From 63810c0c84402ace38ac88dbdb3eddfdffe532b8 Mon Sep 17 00:00:00 2001 From: Gustaf Rydholm Date: Sat, 27 Nov 2021 12:48:35 +0100 Subject: Revert "Remove ViT" This reverts commit 4a6550ddef7d1f1971737bc22715db6381441f79. --- text_recognizer/networks/transformer/vit.py | 46 +++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) create mode 100644 text_recognizer/networks/transformer/vit.py diff --git a/text_recognizer/networks/transformer/vit.py b/text_recognizer/networks/transformer/vit.py new file mode 100644 index 0000000..ab331f8 --- /dev/null +++ b/text_recognizer/networks/transformer/vit.py @@ -0,0 +1,46 @@ +"""Vision Transformer.""" +from typing import Tuple, Type + +from einops.layers.torch import Rearrange +import torch +from torch import nn, Tensor + + +class ViT(nn.Module): + def __init__( + self, + image_size: Tuple[int, int], + patch_size: Tuple[int, int], + dim: int, + transformer: Type[nn.Module], + channels: int = 1, + ) -> None: + super().__init__() + img_height, img_width = image_size + patch_height, patch_width = patch_size + assert img_height % patch_height == 0 + assert img_width % patch_width == 0 + + num_patches = (img_height // patch_height) * (img_width // patch_width) + patch_dim = channels * patch_height * patch_width + + self.to_patch_embedding = nn.Sequential( + Rearrange( + "b c (h p1) (w p2) -> b (h w) (p1 p2 c)", + p1=patch_height, + p2=patch_width, + c=channels, + ), + nn.Linear(patch_dim, dim), + ) + + self.pos_embedding = nn.Parameter(torch.randn(1, num_patches, dim)) + self.transformer = transformer + self.norm = nn.LayerNorm(dim) + + def forward(self, img: Tensor) -> Tensor: + x = self.to_patch_embedding(img) + _, n, _ = x.shape + x += self.pos_embedding[:, :n] + x = self.transformer(x) + return x -- cgit v1.2.3-70-g09d2