Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,54 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
---
|
4 |
+
|
5 |
+
|
6 |
+
|
7 |
+
<br>
|
8 |
+
|
9 |
+
# ViG Model Card
|
10 |
+
|
11 |
+
## Model Details
|
12 |
+
|
13 |
+
ViG is a generic backbone trained on the ImageNet-1K dataset for vision tasks.
|
14 |
+
|
15 |
+
- **Developed by:** [HUST](https://english.hust.edu.cn/), [Horizon Robotics](https://en.horizon.cc/)
|
16 |
+
- **Model type:** A generic vision backbone based on the Gated Linear Attention (GLA) architecture.
|
17 |
+
- **License:** Non-commercial license
|
18 |
+
|
19 |
+
|
20 |
+
### Model Sources
|
21 |
+
|
22 |
+
- **Repository:** https://github.com/hustvl/ViG
|
23 |
+
- **Paper:** https://arxiv.org/abs/2405.18425
|
24 |
+
|
25 |
+
## Uses
|
26 |
+
|
27 |
+
The primary use of ViG is research on vision tasks, e.g., classification, segmentation, detection, and instance segmentation, with an GLA-based backbone.
|
28 |
+
The primary intended users of the model are researchers and hobbyists in computer vision, machine learning, and artificial intelligence.
|
29 |
+
|
30 |
+
|
31 |
+
## Training Details
|
32 |
+
|
33 |
+
ViG is pretrained on ImageNet-1K with classification supervision.
|
34 |
+
The training data is around 1.3M images from [ImageNet-1K dataset](https://www.image-net.org/challenges/LSVRC/2012/).
|
35 |
+
See more details in this [paper](https://arxiv.org/abs/2405.18425).
|
36 |
+
|
37 |
+
## Evaluation
|
38 |
+
|
39 |
+
ViG is evaluated on ImageNet-1K val set, more details can be found in this [paper](https://arxiv.org/abs/2405.18425).
|
40 |
+
|
41 |
+
## Additional Information
|
42 |
+
|
43 |
+
## Citation Information
|
44 |
+
|
45 |
+
```
|
46 |
+
@article{vig,
|
47 |
+
title={ViG: Linear-complexity Visual Sequence Learning with Gated Linear Attention},
|
48 |
+
author={Bencheng Liao and Xinggang Wang and Lianghui Zhu and Qian Zhang and Chang Huang},
|
49 |
+
journal={arXiv preprint arXiv:2405.18425},
|
50 |
+
year={2024}
|
51 |
+
}
|
52 |
+
```
|
53 |
+
|
54 |
+
|