MetaFormer is a general architecture abstracted from Transformers by not specifying the token mixer.
Source: MetaFormer Is Actually What You Need for VisionPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Image Classification | 6 | 28.57% |
Semantic Segmentation | 3 | 14.29% |
Object Detection | 2 | 9.52% |
Object Recognition | 1 | 4.76% |
Image Categorization | 1 | 4.76% |
Decision Making | 1 | 4.76% |
Reinforcement Learning (RL) | 1 | 4.76% |
Decoder | 1 | 4.76% |
Image Segmentation | 1 | 4.76% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |