross Wightman using RandAugment RA3 recipe. Related to B recipe in ResNet Strikes Back.timm specific pre-activation Xception block.1from urllib.request import urlopen
2from PIL import Image
3import timm
4
5img = Image.open(urlopen(
6 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
7))
8
9model = timm.create_model('xception65p.ra3_in1k', pretrained=True)
10model = model.eval()
11
12# get model specific transforms (normalization, resize)
13data_config = timm.data.resolve_model_data_config(model)
14transforms = timm.data.create_transform(**data_config, is_training=False)
15
16output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1
17
18top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5)1from urllib.request import urlopen
2from PIL import Image
3import timm
4
5img = Image.open(urlopen(
6 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
7))
8
9model = timm.create_model(
10 'xception65p.ra3_in1k',
11 pretrained=True,
12 features_only=True,
13)
14model = model.eval()
15
16# get model specific transforms (normalization, resize)
17data_config = timm.data.resolve_model_data_config(model)
18transforms = timm.data.create_transform(**data_config, is_training=False)
19
20output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1
21
22for o in output:
23 # print shape of each feature map in output
24 # e.g.:
25 # torch.Size([1, 128, 150, 150])
26 # torch.Size([1, 256, 75, 75])
27 # torch.Size([1, 728, 38, 38])
28 # torch.Size([1, 1024, 19, 19])
29 # torch.Size([1, 2048, 10, 10])
30
31 print(o.shape)1from urllib.request import urlopen
2from PIL import Image
3import timm
4
5img = Image.open(urlopen(
6 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
7))
8
9model = timm.create_model(
10 'xception65p.ra3_in1k',
11 pretrained=True,
12 num_classes=0, # remove classifier nn.Linear
13)
14model = model.eval()
15
16# get model specific transforms (normalization, resize)
17data_config = timm.data.resolve_model_data_config(model)
18transforms = timm.data.create_transform(**data_config, is_training=False)
19
20output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor
21
22# or equivalently (without needing to set num_classes=0)
23
24output = model.forward_features(transforms(img).unsqueeze(0))
25# output is unpooled, a (1, 2048, 10, 10) shaped tensor
26
27output = model.forward_head(output, pre_logits=True)
28# output is a (1, num_features) shaped tensor1@inproceedings{deeplabv3plus2018,
2 title={Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation},
3 author={Liang-Chieh Chen and Yukun Zhu and George Papandreou and Florian Schroff and Hartwig Adam},
4 booktitle={ECCV},
5 year={2018}
6}1@misc{chollet2017xception,
2 title={Xception: Deep Learning with Depthwise Separable Convolutions},
3 author={François Chollet},
4 year={2017},
5 eprint={1610.02357},
6 archivePrefix={arXiv},
7 primaryClass={cs.CV}
8}1@inproceedings{wightman2021resnet,
2 title={ResNet strikes back: An improved training procedure in timm},
3 author={Wightman, Ross and Touvron, Hugo and Jegou, Herve},
4 booktitle={NeurIPS 2021 Workshop on ImageNet: Past, Present, and Future}
5}