File size: 3,972 Bytes
a4f6b1d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 |
---
tags:
- image-classification
- birder
- pytorch
library_name: birder
license: apache-2.0
---
# Model Card for vit_parallel_s16_18x2_ls_avg_data2vec-intermediate-il-all
A ViT Parallel s16 18x2 image classification model. The model follows a three-stage training process: first, data2vec pretraining, next intermediate training on a large-scale dataset containing diverse bird species from around the world, finally fine-tuned specifically on the `il-all` dataset. The dataset, encompassing all relevant bird species found in Israel, including rarities.
The species list is derived from data available at <https://www.israbirding.com/checklist/>.
## Model Details
- **Model Type:** Image classification and detection backbone
- **Model Stats:**
- Params (M): 64.6
- Input image size: 384 x 384
- **Dataset:** il-all (550 classes)
- Intermediate training involved ~8000 species from all over the world
- **Papers:**
- Three things everyone should know about Vision Transformers: <https://arxiv.org/abs/2203.09795>
- data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language: <https://arxiv.org/abs/2202.03555>
## Model Usage
### Image Classification
```python
import birder
from birder.inference.classification import infer_image
(net, model_info) = birder.load_pretrained_model("vit_parallel_s16_18x2_ls_avg_data2vec-intermediate-il-all", inference=True)
# Get the image size the model was trained on
size = birder.get_size_from_signature(model_info.signature)
# Create an inference transform
transform = birder.classification_transform(size, model_info.rgb_stats)
image = "path/to/image.jpeg" # or a PIL image, must be loaded in RGB format
(out, _) = infer_image(net, image, transform)
# out is a NumPy array with shape of (1, 550), representing class probabilities.
```
### Image Embeddings
```python
import birder
from birder.inference.classification import infer_image
(net, model_info) = birder.load_pretrained_model("vit_parallel_s16_18x2_ls_avg_data2vec-intermediate-il-all", inference=True)
# Get the image size the model was trained on
size = birder.get_size_from_signature(model_info.signature)
# Create an inference transform
transform = birder.classification_transform(size, model_info.rgb_stats)
image = "path/to/image.jpeg" # or a PIL image
(out, embedding) = infer_image(net, image, transform, return_embedding=True)
# embedding is a NumPy array with shape of (1, 384)
```
### Detection Feature Map
```python
from PIL import Image
import birder
(net, model_info) = birder.load_pretrained_model("vit_parallel_s16_18x2_ls_avg_data2vec-intermediate-il-all", inference=True)
# Get the image size the model was trained on
size = birder.get_size_from_signature(model_info.signature)
# Create an inference transform
transform = birder.classification_transform(size, model_info.rgb_stats)
image = Image.open("path/to/image.jpeg")
features = net.detection_features(transform(image).unsqueeze(0))
# features is a dict (stage name -> torch.Tensor)
print([(k, v.size()) for k, v in features.items()])
# Output example:
# [('neck', torch.Size([1, 384, 24, 24]))]
```
## Citation
```bibtex
@misc{touvron2022thingsknowvisiontransformers,
title={Three things everyone should know about Vision Transformers},
author={Hugo Touvron and Matthieu Cord and Alaaeldin El-Nouby and Jakob Verbeek and Hervé Jégou},
year={2022},
eprint={2203.09795},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2203.09795},
}
@misc{https://doi.org/10.48550/arxiv.2202.03555,
title={data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language},
author={Alexei Baevski and Wei-Ning Hsu and Qiantong Xu and Arun Babu and Jiatao Gu and Michael Auli},
year={2022},
eprint={2202.03555},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2202.03555},
}
```
|