From c90a69b41493a280085b373420031184886fefa1 Mon Sep 17 00:00:00 2001
From: Gustaf Rydholm <gustaf.rydholm@gmail.com>
Date: Mon, 11 Sep 2023 22:14:58 +0200
Subject: Add MaMMUT network

---
 text_recognizer/network/mammut.py | 99 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 99 insertions(+)
 create mode 100644 text_recognizer/network/mammut.py

(limited to 'text_recognizer/network')

diff --git a/text_recognizer/network/mammut.py b/text_recognizer/network/mammut.py
new file mode 100644
index 0000000..761e5f9
--- /dev/null
+++ b/text_recognizer/network/mammut.py
@@ -0,0 +1,99 @@
+from typing import Type
+
+from einops import repeat
+import torch
+from torch import nn, Tensor
+import torch.nn.functional as F
+
+from text_recognizer.network.transformer.decoder import Decoder
+from text_recognizer.network.transformer.attention import Attention
+
+
+class ToLatent(nn.Module):
+    def __init__(self, dim: int, dim_latent: int) -> None:
+        super().__init__()
+        self.to_latent = nn.Linear(dim, dim_latent, bias=False)
+
+    def forward(self, x: Tensor) -> Tensor:
+        return F.normalize(self.to_latent(x), dim=-1)
+
+
+class MaMMUT(nn.Module):
+    def __init__(
+        self,
+        encoder: Type[nn.Module],
+        image_attn_pool: Attention,
+        decoder: Decoder,
+        dim: int,
+        dim_latent: int,
+        num_tokens: int,
+        pad_index: int,
+        num_image_queries: int = 256,
+    ) -> None:
+        super().__init__()
+        # Image encoder
+        self.encoder = encoder
+        self.image_queries = nn.Parameter(torch.randn(num_image_queries + 1, dim))
+        self.image_attn_pool = image_attn_pool
+        self.image_attn_pool_norm = nn.LayerNorm(dim)
+
+        # Text decoder
+        self.decoder = decoder
+        self.pad_index = pad_index
+        self.token_embeddings = nn.Embedding(num_tokens, dim)
+        self.text_cls_token = nn.Parameter(torch.randn(dim))
+        self.text_cls_norm = nn.LayerNorm(dim)
+
+        # To latents
+        self.to_image_latent = ToLatent(dim, dim_latent)
+        self.to_text_latent = ToLatent(dim, dim_latent)
+
+        self.to_logits = nn.Sequential(
+            nn.LayerNorm(dim), nn.Linear(dim, num_tokens, bias=False)
+        )
+        self.to_logits[-1].weight = self.token_embeddings.weight
+        nn.init.normal_(self.token_embeddings.weight, std=0.02)
+
+    def to_image_features(self, images: Tensor) -> Tensor:
+        image_tokens = self.encoder(images)
+        queries = repeat(self.image_queries, "n d -> b n d", b=image_tokens.shape[0])
+        queries = self.image_attn_pool(queries, image_tokens)
+        queries = self.image_attn_pool_norm(queries)
+        return queries[:, 0], queries[:, 1:]
+
+    def to_text_cls_features(self, text: Tensor) -> Tensor:
+        b = text.shape[0]
+        text = text.long()
+        mask = text != self.pad_index
+        tokens = self.token_embeddings(text)
+
+        # Append cls token
+        cls_tokens = repeat(self.text_cls_token, "d -> b 1 d", b=b)
+        tokens = torch.cat([tokens, cls_tokens], dim=-2)
+
+        # Create input mask
+        mask = F.pad(mask, pad=(0, 1), value=True)
+
+        features = self.decoder.self_attn(tokens, mask)
+        cls_embedding = features[:, -1]
+        return self.text_cls_norm(cls_embedding)
+
+    def to_latents(self, image_embeddings: Tensor, text_embeddings: Tensor) -> Tensor:
+        image_latent = self.to_image_latent(image_embeddings)
+        text_latent = self.to_text_latent(text_embeddings)
+        return image_latent, text_latent
+
+    def decode(self, text: Tensor, image_features: Tensor) -> Tensor:
+        text = text.long()
+        mask = text != self.pad_index
+        tokens = self.token_embeddings(text)
+        text_features = self.decoder(tokens, context=image_features, mask=mask)
+        return self.to_logits(text_features)
+
+    def encode(self, images: Tensor) -> Tensor:
+        _, image_features = self.to_image_features(images)
+        return image_features
+
+    def forward(self, images: Tensor, text: Tensor) -> Tensor:
+        image_features = self.encode(images)
+        return self.decode(text, image_features)
-- 
cgit v1.2.3-70-g09d2