from typing import Sequence, Union
import torch
import torch.nn as nn
from fusionlab.layers import (
PatchEmbedding,
SelfAttention,
)
[docs]
class MLPBlock(nn.Module):
"""
A multi-layer perceptron block, based on: "Dosovitskiy et al.,
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale <https://arxiv.org/abs/2010.11929>"
"""
[docs]
def __init__(
self,
hidden_size: int,
mlp_dim: int,
dropout_rate: float = 0.0,
act: nn.Module = nn.GELU,
) -> None:
"""
Args:
hidden_size: dimension of hidden layer.
mlp_dim: dimension of feedforward layer. If 0, `hidden_size` will be used.
dropout_rate: faction of the input units to drop.
act: activation type and arguments. Defaults to nn.GELU
"""
super().__init__()
if not (0 <= dropout_rate <= 1):
raise ValueError("dropout_rate should be between 0 and 1.")
mlp_dim = mlp_dim or hidden_size
self.linear1 = nn.Linear(hidden_size, mlp_dim)
self.linear2 = nn.Linear(mlp_dim, hidden_size)
self.act = act()
self.drop1 = nn.Dropout(dropout_rate)
self.drop2 = nn.Dropout(dropout_rate)
[docs]
def forward(self, x):
x = self.act(self.linear1(x))
x = self.drop1(x)
x = self.linear2(x)
x = self.drop2(x)
return x
[docs]
class ViT(nn.Module):
"""
Vision Transformer (ViT), based on: "Dosovitskiy et al.,
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale <https://arxiv.org/abs/2010.11929>"
ViT supports Torchscript but only works for Pytorch after 1.8.
source code: https://github.com/Project-MONAI/MONAI/blob/main/monai/networks/nets/vit.py
"""
[docs]
def __init__(
self,
in_channels: int,
img_size: Union[Sequence[int], int],
patch_size: Union[Sequence[int], int],
hidden_size: int = 768,
mlp_dim: int = 3072,
num_layers: int = 12,
num_heads: int = 12,
pos_embed: str = "conv",
dropout_rate: float = 0.0,
spatial_dims: int = 2,
qkv_bias: bool = False,
save_attn: bool = False,
) -> None:
"""
Args:
in_channels (int): dimension of input channels.
img_size (Union[Sequence[int], int]): dimension of input image.
patch_size (Union[Sequence[int], int]): dimension of patch size.
hidden_size (int, optional): dimension of hidden layer. Defaults to 768.
mlp_dim (int, optional): dimension of feedforward layer. Defaults to 3072.
num_layers (int, optional): number of transformer blocks. Defaults to 12.
num_heads (int, optional): number of attention heads. Defaults to 12.
pos_embed (str, optional): position embedding layer type. Defaults to "conv".
num_classes (int, optional): number of classes if classification is used. Defaults to 2.
dropout_rate (float, optional): faction of the input units to drop. Defaults to 0.0.
spatial_dims (int, optional): number of spatial dimensions. Defaults to 3.
qkv_bias (bool, optional): apply bias to the qkv linear layer in self attention block. Defaults to False.
save_attn (bool, optional): to make accessible the attention in self attention block. Defaults to False.
"""
super().__init__()
if hidden_size % num_heads != 0:
raise ValueError("hidden_size should be divisible by num_heads.")
self.patch_embedding = PatchEmbedding(
in_channels=in_channels,
img_size=img_size,
patch_size=patch_size,
hidden_size=hidden_size,
pos_embed_type=pos_embed,
dropout_rate=dropout_rate,
spatial_dims=spatial_dims,
)
self.blocks = nn.ModuleList(
[
TransformerBlock(hidden_size, mlp_dim, num_heads, dropout_rate, qkv_bias, save_attn)
for _ in range(num_layers)
]
)
self.norm = nn.LayerNorm(hidden_size)
[docs]
def forward(self, x, return_features=False):
x = self.patch_embedding(x)
features = []
for block in self.blocks:
x = block(x)
features.append(x)
x = self.norm(x)
if return_features:
return x, features
else:
return x
VisionTransformer = ViT
if __name__ == '__main__':
inputs = torch.randn(1, 3, 224, 224)
model = ViT(
in_channels=3,
img_size=224,
patch_size=16,
hidden_size=768,
mlp_dim=3072,
num_layers=2,
# num_layers=12,
num_heads=12,
)
outputs = model(inputs)
print(outputs.shape)
outputs, hidden = model(inputs, return_features=True)
print(outputs.shape)
[print(i.shape) for i in hidden]