https://github.com/microsoft/Swin-Transformer

This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows".
https://github.com/microsoft/Swin-Transformer

ade20k image-classification imagenet mask-rcnn mscoco object-detection semantic-segmentation swin-transformer

Last synced: about 2 months ago
JSON representation

This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows".

• Host: GitHub
• URL: https://github.com/microsoft/Swin-Transformer
• Owner: microsoft
• License: mit
• Created: 2021-03-25T12:42:36.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2024-07-15T15:00:33.000Z (2 months ago)
• Last Synced: 2024-07-15T18:12:06.277Z (2 months ago)
• Topics: ade20k, image-classification, imagenet, mask-rcnn, mscoco, object-detection, semantic-segmentation, swin-transformer
• Language: Python
• Homepage: https://arxiv.org/abs/2103.14030
• Size: 1.05 MB
• Stars: 13,331
• Watchers: 128
• Forks: 2,025
• Open Issues: 183
• Metadata Files:
  • Readme: README.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md
  • Security: SECURITY.md
  • Support: SUPPORT.md

Awesome Lists containing this project

• Awesome-Reasoning-Foundation-Models - [code
• StarryDivineSky - microsoft/Swin-Transformer - attention 计算限制在不重叠的本地窗口上，同时还允许跨窗口连接，从而带来更高的效率。将 CLIP 预训练的 ViT-L 提高了 +1.6%，以达到ImageNet-1K 图像分类，这是最准确的 ViT-L 模型。在 COCO 对象检测（58.7 box AP和51.1 mask APtest-dev）和 ADE20K 语义分割（53.5 mIoU在 val）上实现了强大的性能，大大超过了以前的模型。 (对象检测、分割 / 网络服务_其他)