Views
No views yet
1import birder
2from birder.inference.classification import infer_image
3
4# Option 1: manual setup (more control over preprocessing)
5(net, model_info) = birder.load_pretrained_model("vit_reg4_b14_nps_ls_dino-v2-lvd142m", inference=True)
6
7# Get the image size the model was trained on
8size = birder.get_size_from_signature(model_info.signature)
9
10# Create an inference transform
11transform = birder.classification_transform(size, model_info.rgb_stats)
12
13# Option 2: helper (quick start with default preprocessing)
14(net, model_info, transform) = birder.load_pretrained_model_and_transform("vit_reg4_b14_nps_ls_dino-v2-lvd142m", inference=True)
15
16image = "path/to/image.jpeg" # or a PIL image
17(out, embedding) = infer_image(net, image, transform, return_embedding=True)
18# embedding is a NumPy array with shape of (1, 768)1from PIL import Image
2import birder
3
4(net, model_info, transform) = birder.load_pretrained_model_and_transform("vit_reg4_b14_nps_ls_dino-v2-lvd142m", inference=True)
5
6image = Image.open("path/to/image.jpeg")
7features = net.detection_features(transform(image).unsqueeze(0))
8# features is a dict (stage name -> torch.Tensor)
9print([(k, v.size()) for k, v in features.items()])
10# Output example:
11# [('stage1', torch.Size([1, 768, 37, 37]))]1@misc{dosovitskiy2021imageworth16x16words,
2 title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},
3 author={Alexey Dosovitskiy and Lucas Beyer and Alexander Kolesnikov and Dirk Weissenborn and Xiaohua Zhai and Thomas Unterthiner and Mostafa Dehghani and Matthias Minderer and Georg Heigold and Sylvain Gelly and Jakob Uszkoreit and Neil Houlsby},
4 year={2021},
5 eprint={2010.11929},
6 archivePrefix={arXiv},
7 primaryClass={cs.CV},
8 url={https://arxiv.org/abs/2010.11929},
9}
10
11@misc{darcet2024visiontransformersneedregisters,
12 title={Vision Transformers Need Registers},
13 author={Timothée Darcet and Maxime Oquab and Julien Mairal and Piotr Bojanowski},
14 year={2024},
15 eprint={2309.16588},
16 archivePrefix={arXiv},
17 primaryClass={cs.CV},
18 url={https://arxiv.org/abs/2309.16588},
19}
20
21@misc{oquab2024dinov2learningrobustvisual,
22 title={DINOv2: Learning Robust Visual Features without Supervision},
23 author={Maxime Oquab and Timothée Darcet and Théo Moutakanni and Huy Vo and Marc Szafraniec and Vasil Khalidov and Pierre Fernandez and Daniel Haziza and Francisco Massa and Alaaeldin El-Nouby and Mahmoud Assran and Nicolas Ballas and Wojciech Galuba and Russell Howes and Po-Yao Huang and Shang-Wen Li and Ishan Misra and Michael Rabbat and Vasu Sharma and Gabriel Synnaeve and Hu Xu and Hervé Jegou and Julien Mairal and Patrick Labatut and Armand Joulin and Piotr Bojanowski},
24 year={2024},
25 eprint={2304.07193},
26 archivePrefix={arXiv},
27 primaryClass={cs.CV},
28 url={https://arxiv.org/abs/2304.07193},
29}