【Timm】create_model所提供的ViT模型概览

 ⚪查看代码：python xxx.py 

import timm
 
if __name__ == '__main__':
    model_vit = timm.list_models('*vit*')
    print(len(model_vit),model_vit[:])

⚪结合vision transformer理解

7 ResNets:

• R50x1, R50x2 R101x1, R152x1, R152x2, pre-trained for 7 epochs,
• plus R152x2 and R200x3 pre-trained for 14 epochs;

6 Vision Transformers: 

• ViT-B/32, B/16, L/32, L/16, pre-trained for 7 epochs,
• plus L/16 and H/14 pre-trained for 14 epochs;

5 hybrids,

• R50+ViT-B/32, B/16, L/32, L/16 pretrained for 7 epochs,
• plus R50+ViT-L/16 pre-trained for 14 epochs

参数解读：

• 以ViT-L/16为例，表示ViT Large模型，对应patch_size为16。
• 但是，混合模型的数值不是对应patch_size，而是ResNet的总取样率。
• 采样：模拟信号进行取样时的快慢次数
• 这里就能对Timm库所提供的预训练模型有所理解。

⚪ViT_model概览-28个

1. 'vit_base_patch16_224',
2. 'vit_base_patch16_224_in21k',
3. 'vit_base_patch16_384',
4. 'vit_base_patch32_224',
5. 'vit_base_patch32_224_in21k',
6. 'vit_base_patch32_384',
7. 'vit_base_resnet26d_224',
8. 'vit_base_resnet50_224_in21k',
9. 'vit_base_resnet50_384',
10. 'vit_base_resnet50d_224',
11. 'vit_deit_base_distilled_patch16_224',
12. 'vit_deit_base_distilled_patch16_384',
13. 'vit_deit_base_patch16_224',
14. 'vit_deit_base_patch16_384',
15. 'vit_deit_small_distilled_patch16_224',
16. 'vit_deit_small_patch16_224',
17. 'vit_deit_tiny_distilled_patch16_224',
18. 'vit_deit_tiny_patch16_224',
19. 'vit_huge_patch14_224_in21k',
20. 'vit_large_patch16_224',
21. 'vit_large_patch16_224_in21k',
22. 'vit_large_patch16_384',
23. 'vit_large_patch32_224',
24. 'vit_large_patch32_224_in21k',
25. 'vit_large_patch32_384',
26. 'vit_small_patch16_224',
27. 'vit_small_resnet26d_224',
28. 'vit_small_resnet50d_s3_224'

文章推荐：

• Pytorch视觉模型库--timm_pytorch 模型库
• pytorch下的迁移学习模型库·详细使用教程