|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
from typing import Sequence, Tuple, Type, Union |
|
|
|
import numpy as np |
|
import torch |
|
import torch.nn as nn |
|
import torch.nn.functional as F |
|
import torch.utils.checkpoint as checkpoint |
|
from torch.nn import LayerNorm |
|
|
|
from monai.networks.blocks import MLPBlock as Mlp |
|
from monai.networks.blocks import PatchEmbed, UnetOutBlock, UnetrBasicBlock, UnetrUpBlock |
|
from monai.networks.layers import DropPath, trunc_normal_ |
|
from monai.utils import ensure_tuple_rep, optional_import |
|
import math |
|
|
|
rearrange, _ = optional_import("einops", name="rearrange") |
|
|
|
|
|
class TextSwinUNETR(nn.Module): |
|
""" |
|
Swin UNETR based on: "Hatamizadeh et al., |
|
Swin UNETR: Swin Transformers for Semantic Segmentation of Brain Tumors in MRI Images |
|
<https://arxiv.org/abs/2201.01266>" |
|
""" |
|
|
|
def __init__( |
|
self, |
|
img_size: Union[Sequence[int], int], |
|
in_channels: int, |
|
out_channels: int, |
|
text_dim: int, |
|
depths: Sequence[int] = (2, 2, 2, 2), |
|
num_heads: Sequence[int] = (3, 6, 12, 24), |
|
feature_size: int = 24, |
|
norm_name: Union[Tuple, str] = "instance", |
|
drop_rate: float = 0.0, |
|
attn_drop_rate: float = 0.0, |
|
dropout_path_rate: float = 0.0, |
|
normalize: bool = True, |
|
use_checkpoint: bool = False, |
|
spatial_dims: int = 3, |
|
) -> None: |
|
""" |
|
Args: |
|
img_size: dimension of input image. |
|
in_channels: dimension of input channels. |
|
out_channels: dimension of output channels. |
|
feature_size: dimension of network feature size. |
|
depths: number of layers in each stage. |
|
num_heads: number of attention heads. |
|
norm_name: feature normalization type and arguments. |
|
drop_rate: dropout rate. |
|
attn_drop_rate: attention dropout rate. |
|
dropout_path_rate: drop path rate. |
|
normalize: normalize output intermediate features in each stage. |
|
use_checkpoint: use gradient checkpointing for reduced memory usage. |
|
spatial_dims: number of spatial dims. |
|
|
|
Examples:: |
|
|
|
# for 3D single channel input with size (96,96,96), 4-channel output and feature size of 48. |
|
#>>> net = SwinUNETR(img_size=(96,96,96), in_channels=1, out_channels=4, feature_size=48) |
|
|
|
# for 3D 4-channel input with size (128,128,128), 3-channel output and (2,4,2,2) layers in each stage. |
|
#>>> net = SwinUNETR(img_size=(128,128,128), in_channels=4, out_channels=3, depths=(2,4,2,2)) |
|
|
|
# for 2D single channel input with size (96,96), 2-channel output and gradient checkpointing. |
|
#>>> net = SwinUNETR(img_size=(96,96), in_channels=3, out_channels=2, use_checkpoint=True, spatial_dims=2) |
|
|
|
""" |
|
|
|
super().__init__() |
|
|
|
img_size = ensure_tuple_rep(img_size, spatial_dims) |
|
patch_size = ensure_tuple_rep(2, spatial_dims) |
|
window_size = ensure_tuple_rep(7, spatial_dims) |
|
|
|
if not (spatial_dims == 2 or spatial_dims == 3): |
|
raise ValueError("spatial dimension should be 2 or 3.") |
|
|
|
for m, p in zip(img_size, patch_size): |
|
for i in range(5): |
|
if m % np.power(p, i + 1) != 0: |
|
raise ValueError("input image size (img_size) should be divisible by stage-wise image resolution.") |
|
|
|
if not (0 <= drop_rate <= 1): |
|
raise ValueError("dropout rate should be between 0 and 1.") |
|
|
|
if not (0 <= attn_drop_rate <= 1): |
|
raise ValueError("attention dropout rate should be between 0 and 1.") |
|
|
|
if not (0 <= dropout_path_rate <= 1): |
|
raise ValueError("drop path rate should be between 0 and 1.") |
|
|
|
if feature_size % 12 != 0: |
|
raise ValueError("feature_size should be divisible by 12.") |
|
|
|
self.normalize = normalize |
|
|
|
self.swinViT = SwinTransformer( |
|
in_chans=in_channels, |
|
embed_dim=feature_size, |
|
window_size=window_size, |
|
patch_size=patch_size, |
|
depths=depths, |
|
num_heads=num_heads, |
|
mlp_ratio=4.0, |
|
qkv_bias=True, |
|
drop_rate=drop_rate, |
|
attn_drop_rate=attn_drop_rate, |
|
drop_path_rate=dropout_path_rate, |
|
norm_layer=nn.LayerNorm, |
|
use_checkpoint=use_checkpoint, |
|
spatial_dims=spatial_dims, |
|
text_dim=text_dim, |
|
) |
|
|
|
self.encoder1 = UnetrBasicBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=in_channels, |
|
out_channels=feature_size, |
|
kernel_size=3, |
|
stride=1, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.encoder2 = UnetrBasicBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=feature_size, |
|
out_channels=feature_size, |
|
kernel_size=3, |
|
stride=1, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.encoder3 = UnetrBasicBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=2 * feature_size, |
|
out_channels=2 * feature_size, |
|
kernel_size=3, |
|
stride=1, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.encoder4 = UnetrBasicBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=4 * feature_size, |
|
out_channels=4 * feature_size, |
|
kernel_size=3, |
|
stride=1, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.encoder10 = UnetrBasicBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=16 * feature_size, |
|
out_channels=16 * feature_size, |
|
kernel_size=3, |
|
stride=1, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.decoder5 = UnetrUpBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=16 * feature_size, |
|
out_channels=8 * feature_size, |
|
kernel_size=3, |
|
upsample_kernel_size=2, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.decoder4 = UnetrUpBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=feature_size * 8, |
|
out_channels=feature_size * 4, |
|
kernel_size=3, |
|
upsample_kernel_size=2, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.decoder3 = UnetrUpBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=feature_size * 4, |
|
out_channels=feature_size * 2, |
|
kernel_size=3, |
|
upsample_kernel_size=2, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
self.decoder2 = UnetrUpBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=feature_size * 2, |
|
out_channels=feature_size, |
|
kernel_size=3, |
|
upsample_kernel_size=2, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.decoder1 = UnetrUpBlock( |
|
spatial_dims=spatial_dims, |
|
in_channels=feature_size, |
|
out_channels=feature_size, |
|
kernel_size=3, |
|
upsample_kernel_size=2, |
|
norm_name=norm_name, |
|
res_block=True, |
|
) |
|
|
|
self.out = UnetOutBlock( |
|
spatial_dims=spatial_dims, in_channels=feature_size, out_channels=out_channels |
|
) |
|
|
|
def load_from(self, weights): |
|
|
|
with torch.no_grad(): |
|
self.swinViT.patch_embed.proj.weight.copy_(weights["state_dict"]["module.patch_embed.proj.weight"]) |
|
self.swinViT.patch_embed.proj.bias.copy_(weights["state_dict"]["module.patch_embed.proj.bias"]) |
|
for bname, block in self.swinViT.layers1[0].blocks.named_children(): |
|
block.load_from(weights, n_block=bname, layer="layers1") |
|
self.swinViT.layers1[0].downsample.reduction.weight.copy_( |
|
weights["state_dict"]["module.layers1.0.downsample.reduction.weight"] |
|
) |
|
self.swinViT.layers1[0].downsample.norm.weight.copy_( |
|
weights["state_dict"]["module.layers1.0.downsample.norm.weight"] |
|
) |
|
self.swinViT.layers1[0].downsample.norm.bias.copy_( |
|
weights["state_dict"]["module.layers1.0.downsample.norm.bias"] |
|
) |
|
for bname, block in self.swinViT.layers2[0].blocks.named_children(): |
|
block.load_from(weights, n_block=bname, layer="layers2") |
|
self.swinViT.layers2[0].downsample.reduction.weight.copy_( |
|
weights["state_dict"]["module.layers2.0.downsample.reduction.weight"] |
|
) |
|
self.swinViT.layers2[0].downsample.norm.weight.copy_( |
|
weights["state_dict"]["module.layers2.0.downsample.norm.weight"] |
|
) |
|
self.swinViT.layers2[0].downsample.norm.bias.copy_( |
|
weights["state_dict"]["module.layers2.0.downsample.norm.bias"] |
|
) |
|
for bname, block in self.swinViT.layers3[0].blocks.named_children(): |
|
block.load_from(weights, n_block=bname, layer="layers3") |
|
self.swinViT.layers3[0].downsample.reduction.weight.copy_( |
|
weights["state_dict"]["module.layers3.0.downsample.reduction.weight"] |
|
) |
|
self.swinViT.layers3[0].downsample.norm.weight.copy_( |
|
weights["state_dict"]["module.layers3.0.downsample.norm.weight"] |
|
) |
|
self.swinViT.layers3[0].downsample.norm.bias.copy_( |
|
weights["state_dict"]["module.layers3.0.downsample.norm.bias"] |
|
) |
|
for bname, block in self.swinViT.layers4[0].blocks.named_children(): |
|
block.load_from(weights, n_block=bname, layer="layers4") |
|
self.swinViT.layers4[0].downsample.reduction.weight.copy_( |
|
weights["state_dict"]["module.layers4.0.downsample.reduction.weight"] |
|
) |
|
self.swinViT.layers4[0].downsample.norm.weight.copy_( |
|
weights["state_dict"]["module.layers4.0.downsample.norm.weight"] |
|
) |
|
self.swinViT.layers4[0].downsample.norm.bias.copy_( |
|
weights["state_dict"]["module.layers4.0.downsample.norm.bias"] |
|
) |
|
|
|
def forward(self, x_in, text_in): |
|
hidden_states_out = self.swinViT(x_in, text_in, self.normalize) |
|
enc0 = self.encoder1(x_in) |
|
enc1 = self.encoder2(hidden_states_out[0]) |
|
enc2 = self.encoder3(hidden_states_out[1]) |
|
enc3 = self.encoder4(hidden_states_out[2]) |
|
dec4 = self.encoder10(hidden_states_out[4]) |
|
dec3 = self.decoder5(dec4, hidden_states_out[3]) |
|
dec2 = self.decoder4(dec3, enc3) |
|
dec1 = self.decoder3(dec2, enc2) |
|
dec0 = self.decoder2(dec1, enc1) |
|
out = self.decoder1(dec0, enc0) |
|
logits = self.out(out) |
|
return logits |
|
|
|
|
|
def window_partition(x, window_size): |
|
"""window partition operation based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
|
|
Args: |
|
x: input tensor. |
|
window_size: local window size. |
|
""" |
|
x_shape = x.size() |
|
if len(x_shape) == 5: |
|
b, d, h, w, c = x_shape |
|
x = x.view( |
|
b, |
|
d // window_size[0], |
|
window_size[0], |
|
h // window_size[1], |
|
window_size[1], |
|
w // window_size[2], |
|
window_size[2], |
|
c, |
|
) |
|
windows = ( |
|
x.permute(0, 1, 3, 5, 2, 4, 6, 7).contiguous().view(-1, window_size[0] * window_size[1] * window_size[2], c) |
|
) |
|
elif len(x_shape) == 4: |
|
b, h, w, c = x.shape |
|
x = x.view(b, h // window_size[0], window_size[0], w // window_size[1], window_size[1], c) |
|
windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size[0] * window_size[1], c) |
|
return windows |
|
|
|
|
|
def window_reverse(windows, window_size, dims): |
|
"""window reverse operation based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
|
|
Args: |
|
windows: windows tensor. |
|
window_size: local window size. |
|
dims: dimension values. |
|
""" |
|
if len(dims) == 4: |
|
b, d, h, w = dims |
|
x = windows.view( |
|
b, |
|
d // window_size[0], |
|
h // window_size[1], |
|
w // window_size[2], |
|
window_size[0], |
|
window_size[1], |
|
window_size[2], |
|
-1, |
|
) |
|
x = x.permute(0, 1, 4, 2, 5, 3, 6, 7).contiguous().view(b, d, h, w, -1) |
|
|
|
elif len(dims) == 3: |
|
b, h, w = dims |
|
x = windows.view(b, h // window_size[0], w // window_size[0], window_size[0], window_size[1], -1) |
|
x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(b, h, w, -1) |
|
return x |
|
|
|
|
|
def get_window_size(x_size, window_size, shift_size=None): |
|
"""Computing window size based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
|
|
Args: |
|
x_size: input size. |
|
window_size: local window size. |
|
shift_size: window shifting size. |
|
""" |
|
|
|
use_window_size = list(window_size) |
|
if shift_size is not None: |
|
use_shift_size = list(shift_size) |
|
for i in range(len(x_size)): |
|
if x_size[i] <= window_size[i]: |
|
use_window_size[i] = x_size[i] |
|
if shift_size is not None: |
|
use_shift_size[i] = 0 |
|
|
|
if shift_size is None: |
|
return tuple(use_window_size) |
|
else: |
|
return tuple(use_window_size), tuple(use_shift_size) |
|
|
|
|
|
class WindowAttention(nn.Module): |
|
""" |
|
Window based multi-head self attention module with relative position bias based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
""" |
|
|
|
def __init__( |
|
self, |
|
dim: int, |
|
num_heads: int, |
|
window_size: Sequence[int], |
|
qkv_bias: bool = False, |
|
attn_drop: float = 0.0, |
|
proj_drop: float = 0.0, |
|
) -> None: |
|
""" |
|
Args: |
|
dim: number of feature channels. |
|
num_heads: number of attention heads. |
|
window_size: local window size. |
|
qkv_bias: add a learnable bias to query, key, value. |
|
attn_drop: attention dropout rate. |
|
proj_drop: dropout rate of output. |
|
""" |
|
|
|
super().__init__() |
|
self.dim = dim |
|
self.window_size = window_size |
|
self.num_heads = num_heads |
|
head_dim = dim // num_heads |
|
self.scale = head_dim**-0.5 |
|
mesh_args = torch.meshgrid.__kwdefaults__ |
|
|
|
if len(self.window_size) == 3: |
|
self.relative_position_bias_table = nn.Parameter( |
|
torch.zeros( |
|
(2 * self.window_size[0] - 1) * (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1), |
|
num_heads, |
|
) |
|
) |
|
coords_d = torch.arange(self.window_size[0]) |
|
coords_h = torch.arange(self.window_size[1]) |
|
coords_w = torch.arange(self.window_size[2]) |
|
if mesh_args is not None: |
|
coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w, indexing="ij")) |
|
else: |
|
coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w)) |
|
coords_flatten = torch.flatten(coords, 1) |
|
|
|
relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] |
|
relative_coords = relative_coords.permute(1, 2, 0).contiguous() |
|
relative_coords[:, :, 0] += self.window_size[0] - 1 |
|
relative_coords[:, :, 1] += self.window_size[1] - 1 |
|
relative_coords[:, :, 2] += self.window_size[2] - 1 |
|
relative_coords[:, :, 0] *= (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1) |
|
relative_coords[:, :, 1] *= 2 * self.window_size[2] - 1 |
|
elif len(self.window_size) == 2: |
|
self.relative_position_bias_table = nn.Parameter( |
|
torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads) |
|
) |
|
coords_h = torch.arange(self.window_size[0]) |
|
coords_w = torch.arange(self.window_size[1]) |
|
if mesh_args is not None: |
|
coords = torch.stack(torch.meshgrid(coords_h, coords_w, indexing="ij")) |
|
else: |
|
coords = torch.stack(torch.meshgrid(coords_h, coords_w)) |
|
coords_flatten = torch.flatten(coords, 1) |
|
relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] |
|
relative_coords = relative_coords.permute(1, 2, 0).contiguous() |
|
relative_coords[:, :, 0] += self.window_size[0] - 1 |
|
relative_coords[:, :, 1] += self.window_size[1] - 1 |
|
relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1 |
|
|
|
relative_position_index = relative_coords.sum(-1) |
|
self.register_buffer("relative_position_index", relative_position_index) |
|
self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias) |
|
self.attn_drop = nn.Dropout(attn_drop) |
|
self.proj = nn.Linear(dim, dim) |
|
self.proj_drop = nn.Dropout(proj_drop) |
|
trunc_normal_(self.relative_position_bias_table, std=0.02) |
|
self.softmax = nn.Softmax(dim=-1) |
|
|
|
def forward(self, x, mask): |
|
b, n, c = x.shape |
|
qkv = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4) |
|
q, k, v = qkv[0], qkv[1], qkv[2] |
|
q = q * self.scale |
|
attn = q @ k.transpose(-2, -1) |
|
relative_position_bias = self.relative_position_bias_table[ |
|
self.relative_position_index[:n, :n].reshape(-1) |
|
].reshape(n, n, -1) |
|
relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous() |
|
attn = attn + relative_position_bias.unsqueeze(0) |
|
if mask is not None: |
|
nw = mask.shape[0] |
|
attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0) |
|
attn = attn.view(-1, self.num_heads, n, n) |
|
attn = self.softmax(attn) |
|
else: |
|
attn = self.softmax(attn) |
|
|
|
attn = self.attn_drop(attn) |
|
x = (attn @ v).transpose(1, 2).reshape(b, n, c) |
|
x = self.proj(x) |
|
x = self.proj_drop(x) |
|
return x |
|
|
|
|
|
class SwinTransformerBlock(nn.Module): |
|
""" |
|
Swin Transformer block based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
""" |
|
|
|
def __init__( |
|
self, |
|
dim: int, |
|
num_heads: int, |
|
window_size: Sequence[int], |
|
shift_size: Sequence[int], |
|
mlp_ratio: float = 4.0, |
|
qkv_bias: bool = True, |
|
drop: float = 0.0, |
|
attn_drop: float = 0.0, |
|
drop_path: float = 0.0, |
|
act_layer: str = "GELU", |
|
norm_layer: Type[LayerNorm] = nn.LayerNorm, |
|
use_checkpoint: bool = False, |
|
) -> None: |
|
""" |
|
Args: |
|
dim: number of feature channels. |
|
num_heads: number of attention heads. |
|
window_size: local window size. |
|
shift_size: window shift size. |
|
mlp_ratio: ratio of mlp hidden dim to embedding dim. |
|
qkv_bias: add a learnable bias to query, key, value. |
|
drop: dropout rate. |
|
attn_drop: attention dropout rate. |
|
drop_path: stochastic depth rate. |
|
act_layer: activation layer. |
|
norm_layer: normalization layer. |
|
use_checkpoint: use gradient checkpointing for reduced memory usage. |
|
""" |
|
|
|
super().__init__() |
|
self.dim = dim |
|
self.num_heads = num_heads |
|
self.window_size = window_size |
|
self.shift_size = shift_size |
|
self.mlp_ratio = mlp_ratio |
|
self.use_checkpoint = use_checkpoint |
|
self.norm1 = norm_layer(dim) |
|
self.attn = WindowAttention( |
|
dim, |
|
window_size=self.window_size, |
|
num_heads=num_heads, |
|
qkv_bias=qkv_bias, |
|
attn_drop=attn_drop, |
|
proj_drop=drop, |
|
) |
|
|
|
self.drop_path = DropPath(drop_path) if drop_path > 0.0 else nn.Identity() |
|
self.norm2 = norm_layer(dim) |
|
mlp_hidden_dim = int(dim * mlp_ratio) |
|
self.mlp = Mlp(hidden_size=dim, mlp_dim=mlp_hidden_dim, act=act_layer, dropout_rate=drop, dropout_mode="swin") |
|
|
|
def forward_part1(self, x, mask_matrix): |
|
x_shape = x.size() |
|
x = self.norm1(x) |
|
if len(x_shape) == 5: |
|
b, d, h, w, c = x.shape |
|
window_size, shift_size = get_window_size((d, h, w), self.window_size, self.shift_size) |
|
pad_l = pad_t = pad_d0 = 0 |
|
pad_d1 = (window_size[0] - d % window_size[0]) % window_size[0] |
|
pad_b = (window_size[1] - h % window_size[1]) % window_size[1] |
|
pad_r = (window_size[2] - w % window_size[2]) % window_size[2] |
|
x = F.pad(x, (0, 0, pad_l, pad_r, pad_t, pad_b, pad_d0, pad_d1)) |
|
_, dp, hp, wp, _ = x.shape |
|
dims = [b, dp, hp, wp] |
|
|
|
elif len(x_shape) == 4: |
|
b, h, w, c = x.shape |
|
window_size, shift_size = get_window_size((h, w), self.window_size, self.shift_size) |
|
pad_l = pad_t = 0 |
|
pad_r = (window_size[0] - h % window_size[0]) % window_size[0] |
|
pad_b = (window_size[1] - w % window_size[1]) % window_size[1] |
|
x = F.pad(x, (0, 0, pad_l, pad_r, pad_t, pad_b)) |
|
_, hp, wp, _ = x.shape |
|
dims = [b, hp, wp] |
|
|
|
if any(i > 0 for i in shift_size): |
|
if len(x_shape) == 5: |
|
shifted_x = torch.roll(x, shifts=(-shift_size[0], -shift_size[1], -shift_size[2]), dims=(1, 2, 3)) |
|
elif len(x_shape) == 4: |
|
shifted_x = torch.roll(x, shifts=(-shift_size[0], -shift_size[1]), dims=(1, 2)) |
|
attn_mask = mask_matrix |
|
else: |
|
shifted_x = x |
|
attn_mask = None |
|
x_windows = window_partition(shifted_x, window_size) |
|
attn_windows = self.attn(x_windows, mask=attn_mask) |
|
attn_windows = attn_windows.view(-1, *(window_size + (c,))) |
|
shifted_x = window_reverse(attn_windows, window_size, dims) |
|
if any(i > 0 for i in shift_size): |
|
if len(x_shape) == 5: |
|
x = torch.roll(shifted_x, shifts=(shift_size[0], shift_size[1], shift_size[2]), dims=(1, 2, 3)) |
|
elif len(x_shape) == 4: |
|
x = torch.roll(shifted_x, shifts=(shift_size[0], shift_size[1]), dims=(1, 2)) |
|
else: |
|
x = shifted_x |
|
|
|
if len(x_shape) == 5: |
|
if pad_d1 > 0 or pad_r > 0 or pad_b > 0: |
|
x = x[:, :d, :h, :w, :].contiguous() |
|
elif len(x_shape) == 4: |
|
if pad_r > 0 or pad_b > 0: |
|
x = x[:, :h, :w, :].contiguous() |
|
|
|
return x |
|
|
|
def forward_part2(self, x): |
|
return self.drop_path(self.mlp(self.norm2(x))) |
|
|
|
def load_from(self, weights, n_block, layer): |
|
root = f"module.{layer}.0.blocks.{n_block}." |
|
block_names = [ |
|
"norm1.weight", |
|
"norm1.bias", |
|
"attn.relative_position_bias_table", |
|
"attn.relative_position_index", |
|
"attn.qkv.weight", |
|
"attn.qkv.bias", |
|
"attn.proj.weight", |
|
"attn.proj.bias", |
|
"norm2.weight", |
|
"norm2.bias", |
|
"mlp.fc1.weight", |
|
"mlp.fc1.bias", |
|
"mlp.fc2.weight", |
|
"mlp.fc2.bias", |
|
] |
|
with torch.no_grad(): |
|
self.norm1.weight.copy_(weights["state_dict"][root + block_names[0]]) |
|
self.norm1.bias.copy_(weights["state_dict"][root + block_names[1]]) |
|
self.attn.relative_position_bias_table.copy_(weights["state_dict"][root + block_names[2]]) |
|
self.attn.relative_position_index.copy_(weights["state_dict"][root + block_names[3]]) |
|
self.attn.qkv.weight.copy_(weights["state_dict"][root + block_names[4]]) |
|
self.attn.qkv.bias.copy_(weights["state_dict"][root + block_names[5]]) |
|
self.attn.proj.weight.copy_(weights["state_dict"][root + block_names[6]]) |
|
self.attn.proj.bias.copy_(weights["state_dict"][root + block_names[7]]) |
|
self.norm2.weight.copy_(weights["state_dict"][root + block_names[8]]) |
|
self.norm2.bias.copy_(weights["state_dict"][root + block_names[9]]) |
|
self.mlp.linear1.weight.copy_(weights["state_dict"][root + block_names[10]]) |
|
self.mlp.linear1.bias.copy_(weights["state_dict"][root + block_names[11]]) |
|
self.mlp.linear2.weight.copy_(weights["state_dict"][root + block_names[12]]) |
|
self.mlp.linear2.bias.copy_(weights["state_dict"][root + block_names[13]]) |
|
|
|
def forward(self, x, mask_matrix): |
|
shortcut = x |
|
if self.use_checkpoint: |
|
x = checkpoint.checkpoint(self.forward_part1, x, mask_matrix) |
|
else: |
|
x = self.forward_part1(x, mask_matrix) |
|
x = shortcut + self.drop_path(x) |
|
if self.use_checkpoint: |
|
x = x + checkpoint.checkpoint(self.forward_part2, x) |
|
else: |
|
x = x + self.forward_part2(x) |
|
return x |
|
|
|
|
|
class PatchMerging(nn.Module): |
|
""" |
|
Patch merging layer based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
""" |
|
|
|
def __init__( |
|
self, dim: int, norm_layer: Type[LayerNorm] = nn.LayerNorm, spatial_dims: int = 3 |
|
) -> None: |
|
""" |
|
Args: |
|
dim: number of feature channels. |
|
norm_layer: normalization layer. |
|
spatial_dims: number of spatial dims. |
|
""" |
|
|
|
super().__init__() |
|
self.dim = dim |
|
if spatial_dims == 3: |
|
self.reduction = nn.Linear(8 * dim, 2 * dim, bias=False) |
|
self.norm = norm_layer(8 * dim) |
|
elif spatial_dims == 2: |
|
self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False) |
|
self.norm = norm_layer(4 * dim) |
|
|
|
def forward(self, x): |
|
|
|
x_shape = x.size() |
|
if len(x_shape) == 5: |
|
b, d, h, w, c = x_shape |
|
pad_input = (h % 2 == 1) or (w % 2 == 1) or (d % 2 == 1) |
|
if pad_input: |
|
x = F.pad(x, (0, 0, 0, d % 2, 0, w % 2, 0, h % 2)) |
|
x0 = x[:, 0::2, 0::2, 0::2, :] |
|
x1 = x[:, 1::2, 0::2, 0::2, :] |
|
x2 = x[:, 0::2, 1::2, 0::2, :] |
|
x3 = x[:, 0::2, 0::2, 1::2, :] |
|
x4 = x[:, 1::2, 0::2, 1::2, :] |
|
x5 = x[:, 0::2, 1::2, 0::2, :] |
|
x6 = x[:, 0::2, 0::2, 1::2, :] |
|
x7 = x[:, 1::2, 1::2, 1::2, :] |
|
x = torch.cat([x0, x1, x2, x3, x4, x5, x6, x7], -1) |
|
|
|
elif len(x_shape) == 4: |
|
b, h, w, c = x_shape |
|
pad_input = (h % 2 == 1) or (w % 2 == 1) |
|
if pad_input: |
|
x = F.pad(x, (0, 0, 0, w % 2, 0, h % 2)) |
|
x0 = x[:, 0::2, 0::2, :] |
|
x1 = x[:, 1::2, 0::2, :] |
|
x2 = x[:, 0::2, 1::2, :] |
|
x3 = x[:, 1::2, 1::2, :] |
|
x = torch.cat([x0, x1, x2, x3], -1) |
|
|
|
x = self.norm(x) |
|
x = self.reduction(x) |
|
return x |
|
|
|
|
|
def compute_mask(dims, window_size, shift_size, device): |
|
"""Computing region masks based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
|
|
Args: |
|
dims: dimension values. |
|
window_size: local window size. |
|
shift_size: shift size. |
|
device: device. |
|
""" |
|
|
|
cnt = 0 |
|
|
|
if len(dims) == 3: |
|
d, h, w = dims |
|
img_mask = torch.zeros((1, d, h, w, 1), device=device) |
|
for d in slice(-window_size[0]), slice(-window_size[0], -shift_size[0]), slice(-shift_size[0], None): |
|
for h in slice(-window_size[1]), slice(-window_size[1], -shift_size[1]), slice(-shift_size[1], None): |
|
for w in slice(-window_size[2]), slice(-window_size[2], -shift_size[2]), slice(-shift_size[2], None): |
|
img_mask[:, d, h, w, :] = cnt |
|
cnt += 1 |
|
|
|
elif len(dims) == 2: |
|
h, w = dims |
|
img_mask = torch.zeros((1, h, w, 1), device=device) |
|
for h in slice(-window_size[0]), slice(-window_size[0], -shift_size[0]), slice(-shift_size[0], None): |
|
for w in slice(-window_size[1]), slice(-window_size[1], -shift_size[1]), slice(-shift_size[1], None): |
|
img_mask[:, h, w, :] = cnt |
|
cnt += 1 |
|
|
|
mask_windows = window_partition(img_mask, window_size) |
|
mask_windows = mask_windows.squeeze(-1) |
|
attn_mask = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2) |
|
attn_mask = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0)) |
|
|
|
return attn_mask |
|
|
|
|
|
class BasicLayer(nn.Module): |
|
""" |
|
Basic Swin Transformer layer in one stage based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
""" |
|
|
|
def __init__( |
|
self, |
|
dim: int, |
|
depth: int, |
|
num_heads: int, |
|
window_size: Sequence[int], |
|
drop_path: list, |
|
mlp_ratio: float = 4.0, |
|
qkv_bias: bool = False, |
|
drop: float = 0.0, |
|
attn_drop: float = 0.0, |
|
norm_layer: Type[LayerNorm] = nn.LayerNorm, |
|
downsample: isinstance = None, |
|
use_checkpoint: bool = False, |
|
) -> None: |
|
""" |
|
Args: |
|
dim: number of feature channels. |
|
depths: number of layers in each stage. |
|
num_heads: number of attention heads. |
|
window_size: local window size. |
|
drop_path: stochastic depth rate. |
|
mlp_ratio: ratio of mlp hidden dim to embedding dim. |
|
qkv_bias: add a learnable bias to query, key, value. |
|
drop: dropout rate. |
|
attn_drop: attention dropout rate. |
|
norm_layer: normalization layer. |
|
downsample: downsample layer at the end of the layer. |
|
use_checkpoint: use gradient checkpointing for reduced memory usage. |
|
""" |
|
|
|
super().__init__() |
|
self.window_size = window_size |
|
self.shift_size = tuple(i // 2 for i in window_size) |
|
self.no_shift = tuple(0 for i in window_size) |
|
self.depth = depth |
|
self.use_checkpoint = use_checkpoint |
|
self.blocks = nn.ModuleList( |
|
[ |
|
SwinTransformerBlock( |
|
dim=dim, |
|
num_heads=num_heads, |
|
window_size=self.window_size, |
|
shift_size=self.no_shift if (i % 2 == 0) else self.shift_size, |
|
mlp_ratio=mlp_ratio, |
|
qkv_bias=qkv_bias, |
|
drop=drop, |
|
attn_drop=attn_drop, |
|
drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path, |
|
norm_layer=norm_layer, |
|
use_checkpoint=use_checkpoint, |
|
) |
|
for i in range(depth) |
|
] |
|
) |
|
self.downsample = downsample |
|
if self.downsample is not None: |
|
self.downsample = downsample(dim=dim, norm_layer=norm_layer, spatial_dims=len(self.window_size)) |
|
|
|
def forward(self, x): |
|
x_shape = x.size() |
|
if len(x_shape) == 5: |
|
b, c, d, h, w = x_shape |
|
window_size, shift_size = get_window_size((d, h, w), self.window_size, self.shift_size) |
|
x = rearrange(x, "b c d h w -> b d h w c") |
|
dp = int(np.ceil(d / window_size[0])) * window_size[0] |
|
hp = int(np.ceil(h / window_size[1])) * window_size[1] |
|
wp = int(np.ceil(w / window_size[2])) * window_size[2] |
|
attn_mask = compute_mask([dp, hp, wp], window_size, shift_size, x.device) |
|
for blk in self.blocks: |
|
x = blk(x, attn_mask) |
|
x = x.view(b, d, h, w, -1) |
|
if self.downsample is not None: |
|
x = self.downsample(x) |
|
x = rearrange(x, "b d h w c -> b c d h w") |
|
|
|
elif len(x_shape) == 4: |
|
b, c, h, w = x_shape |
|
window_size, shift_size = get_window_size((h, w), self.window_size, self.shift_size) |
|
x = rearrange(x, "b c h w -> b h w c") |
|
hp = int(np.ceil(h / window_size[0])) * window_size[0] |
|
wp = int(np.ceil(w / window_size[1])) * window_size[1] |
|
attn_mask = compute_mask([hp, wp], window_size, shift_size, x.device) |
|
for blk in self.blocks: |
|
x = blk(x, attn_mask) |
|
x = x.view(b, h, w, -1) |
|
if self.downsample is not None: |
|
x = self.downsample(x) |
|
x = rearrange(x, "b h w c -> b c h w") |
|
return x |
|
|
|
|
|
class SwinTransformer(nn.Module): |
|
""" |
|
Swin Transformer based on: "Liu et al., |
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows |
|
<https://arxiv.org/abs/2103.14030>" |
|
https://github.com/microsoft/Swin-Transformer |
|
""" |
|
|
|
def __init__( |
|
self, |
|
in_chans: int, |
|
embed_dim: int, |
|
text_dim: int, |
|
window_size: Sequence[int], |
|
patch_size: Sequence[int], |
|
depths: Sequence[int], |
|
num_heads: Sequence[int], |
|
mlp_ratio: float = 4.0, |
|
qkv_bias: bool = True, |
|
drop_rate: float = 0.0, |
|
attn_drop_rate: float = 0.0, |
|
drop_path_rate: float = 0.0, |
|
norm_layer: Type[LayerNorm] = nn.LayerNorm, |
|
patch_norm: bool = False, |
|
use_checkpoint: bool = False, |
|
spatial_dims: int = 3, |
|
) -> None: |
|
""" |
|
Args: |
|
in_chans: dimension of input channels. |
|
embed_dim: number of linear projection output channels. |
|
window_size: local window size. |
|
patch_size: patch size. |
|
depths: number of layers in each stage. |
|
num_heads: number of attention heads. |
|
mlp_ratio: ratio of mlp hidden dim to embedding dim. |
|
qkv_bias: add a learnable bias to query, key, value. |
|
drop_rate: dropout rate. |
|
attn_drop_rate: attention dropout rate. |
|
drop_path_rate: stochastic depth rate. |
|
norm_layer: normalization layer. |
|
patch_norm: add normalization after patch embedding. |
|
use_checkpoint: use gradient checkpointing for reduced memory usage. |
|
spatial_dims: spatial dimension. |
|
""" |
|
|
|
super().__init__() |
|
self.num_layers = len(depths) |
|
self.embed_dim = embed_dim |
|
self.patch_norm = patch_norm |
|
self.window_size = window_size |
|
self.patch_size = patch_size |
|
self.patch_embed = PatchEmbed( |
|
patch_size=self.patch_size, |
|
in_chans=in_chans, |
|
embed_dim=embed_dim, |
|
norm_layer=norm_layer if self.patch_norm else None, |
|
spatial_dims=spatial_dims, |
|
) |
|
self.pos_drop = nn.Dropout(p=drop_rate) |
|
dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] |
|
self.layers1 = nn.ModuleList() |
|
self.layers2 = nn.ModuleList() |
|
self.layers3 = nn.ModuleList() |
|
self.layers4 = nn.ModuleList() |
|
for i_layer in range(self.num_layers): |
|
layer = BasicLayer( |
|
dim=int(embed_dim * 2**i_layer), |
|
depth=depths[i_layer], |
|
num_heads=num_heads[i_layer], |
|
window_size=self.window_size, |
|
drop_path=dpr[sum(depths[:i_layer]) : sum(depths[: i_layer + 1])], |
|
mlp_ratio=mlp_ratio, |
|
qkv_bias=qkv_bias, |
|
drop=drop_rate, |
|
attn_drop=attn_drop_rate, |
|
norm_layer=norm_layer, |
|
downsample=PatchMerging, |
|
use_checkpoint=use_checkpoint, |
|
) |
|
if i_layer == 0: |
|
self.layers1.append(layer) |
|
elif i_layer == 1: |
|
self.layers2.append(layer) |
|
elif i_layer == 2: |
|
self.layers3.append(layer) |
|
elif i_layer == 3: |
|
self.layers4.append(layer) |
|
self.num_features = int(embed_dim * 2 ** (self.num_layers - 1)) |
|
self.mlp_text = nn.Sequential( |
|
nn.Conv1d(text_dim, 1024, kernel_size=1), |
|
nn.ReLU(), |
|
nn.Conv1d(1024, 768, kernel_size=1), |
|
) |
|
|
|
self.fw_mlp = nn.Sequential( |
|
nn.Conv3d(768, 768, kernel_size=1), |
|
nn.ReLU()) |
|
|
|
self.mlpk = nn.Conv1d(768, 768, kernel_size=1) |
|
self.mlpv = nn.Conv1d(768, 768, kernel_size=1) |
|
|
|
self.ly_norm = nn.LayerNorm(normalized_shape=(4,4,4)) |
|
|
|
self.mlp_text_q = nn.Conv1d(768, 768, kernel_size=1) |
|
self.mlp_image_k = nn.Conv1d(768, 768, kernel_size=1) |
|
self.mlp_image_v = nn.Conv1d(768, 768, kernel_size=1) |
|
self.mlp_image_q = nn.Conv1d(768, 768, kernel_size=1) |
|
|
|
|
|
def proj_out(self, x, normalize=False): |
|
if normalize: |
|
x_shape = x.size() |
|
if len(x_shape) == 5: |
|
n, ch, d, h, w = x_shape |
|
x = rearrange(x, "n c d h w -> n d h w c") |
|
x = F.layer_norm(x, [ch]) |
|
x = rearrange(x, "n d h w c -> n c d h w") |
|
elif len(x_shape) == 4: |
|
n, ch, h, w = x_shape |
|
x = rearrange(x, "n c h w -> n h w c") |
|
x = F.layer_norm(x, [ch]) |
|
x = rearrange(x, "n h w c -> n c h w") |
|
return x |
|
|
|
def sequential_cross_attention(self, image_features, text_features): |
|
|
|
""" |
|
Cross attention between image and text features. |
|
Args: |
|
image_features: Tensor of shape (B, C, H, W, D) |
|
text_features: Tensor of shape (B, T_dim, T_len) |
|
Returns: |
|
Processed image features with the same shape as input (B, C, H, W, D) |
|
""" |
|
B, C, H, W, D = image_features.shape |
|
_, T_dim, T_len = text_features.shape |
|
|
|
|
|
|
|
text_features = self.mlp_text(text_features.permute(0,2,1).contiguous()) |
|
text_q = self.mlp_text_q(text_features).permute(0,2,1).contiguous() |
|
|
|
|
|
image_features_flat = image_features.view(B, C, -1).contiguous() |
|
image_k = self.mlp_image_k(image_features_flat).permute(0,2,1).contiguous() |
|
image_v = self.mlp_image_v(image_features_flat).permute(0,2,1).contiguous() |
|
|
|
|
|
attn_scores_t2i = torch.matmul(text_q, image_k.transpose(-2, -1)) / math.sqrt( |
|
text_q.size(-1)) |
|
attn_weights_t2i = F.softmax(attn_scores_t2i, dim=-1) |
|
|
|
|
|
attended_image_features = torch.matmul(attn_weights_t2i, image_v).permute(0,2,1).contiguous() |
|
|
|
|
|
|
|
image_q = self.mlp_image_q(image_features_flat).permute(0,2,1).contiguous() |
|
|
|
|
|
attended_image_k = self.mlpk(attended_image_features).permute(0,2,1).contiguous() |
|
attended_image_v = self.mlpv(attended_image_features).permute(0,2,1).contiguous() |
|
|
|
|
|
attn_scores_i2t = torch.matmul(image_q, attended_image_k.transpose(-2, -1)) / math.sqrt( |
|
image_q.size(-1)) |
|
attn_weights_i2t = F.softmax(attn_scores_i2t, dim=-1) |
|
|
|
|
|
attn_output_image = torch.matmul(attn_weights_i2t, attended_image_v) |
|
|
|
|
|
attn_output_image = attn_output_image.permute(0, 2, 1).contiguous() |
|
attn_output_image = attn_output_image.view(B, C, H, W, D) |
|
|
|
|
|
processed_image_features = self.ly_norm(attn_output_image) |
|
processed_image_features = self.fw_mlp(processed_image_features.float()) |
|
processed_image_features = self.ly_norm(processed_image_features) |
|
|
|
return processed_image_features |
|
|
|
|
|
def forward(self, x, text, normalize=True): |
|
x0 = self.patch_embed(x) |
|
x0 = self.pos_drop(x0) |
|
x0_out = self.proj_out(x0, normalize) |
|
x1 = self.layers1[0](x0.contiguous()) |
|
x1_out = self.proj_out(x1, normalize) |
|
x2 = self.layers2[0](x1.contiguous()) |
|
x2_out = self.proj_out(x2, normalize) |
|
x3 = self.layers3[0](x2.contiguous()) |
|
x3_out = self.proj_out(x3, normalize) |
|
x4 = self.layers4[0](x3.contiguous()) |
|
|
|
x4 = self.sequential_cross_attention(x4,text) |
|
x4_out = self.proj_out(x4, normalize) |
|
return [x0_out, x1_out, x2_out, x3_out, x4_out] |
|
|
|
|
|
if __name__ == "__main__": |
|
model = TextSwinUNETR( |
|
img_size=(128,128,128), |
|
in_channels=4, |
|
out_channels=3, |
|
feature_size=48, |
|
text_dim=768, |
|
use_checkpoint=False, |
|
).cuda() |
|
|
|
input = torch.randn(1,4,128,128,128).cuda() |
|
text = torch.randn(1,128,768).cuda() |
|
output = model(input,text) |
|
print(output[0].shape) |
|
|