2D-Positional-Encoding-Vision-Transformer icon indicating copy to clipboard operation
2D-Positional-Encoding-Vision-Transformer copied to clipboard

Published 1 year ago • verified publisher icon s-chh
→

Metadata

PyTorch implementation of 2D Positional Encodings for Vision Transformers (ViT). Positional Encodings/Embeddings: Sinusoidal (Absolute), Learnable, Relative and Rotation (Rope).

← Metadata

36
Stars
4
Forks
36
Watchers

Owner

verified publisher icon s-chh

Metadata

PyTorch implementation of 2D Positional Encodings for Vision Transformers (ViT). Positional Encodings/Embeddings: Sinusoidal (Absolute), Learnable, Relative and Rotation (Rope).

Back