Source code for monai.networks.nets.vit

# Copyright (c) MONAI Consortium
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# See the License for the specific language governing permissions and
# limitations under the License.

from __future__ import annotations

from import Sequence

import torch
import torch.nn as nn

from monai.networks.blocks.patchembedding import PatchEmbeddingBlock
from monai.networks.blocks.transformerblock import TransformerBlock

__all__ = ["ViT"]

[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 <>" ViT supports Torchscript but only works for Pytorch after 1.8. """
[docs] def __init__( self, in_channels: int, img_size: Sequence[int] | int, patch_size: Sequence[int] | int, hidden_size: int = 768, mlp_dim: int = 3072, num_layers: int = 12, num_heads: int = 12, pos_embed: str = "conv", classification: bool = False, num_classes: int = 2, dropout_rate: float = 0.0, spatial_dims: int = 3, post_activation="Tanh", 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". classification (bool, optional): bool argument to determine if classification is used. Defaults to False. 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. post_activation (str, optional): add a final acivation function to the classification head when `classification` is True. Default to "Tanh" for `nn.Tanh()`. Set to other values to remove this function. 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. Examples:: # for single channel input with image size of (96,96,96), conv position embedding and segmentation backbone >>> net = ViT(in_channels=1, img_size=(96,96,96), pos_embed='conv') # for 3-channel with image size of (128,128,128), 24 layers and classification backbone >>> net = ViT(in_channels=3, img_size=(128,128,128), pos_embed='conv', classification=True) # for 3-channel with image size of (224,224), 12 layers and classification backbone >>> net = ViT(in_channels=3, img_size=(224,224), pos_embed='conv', classification=True, spatial_dims=2) """ super().__init__() if not (0 <= dropout_rate <= 1): raise ValueError("dropout_rate should be between 0 and 1.") if hidden_size % num_heads != 0: raise ValueError("hidden_size should be divisible by num_heads.") self.classification = classification self.patch_embedding = PatchEmbeddingBlock( in_channels=in_channels, img_size=img_size, patch_size=patch_size, hidden_size=hidden_size, num_heads=num_heads, pos_embed=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 i in range(num_layers) ] ) self.norm = nn.LayerNorm(hidden_size) if self.classification: self.cls_token = nn.Parameter(torch.zeros(1, 1, hidden_size)) if post_activation == "Tanh": self.classification_head = nn.Sequential(nn.Linear(hidden_size, num_classes), nn.Tanh()) else: self.classification_head = nn.Linear(hidden_size, num_classes) # type: ignore
[docs] def forward(self, x): x = self.patch_embedding(x) if hasattr(self, "cls_token"): cls_token = self.cls_token.expand(x.shape[0], -1, -1) x =, x), dim=1) hidden_states_out = [] for blk in self.blocks: x = blk(x) hidden_states_out.append(x) x = self.norm(x) if hasattr(self, "classification_head"): x = self.classification_head(x[:, 0]) return x, hidden_states_out