Views
No views yet
timm scripts using hyper-parameters inspired by the MobileNet-V4 small, mixed with go-to hparams from timm and "ResNet Strikes Back".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('efficientnet_b1.ra4_e3600_r240_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 'efficientnet_b1.ra4_e3600_r240_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, 16, 120, 120])
26 # torch.Size([1, 24, 60, 60])
27 # torch.Size([1, 40, 30, 30])
28 # torch.Size([1, 112, 15, 15])
29 # torch.Size([1, 320, 8, 8])
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 'efficientnet_b1.ra4_e3600_r240_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, 1280, 8, 8) shaped tensor
26
27output = model.forward_head(output, pre_logits=True)
28# output is a (1, num_features) shaped tensor1@misc{rw2019timm,
2 author = {Ross Wightman},
3 title = {PyTorch Image Models},
4 year = {2019},
5 publisher = {GitHub},
6 journal = {GitHub repository},
7 doi = {10.5281/zenodo.4414861},
8 howpublished = {\url{https://github.com/huggingface/pytorch-image-models}}
9}1@inproceedings{tan2019efficientnet,
2 title={Efficientnet: Rethinking model scaling for convolutional neural networks},
3 author={Tan, Mingxing and Le, Quoc},
4 booktitle={International conference on machine learning},
5 pages={6105--6114},
6 year={2019},
7 organization={PMLR}
8}1@article{qin2024mobilenetv4,
2 title={MobileNetV4-Universal Models for the Mobile Ecosystem},
3 author={Qin, Danfeng and Leichner, Chas and Delakis, Manolis and Fornoni, Marco and Luo, Shixin and Yang, Fan and Wang, Weijun and Banbury, Colby and Ye, Chengxi and Akin, Berkin and others},
4 journal={arXiv preprint arXiv:2404.10518},
5 year={2024}
6}