BOAT-Swin Transformer for Classification - Inference with ImageNet1K(mini version) dataset
on Projects
BOAT-Swin Transformer for Classification - Inference with ImageNet1K(mini version) dataset
1. dataset
2. Edit the swin_transformer.py
# --------------------------------------------------------
# Swin Transformer
# Copyright (c) 2021 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Ze Liu
# --------------------------------------------------------
import torch
import torch.nn as nn
import torch.utils.checkpoint as checkpoint
from timm.models.layers import DropPath, to_2tuple, trunc_normal_
import math
from einops import rearrange, repeat
class Mlp(nn.Module):
def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
super().__init__()
out_features = out_features or in_features
hidden_features = hidden_features or in_features
self.fc1 = nn.Linear(in_features, hidden_features)
self.act = act_layer()
self.fc2 = nn.Linear(hidden_features, out_features)
self.drop = nn.Dropout(drop)
def forward(self, x):
x = self.fc1(x)
x = self.act(x)
x = self.drop(x)
x = self.fc2(x)
x = self.drop(x)
return x
def window_partition(x, window_size):
"""
Args:
x: (B, H, W, C)
window_size (int): window size
Returns:
windows: (num_windows*B, window_size, window_size, C)
"""
B, H, W, C = x.shape
x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
return windows
def window_reverse(windows, window_size, H, W):
"""
Args:
windows: (num_windows*B, window_size, window_size, C)
window_size (int): Window size
H (int): Height of image
W (int): Width of image
Returns:
x: (B, H, W, C)
"""
B = int(windows.shape[0] / (H * W / window_size / window_size))
x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
return x
class ContentAttention(nn.Module):
r""" Window based multi-head self attention (W-MSA) module with relative position bias.
It supports both of shifted and non-shifted window.
Args:
dim (int): Number of input channels.
window_size (tuple[int]): The height and width of the window.
num_heads (int): Number of attention heads.
qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
proj_drop (float, optional): Dropout ratio of output. Default: 0.0
"""
def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0., kmeans = False):
super().__init__()
self.dim = dim
self.window_size = window_size # Wh, Ww
self.ws = window_size
self.num_heads = num_heads
head_dim = dim // num_heads
self.scale = qk_scale or head_dim ** -0.5
self.kmeans = kmeans
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)
self.softmax = nn.Softmax(dim=-1)
self.get_v = nn.Conv2d(dim, dim, kernel_size=3, stride=1, padding=1,groups=dim)
def forward(self, x, mask=None):
"""
Args:
x: input features with shape of (num_windows*B, N, C)
mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
"""
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)# 3, B_, self.num_heads,N,D
if True:
q_pre = qkv[0].reshape(B_*self.num_heads,N, C // self.num_heads).permute(0,2,1)#qkv_pre[:,0].reshape(b*self.num_heads,qkvhd//3//self.num_heads,hh*ww)
ntimes = int(math.log(N//49,2))
q_idx_last = torch.arange(N).cuda().unsqueeze(0).expand(B_*self.num_heads,N)
for i in range(ntimes):
bh,d,n = q_pre.shape
q_pre_new = q_pre.reshape(bh,d,2,n//2)
q_avg = q_pre_new.mean(dim=-1)#.reshape(b*self.num_heads,qkvhd//3//self.num_heads,)
q_avg = torch.nn.functional.normalize(q_avg,dim=-2)
iters = 2
for i in range(iters):
q_scores = torch.nn.functional.normalize(q_pre.permute(0,2,1),dim=-1).bmm(q_avg)
soft_assign = torch.nn.functional.softmax(q_scores*100, dim=-1).detach()
q_avg = q_pre.bmm(soft_assign)
q_avg = torch.nn.functional.normalize(q_avg,dim=-2)
q_scores = torch.nn.functional.normalize(q_pre.permute(0,2,1),dim=-1).bmm(q_avg).reshape(bh,n,2)#.unsqueeze(2)
q_idx = (q_scores[:,:,0]+1)/(q_scores[:,:,1]+1)
_,q_idx = torch.sort(q_idx,dim=-1)
q_idx_last = q_idx_last.gather(dim=-1,index=q_idx).reshape(bh*2,n//2)
q_idx = q_idx.unsqueeze(1).expand(q_pre.size())
q_pre = q_pre.gather(dim=-1,index=q_idx).reshape(bh,d,2,n//2).permute(0,2,1,3).reshape(bh*2,d,n//2)
q_idx = q_idx_last.view(B_,self.num_heads,N)
_,q_idx_rev = torch.sort(q_idx,dim=-1)
q_idx = q_idx.unsqueeze(0).unsqueeze(4).expand(qkv.size())
qkv_pre = qkv.gather(dim=-2,index=q_idx)
q, k, v = rearrange(qkv_pre, 'qkv b h (nw ws) c -> qkv (b nw) h ws c', ws=49)
k = k.view(B_*((N//49))//2,2,self.num_heads,49,-1)
k_over1 = k[:,1,:,:20].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
k_over2 = k[:,0,:,29:].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
k_over = torch.cat([k_over1,k_over2],1)
k = torch.cat([k,k_over],3).contiguous().view(B_*((N//49)),self.num_heads,49+20,-1)
v = v.view(B_*((N//49))//2,2,self.num_heads,49,-1)
v_over1 = v[:,1,:,:20].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
v_over2 = v[:,0,:,29:].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
v_over = torch.cat([v_over1,v_over2],1)
v = torch.cat([v,v_over],3).contiguous().view(B_*((N//49)),self.num_heads,49+20,-1)
attn = (q @ k.transpose(-2, -1))*self.scale
attn = self.softmax(attn)
attn = self.attn_drop(attn)
out = attn @ v
if True:
out = rearrange(out, '(b nw) h ws d -> b (h d) nw ws', h=self.num_heads, b=B_)
v = rearrange(v[:,:,:49,:], '(b nw) h ws d -> b h d (nw ws)', h=self.num_heads, b=B_)
W = int(math.sqrt(N))
out = out.reshape(B_,self.num_heads,C//self.num_heads,-1)
q_idx_rev = q_idx_rev.unsqueeze(2).expand(out.size())
x = out.gather(dim=-1,index=q_idx_rev).reshape(B_,C,N).permute(0,2,1)
v = v.gather(dim=-1,index=q_idx_rev).reshape(B_,C,W,W)
v = self.get_v(v)
v = v.reshape(B_,C,N).permute(0,2,1)
x = x + v
x = self.proj(x)
x = self.proj_drop(x)
return x
class WindowAttention(nn.Module):
r""" Window based multi-head self attention (W-MSA) module with relative position bias.
It supports both of shifted and non-shifted window.
Args:
dim (int): Number of input channels.
window_size (tuple[int]): The height and width of the window.
num_heads (int): Number of attention heads.
qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
proj_drop (float, optional): Dropout ratio of output. Default: 0.0
"""
def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0., kmeans = False):
super().__init__()
self.dim = dim
self.window_size = window_size # Wh, Ww
self.ws = window_size
self.num_heads = num_heads
head_dim = dim // num_heads
self.scale = qk_scale or head_dim ** -0.5
self.kmeans = kmeans
# define a parameter table of relative position bias
if True:
self.relative_position_bias_table = nn.Parameter(
torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)) # 2*Wh-1 * 2*Ww-1, nH
# get pair-wise relative position index for each token inside the window
coords_h = torch.arange(self.window_size[0])
coords_w = torch.arange(self.window_size[1])
coords = torch.stack(torch.meshgrid([coords_h, coords_w])) # 2, Wh, Ww
coords_flatten = torch.flatten(coords, 1) # 2, Wh*Ww
relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] # 2, Wh*Ww, Wh*Ww
relative_coords = relative_coords.permute(1, 2, 0).contiguous() # Wh*Ww, Wh*Ww, 2
relative_coords[:, :, 0] += self.window_size[0] - 1 # shift to start from 0
relative_coords[:, :, 1] += self.window_size[1] - 1
relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
relative_position_index = relative_coords.sum(-1) # Wh*Ww, Wh*Ww
self.register_buffer("relative_position_index", relative_position_index)
trunc_normal_(self.relative_position_bias_table, std=.02)
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=.02)
self.softmax = nn.Softmax(dim=-1)
def forward(self, x, mask=None):
"""
Args:
x: input features with shape of (num_windows*B, N, C)
mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
"""
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] # make torchscript happy (cannot use tensor as tuple)
q = q * self.scale
attn = (q @ k.transpose(-2, -1))
relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1) # Wh*Ww,Wh*Ww,nH
relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous() # nH, Wh*Ww, Wh*Ww
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
def extra_repr(self) -> str:
return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'
def flops(self, N):
# calculate flops for 1 window with token length of N
flops = 0
# qkv = self.qkv(x)
flops += N * self.dim * 3 * self.dim
# attn = (q @ k.transpose(-2, -1))
flops += self.num_heads * N * (self.dim // self.num_heads) * N
# x = (attn @ v)
flops += self.num_heads * N * N * (self.dim // self.num_heads)
# x = self.proj(x)
flops += N * self.dim * self.dim
return flops
class SwinTransformerBlock(nn.Module):
r""" Swin Transformer Block.
Args:
dim (int): Number of input channels.
input_resolution (tuple[int]): Input resulotion.
num_heads (int): Number of attention heads.
window_size (int): Window size.
shift_size (int): Shift size for SW-MSA.
mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
drop (float, optional): Dropout rate. Default: 0.0
attn_drop (float, optional): Attention dropout rate. Default: 0.0
drop_path (float, optional): Stochastic depth rate. Default: 0.0
act_layer (nn.Module, optional): Activation layer. Default: nn.GELU
norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
"""
def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
act_layer=nn.GELU, norm_layer=nn.LayerNorm):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.num_heads = num_heads
self.window_size = window_size
self.shift_size = shift_size
self.mlp_ratio = mlp_ratio
if min(self.input_resolution) <= self.window_size:
# if window size is larger than input resolution, we don't partition windows
self.shift_size = 0
self.window_size = min(self.input_resolution)
assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
self.norm1 = norm_layer(dim)
self.attn = WindowAttention(
dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, kmeans=shift_size)
if self.shift_size > 0:
self.attnC = ContentAttention(
dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, kmeans=shift_size)
self.norm4 = norm_layer(dim);
self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
self.norm2 = norm_layer(dim)
mlp_hidden_dim = int(dim * mlp_ratio)
self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
#self.local = nn.Conv2d(dim, dim, window_size, 1, window_size//2, groups=dim, bias=qkv_bias)
#self.norm3 = norm_layer(dim)
#self.norm4 = norm_layer(dim)
if self.shift_size > 0:
# calculate attention mask for SW-MSA
H, W = self.input_resolution
img_mask = torch.zeros((1, H, W, 1)) # 1 H W 1
h_slices = (slice(0, -self.window_size),
slice(-self.window_size, -self.shift_size),
slice(-self.shift_size, None))
w_slices = (slice(0, -self.window_size),
slice(-self.window_size, -self.shift_size),
slice(-self.shift_size, None))
cnt = 0
for h in h_slices:
for w in w_slices:
img_mask[:, h, w, :] = cnt
cnt += 1
mask_windows = window_partition(img_mask, self.window_size) # nW, window_size, window_size, 1
mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
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))
else:
attn_mask = None
self.register_buffer("attn_mask", attn_mask)
def forward(self, x):
H, W = self.input_resolution
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
shortcut = x
x = self.norm1(x)
x = x.view(B, H, W, C)
# cyclic shift
if self.shift_size > 0:
shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
#x = self.attn(x.view(B,H*W,C))
else:
shifted_x = x
if True:
# partition windows
x_windows = window_partition(shifted_x, self.window_size) # nW*B, window_size, window_size, C
x_windows = x_windows.view(-1, self.window_size * self.window_size, C) # nW*B, window_size*window_size, C
# W-MSA/SW-MSA
attn_windows = self.attn(x_windows, mask=self.attn_mask) # nW*B, window_size*window_size, C
# merge windows
attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
shifted_x = window_reverse(attn_windows, self.window_size, H, W) # B H' W' C
# reverse cyclic shift
if self.shift_size > 0:
x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
else:
x = shifted_x
x = x.view(B, H * W, C)
# FFN
x = shortcut + self.drop_path(x)
if self.shift_size > 0:
x = x + self.attnC(self.norm4(x))
b,n,c = x.shape
#w = int(math.sqrt(n))
#x = self.norm3(x)
#x = x.view(b,w,w,c).permute(0,3,1,2)
#x = x + self.local(x)
#x = x.permute(0,2,3,1).reshape(b,n,c)
x = x + self.drop_path(self.mlp(self.norm2(x)))
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"
def flops(self):
flops = 0
H, W = self.input_resolution
# norm1
flops += self.dim * H * W
# W-MSA/SW-MSA
nW = H * W / self.window_size / self.window_size
flops += nW * self.attn.flops(self.window_size * self.window_size)
# mlp
flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
# norm2
flops += self.dim * H * W
return flops
class PatchMerging(nn.Module):
r""" Patch Merging Layer.
Args:
input_resolution (tuple[int]): Resolution of input feature.
dim (int): Number of input channels.
norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
"""
def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
super().__init__()
self.input_resolution = input_resolution
self.dim = dim
self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
self.norm = norm_layer(4 * dim)
def forward(self, x):
"""
x: B, H*W, C
"""
H, W = self.input_resolution
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."
x = x.view(B, H, W, C)
x0 = x[:, 0::2, 0::2, :] # B H/2 W/2 C
x1 = x[:, 1::2, 0::2, :] # B H/2 W/2 C
x2 = x[:, 0::2, 1::2, :] # B H/2 W/2 C
x3 = x[:, 1::2, 1::2, :] # B H/2 W/2 C
x = torch.cat([x0, x1, x2, x3], -1) # B H/2 W/2 4*C
x = x.view(B, -1, 4 * C) # B H/2*W/2 4*C
x = self.norm(x)
x = self.reduction(x)
return x
def extra_repr(self) -> str:
return f"input_resolution={self.input_resolution}, dim={self.dim}"
def flops(self):
H, W = self.input_resolution
flops = H * W * self.dim
flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
return flops
class BasicLayer(nn.Module):
""" A basic Swin Transformer layer for one stage.
Args:
dim (int): Number of input channels.
input_resolution (tuple[int]): Input resolution.
depth (int): Number of blocks.
num_heads (int): Number of attention heads.
window_size (int): Local window size.
mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
drop (float, optional): Dropout rate. Default: 0.0
attn_drop (float, optional): Attention dropout rate. Default: 0.0
drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
"""
def __init__(self, dim, input_resolution, depth, num_heads, window_size,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.depth = depth
self.use_checkpoint = use_checkpoint
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0 if (i % 2 == 0) else window_size // 2,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer)
for i in range(depth)])
# patch merging layer
if downsample is not None:
self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def forward(self, x):
for blk in self.blocks:
if self.use_checkpoint:
x = checkpoint.checkpoint(blk, x)
else:
x = blk(x)
if self.downsample is not None:
x = self.downsample(x)
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
def flops(self):
flops = 0
for blk in self.blocks:
flops += blk.flops()
if self.downsample is not None:
flops += self.downsample.flops()
return flops
class PatchEmbed(nn.Module):
r""" Image to Patch Embedding
Args:
img_size (int): Image size. Default: 224.
patch_size (int): Patch token size. Default: 4.
in_chans (int): Number of input image channels. Default: 3.
embed_dim (int): Number of linear projection output channels. Default: 96.
norm_layer (nn.Module, optional): Normalization layer. Default: None
"""
def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
super().__init__()
img_size = to_2tuple(img_size)
patch_size = to_2tuple(patch_size)
patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
self.img_size = img_size
self.patch_size = patch_size
self.patches_resolution = patches_resolution
self.num_patches = patches_resolution[0] * patches_resolution[1]
self.in_chans = in_chans
self.embed_dim = embed_dim
self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
if norm_layer is not None:
self.norm = norm_layer(embed_dim)
else:
self.norm = None
def forward(self, x):
B, C, H, W = x.shape
# FIXME look at relaxing size constraints
assert H == self.img_size[0] and W == self.img_size[1], \
f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
x = self.proj(x).flatten(2).transpose(1, 2) # B Ph*Pw C
if self.norm is not None:
x = self.norm(x)
return x
def flops(self):
Ho, Wo = self.patches_resolution
flops = Ho * Wo * self.embed_dim * self.in_chans * (self.patch_size[0] * self.patch_size[1])
if self.norm is not None:
flops += Ho * Wo * self.embed_dim
return flops
class SwinTransformer(nn.Module):
r""" Swin Transformer
A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` -
https://arxiv.org/pdf/2103.14030
Args:
img_size (int | tuple(int)): Input image size. Default 224
patch_size (int | tuple(int)): Patch size. Default: 4
in_chans (int): Number of input image channels. Default: 3
num_classes (int): Number of classes for classification head. Default: 1000
embed_dim (int): Patch embedding dimension. Default: 96
depths (tuple(int)): Depth of each Swin Transformer layer.
num_heads (tuple(int)): Number of attention heads in different layers.
window_size (int): Window size. Default: 7
mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4
qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: None
drop_rate (float): Dropout rate. Default: 0
attn_drop_rate (float): Attention dropout rate. Default: 0
drop_path_rate (float): Stochastic depth rate. Default: 0.1
norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.
ape (bool): If True, add absolute position embedding to the patch embedding. Default: False
patch_norm (bool): If True, add normalization after patch embedding. Default: True
use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
"""
def __init__(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000,
embed_dim=96, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24],
window_size=7, mlp_ratio=4., qkv_bias=True, qk_scale=None,
drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
use_checkpoint=False, **kwargs):
super().__init__()
self.num_classes = num_classes
self.num_layers = len(depths)
self.embed_dim = embed_dim
self.ape = ape
self.patch_norm = patch_norm
self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
self.mlp_ratio = mlp_ratio
# split image into non-overlapping patches
self.patch_embed = PatchEmbed(
img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
norm_layer=norm_layer if self.patch_norm else None)
num_patches = self.patch_embed.num_patches
patches_resolution = self.patch_embed.patches_resolution
self.patches_resolution = patches_resolution
# absolute position embedding
if self.ape:
self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
trunc_normal_(self.absolute_pos_embed, std=.02)
self.pos_drop = nn.Dropout(p=drop_rate)
# stochastic depth
dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] # stochastic depth decay rule
# build layers
self.layers = nn.ModuleList()
for i_layer in range(self.num_layers):
layer = BasicLayer(dim=int(embed_dim * 2 ** i_layer),
input_resolution=(patches_resolution[0] // (2 ** i_layer),
patches_resolution[1] // (2 ** i_layer)),
depth=depths[i_layer],
num_heads=num_heads[i_layer],
window_size=window_size,
mlp_ratio=self.mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop_rate, attn_drop=attn_drop_rate,
drop_path=dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],
norm_layer=norm_layer,
downsample=PatchMerging if (i_layer < self.num_layers - 1) else None,
use_checkpoint=use_checkpoint)
self.layers.append(layer)
self.norm = norm_layer(self.num_features)
self.avgpool = nn.AdaptiveAvgPool1d(1)
self.head = nn.Linear(self.num_features, num_classes) if num_classes > 0 else nn.Identity()
self.apply(self._init_weights)
def _init_weights(self, m):
if isinstance(m, nn.Linear):
trunc_normal_(m.weight, std=.02)
if isinstance(m, nn.Linear) and m.bias is not None:
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.LayerNorm):
nn.init.constant_(m.bias, 0)
nn.init.constant_(m.weight, 1.0)
@torch.jit.ignore
def no_weight_decay(self):
return {'absolute_pos_embed'}
@torch.jit.ignore
def no_weight_decay_keywords(self):
return {'relative_position_bias_table'}
def forward_features(self, x):
x = self.patch_embed(x)
if self.ape:
x = x + self.absolute_pos_embed
x = self.pos_drop(x)
for layer in self.layers:
x = layer(x)
x = self.norm(x) # B L C
x = self.avgpool(x.transpose(1, 2)) # B C 1
x = torch.flatten(x, 1)
return x
def forward(self, x):
x = self.forward_features(x)
x = self.head(x)
return x
def flops(self):
flops = 0
flops += self.patch_embed.flops()
for i, layer in enumerate(self.layers):
flops += layer.flops()
flops += self.num_features * self.patches_resolution[0] * self.patches_resolution[1] // (2 ** self.num_layers)
flops += self.num_features * self.num_classes
return flops
3. Edit the siwn_tiny_patch4_window7_224.yaml
MODEL:
TYPE: swin
NAME: swin_tiny_patch4_window7_224
DROP_PATH_RATE: 0.2
SWIN:
EMBED_DIM: 96
DEPTHS: [ 2, 2, 6, 2 ]
NUM_HEADS: [ 3, 6, 12, 24 ]
WINDOW_SIZE: 7
4. Training
python -m torch.distributed.launch --nproc_per_node 4 --master_port 12345 main.py --cfg configs/swin/swin_tiny_patch4_window7_224.yaml --data-path imagenet --batch-size 128
5. Evaluation
python -m torch.distributed.launch --nproc_per_node 1 --master_port 12345 main.py --eval --cfg configs/swin/swin_tiny_patch4_window7_224.yaml --resume checkpoints/BOATSwinTransformer-tiny.pth --data-path imagenet
6. ver2 Inference
# --------------------------------------------------------
# Swin Transformer
# Copyright (c) 2021 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Ze Liu
# --------------------------------------------------------
import torch
import torch.nn as nn
import torch.utils.checkpoint as checkpoint
from timm.models.layers import DropPath, to_2tuple, trunc_normal_
import math
from einops import rearrange, repeat
class Mlp(nn.Module):
def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
super().__init__()
out_features = out_features or in_features
hidden_features = hidden_features or in_features
self.fc1 = nn.Linear(in_features, hidden_features)
self.act = act_layer()
self.fc2 = nn.Linear(hidden_features, out_features)
self.drop = nn.Dropout(drop)
def forward(self, x):
x = self.fc1(x)
x = self.act(x)
x = self.drop(x)
x = self.fc2(x)
x = self.drop(x)
return x
def window_partition(x, window_size):
"""
Args:
x: (B, H, W, C)
window_size (int): window size
Returns:
windows: (num_windows*B, window_size, window_size, C)
"""
B, H, W, C = x.shape
x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
return windows
def window_reverse(windows, window_size, H, W):
"""
Args:
windows: (num_windows*B, window_size, window_size, C)
window_size (int): Window size
H (int): Height of image
W (int): Width of image
Returns:
x: (B, H, W, C)
"""
B = int(windows.shape[0] / (H * W / window_size / window_size))
x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
return x
class ContentAttention(nn.Module):
r""" Window based multi-head self attention (W-MSA) module with relative position bias.
It supports both of shifted and non-shifted window.
Args:
dim (int): Number of input channels.
window_size (tuple[int]): The height and width of the window.
num_heads (int): Number of attention heads.
qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
proj_drop (float, optional): Dropout ratio of output. Default: 0.0
"""
def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0., kmeans = False):
super().__init__()
self.dim = dim
self.window_size = window_size # Wh, Ww
self.ws = window_size
self.num_heads = num_heads
head_dim = dim // num_heads
self.scale = qk_scale or head_dim ** -0.5
self.kmeans = kmeans
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)
self.softmax = nn.Softmax(dim=-1)
self.get_v = nn.Conv2d(dim, dim, kernel_size=3, stride=1, padding=1,groups=dim)
def forward(self, x, mask=None):
"""
Args:
x: input features with shape of (num_windows*B, N, C)
mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
"""
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)# 3, B_, self.num_heads,N,D
if True:
q_pre = qkv[0].reshape(B_*self.num_heads,N, C // self.num_heads).permute(0,2,1)#qkv_pre[:,0].reshape(b*self.num_heads,qkvhd//3//self.num_heads,hh*ww)
ntimes = int(math.log(N//49,2))
q_idx_last = torch.arange(N).cuda().unsqueeze(0).expand(B_*self.num_heads,N)
for i in range(ntimes):
bh,d,n = q_pre.shape
q_pre_new = q_pre.reshape(bh,d,2,n//2)
q_avg = q_pre_new.mean(dim=-1)#.reshape(b*self.num_heads,qkvhd//3//self.num_heads,)
q_avg = torch.nn.functional.normalize(q_avg,dim=-2)
iters = 2
for i in range(iters):
q_scores = torch.nn.functional.normalize(q_pre.permute(0,2,1),dim=-1).bmm(q_avg)
soft_assign = torch.nn.functional.softmax(q_scores*100, dim=-1).detach()
q_avg = q_pre.bmm(soft_assign)
q_avg = torch.nn.functional.normalize(q_avg,dim=-2)
q_scores = torch.nn.functional.normalize(q_pre.permute(0,2,1),dim=-1).bmm(q_avg).reshape(bh,n,2)#.unsqueeze(2)
q_idx = (q_scores[:,:,0]+1)/(q_scores[:,:,1]+1)
_,q_idx = torch.sort(q_idx,dim=-1)
q_idx_last = q_idx_last.gather(dim=-1,index=q_idx).reshape(bh*2,n//2)
q_idx = q_idx.unsqueeze(1).expand(q_pre.size())
q_pre = q_pre.gather(dim=-1,index=q_idx).reshape(bh,d,2,n//2).permute(0,2,1,3).reshape(bh*2,d,n//2)
q_idx = q_idx_last.view(B_,self.num_heads,N)
_,q_idx_rev = torch.sort(q_idx,dim=-1)
q_idx = q_idx.unsqueeze(0).unsqueeze(4).expand(qkv.size())
qkv_pre = qkv.gather(dim=-2,index=q_idx)
q, k, v = rearrange(qkv_pre, 'qkv b h (nw ws) c -> qkv (b nw) h ws c', ws=49)
k = k.view(B_*((N//49))//2,2,self.num_heads,49,-1)
k_over1 = k[:,1,:,:20].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
k_over2 = k[:,0,:,29:].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
k_over = torch.cat([k_over1,k_over2],1)
k = torch.cat([k,k_over],3).contiguous().view(B_*((N//49)),self.num_heads,49+20,-1)
v = v.view(B_*((N//49))//2,2,self.num_heads,49,-1)
v_over1 = v[:,1,:,:20].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
v_over2 = v[:,0,:,29:].unsqueeze(1)#.expand(-1,2,-1,-1,-1)
v_over = torch.cat([v_over1,v_over2],1)
v = torch.cat([v,v_over],3).contiguous().view(B_*((N//49)),self.num_heads,49+20,-1)
attn = (q @ k.transpose(-2, -1))*self.scale
attn = self.softmax(attn)
attn = self.attn_drop(attn)
out = attn @ v
if True:
out = rearrange(out, '(b nw) h ws d -> b (h d) nw ws', h=self.num_heads, b=B_)
v = rearrange(v[:,:,:49,:], '(b nw) h ws d -> b h d (nw ws)', h=self.num_heads, b=B_)
W = int(math.sqrt(N))
out = out.reshape(B_,self.num_heads,C//self.num_heads,-1)
q_idx_rev = q_idx_rev.unsqueeze(2).expand(out.size())
x = out.gather(dim=-1,index=q_idx_rev).reshape(B_,C,N).permute(0,2,1)
v = v.gather(dim=-1,index=q_idx_rev).reshape(B_,C,W,W)
v = self.get_v(v)
v = v.reshape(B_,C,N).permute(0,2,1)
x = x + v
x = self.proj(x)
x = self.proj_drop(x)
return x
class WindowAttention(nn.Module):
r""" Window based multi-head self attention (W-MSA) module with relative position bias.
It supports both of shifted and non-shifted window.
Args:
dim (int): Number of input channels.
window_size (tuple[int]): The height and width of the window.
num_heads (int): Number of attention heads.
qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
proj_drop (float, optional): Dropout ratio of output. Default: 0.0
"""
def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0., kmeans = False):
super().__init__()
self.dim = dim
self.window_size = window_size # Wh, Ww
self.ws = window_size
self.num_heads = num_heads
head_dim = dim // num_heads
self.scale = qk_scale or head_dim ** -0.5
self.kmeans = kmeans
# define a parameter table of relative position bias
if True:
self.relative_position_bias_table = nn.Parameter(
torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)) # 2*Wh-1 * 2*Ww-1, nH
# get pair-wise relative position index for each token inside the window
coords_h = torch.arange(self.window_size[0])
coords_w = torch.arange(self.window_size[1])
coords = torch.stack(torch.meshgrid([coords_h, coords_w])) # 2, Wh, Ww
coords_flatten = torch.flatten(coords, 1) # 2, Wh*Ww
relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] # 2, Wh*Ww, Wh*Ww
relative_coords = relative_coords.permute(1, 2, 0).contiguous() # Wh*Ww, Wh*Ww, 2
relative_coords[:, :, 0] += self.window_size[0] - 1 # shift to start from 0
relative_coords[:, :, 1] += self.window_size[1] - 1
relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
relative_position_index = relative_coords.sum(-1) # Wh*Ww, Wh*Ww
self.register_buffer("relative_position_index", relative_position_index)
trunc_normal_(self.relative_position_bias_table, std=.02)
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=.02)
self.softmax = nn.Softmax(dim=-1)
def forward(self, x, mask=None):
"""
Args:
x: input features with shape of (num_windows*B, N, C)
mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
"""
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] # make torchscript happy (cannot use tensor as tuple)
q = q * self.scale
attn = (q @ k.transpose(-2, -1))
relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1) # Wh*Ww,Wh*Ww,nH
relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous() # nH, Wh*Ww, Wh*Ww
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
def extra_repr(self) -> str:
return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'
def flops(self, N):
# calculate flops for 1 window with token length of N
flops = 0
# qkv = self.qkv(x)
flops += N * self.dim * 3 * self.dim
# attn = (q @ k.transpose(-2, -1))
flops += self.num_heads * N * (self.dim // self.num_heads) * N
# x = (attn @ v)
flops += self.num_heads * N * N * (self.dim // self.num_heads)
# x = self.proj(x)
flops += N * self.dim * self.dim
return flops
class SwinTransformerBlock_WMSA(nn.Module):
def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
act_layer=nn.GELU, norm_layer=nn.LayerNorm):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.num_heads = num_heads
self.window_size = window_size
self.shift_size = shift_size
self.mlp_ratio = mlp_ratio
if min(self.input_resolution) <= self.window_size:
# if window size is larger than input resolution, we don't partition windows
self.shift_size = 0
self.window_size = min(self.input_resolution)
assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
self.norm1 = norm_layer(dim)
self.attn = WindowAttention(
dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, kmeans=shift_size)
self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
self.norm2 = norm_layer(dim)
mlp_hidden_dim = int(dim * mlp_ratio)
self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
attn_mask = None
self.register_buffer("attn_mask", attn_mask)
def forward(self, x):
H, W = self.input_resolution
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
shortcut = x
x = self.norm1(x)
x = x.view(B, H, W, C)
shifted_x = x
# partition windows
x_windows = window_partition(shifted_x, self.window_size) # nW*B, window_size, window_size, C
x_windows = x_windows.view(-1, self.window_size * self.window_size, C) # nW*B, window_size*window_size, C
# W-MSA
attn_windows = self.attn(x_windows, mask=self.attn_mask) # nW*B, window_size*window_size, C
# merge windows
attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
shifted_x = window_reverse(attn_windows, self.window_size, H, W) # B H' W' C
x = shifted_x
x = x.view(B, H * W, C)
# FFN
x = shortcut + self.drop_path(x)
b,n,c = x.shape
#w = int(math.sqrt(n))
#x = self.norm3(x)
#x = x.view(b,w,w,c).permute(0,3,1,2)
#x = x + self.local(x)
#x = x.permute(0,2,3,1).reshape(b,n,c)
x = x + self.drop_path(self.mlp(self.norm2(x)))
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"
def flops(self):
flops = 0
H, W = self.input_resolution
# norm1
flops += self.dim * H * W
# W-MSA/SW-MSA
nW = H * W / self.window_size / self.window_size
flops += nW * self.attn.flops(self.window_size * self.window_size)
# mlp
flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
# norm2
flops += self.dim * H * W
return flops
class BLA_Block(nn.Module):
def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
act_layer=nn.GELU, norm_layer=nn.LayerNorm):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.num_heads = num_heads
self.window_size = window_size
self.shift_size = shift_size
self.mlp_ratio = mlp_ratio
if min(self.input_resolution) <= self.window_size:
# if window size is larger than input resolution, we don't partition windows
self.shift_size = 0
self.window_size = min(self.input_resolution)
assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
self.norm1 = norm_layer(dim)
self.attn = WindowAttention(
dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, kmeans=shift_size)
self.attnC = ContentAttention(
dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, kmeans=shift_size)
self.norm4 = norm_layer(dim);
self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
self.norm2 = norm_layer(dim)
mlp_hidden_dim = int(dim * mlp_ratio)
self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
#self.local = nn.Conv2d(dim, dim, window_size, 1, window_size//2, groups=dim, bias=qkv_bias)
#self.norm3 = norm_layer(dim)
#self.norm4 = norm_layer(dim)
# calculate attention mask for SW-MSA
H, W = self.input_resolution
img_mask = torch.zeros((1, H, W, 1)) # 1 H W 1
h_slices = (slice(0, -self.window_size),
slice(-self.window_size, -self.shift_size),
slice(-self.shift_size, None))
w_slices = (slice(0, -self.window_size),
slice(-self.window_size, -self.shift_size),
slice(-self.shift_size, None))
cnt = 0
for h in h_slices:
for w in w_slices:
img_mask[:, h, w, :] = cnt
cnt += 1
mask_windows = window_partition(img_mask, self.window_size) # nW, window_size, window_size, 1
mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
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))
self.register_buffer("attn_mask", attn_mask)
def forward(self, x):
H, W = self.input_resolution
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
shortcut = x
x = self.norm1(x)
x = x.view(B, H, W, C)
# cyclic shift
shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
# partition windows
x_windows = window_partition(shifted_x, self.window_size) # nW*B, window_size, window_size, C
x_windows = x_windows.view(-1, self.window_size * self.window_size, C) # nW*B, window_size*window_size, C
# SW-MSA
attn_windows = self.attn(x_windows, mask=self.attn_mask) # nW*B, window_size*window_size, C
# merge windows
attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
shifted_x = window_reverse(attn_windows, self.window_size, H, W) # B H' W' C
# reverse cyclic shift
x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
x = x.view(B, H * W, C)
# FFN
x = shortcut + self.drop_path(x)
x = x + self.attnC(self.norm4(x))
b,n,c = x.shape
#w = int(math.sqrt(n))
#x = self.norm3(x)
#x = x.view(b,w,w,c).permute(0,3,1,2)
#x = x + self.local(x)
#x = x.permute(0,2,3,1).reshape(b,n,c)
x = x + self.drop_path(self.mlp(self.norm2(x)))
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"
def flops(self):
flops = 0
H, W = self.input_resolution
# norm1
flops += self.dim * H * W
# W-MSA/SW-MSA
nW = H * W / self.window_size / self.window_size
flops += nW * self.attn.flops(self.window_size * self.window_size)
# mlp
flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
# norm2
flops += self.dim * H * W
return flops
class PatchMerging(nn.Module):
r""" Patch Merging Layer.
Args:
input_resolution (tuple[int]): Resolution of input feature.
dim (int): Number of input channels.
norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
"""
def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
super().__init__()
self.input_resolution = input_resolution
self.dim = dim
self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
self.norm = norm_layer(4 * dim)
def forward(self, x):
"""
x: B, H*W, C
"""
H, W = self.input_resolution
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."
x = x.view(B, H, W, C)
x0 = x[:, 0::2, 0::2, :] # B H/2 W/2 C
x1 = x[:, 1::2, 0::2, :] # B H/2 W/2 C
x2 = x[:, 0::2, 1::2, :] # B H/2 W/2 C
x3 = x[:, 1::2, 1::2, :] # B H/2 W/2 C
x = torch.cat([x0, x1, x2, x3], -1) # B H/2 W/2 4*C
x = x.view(B, -1, 4 * C) # B H/2*W/2 4*C
x = self.norm(x)
x = self.reduction(x)
return x
def extra_repr(self) -> str:
return f"input_resolution={self.input_resolution}, dim={self.dim}"
def flops(self):
H, W = self.input_resolution
flops = H * W * self.dim
flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
return flops
class BasicLayer1(nn.Module):
def __init__(self, dim, input_resolution, depth, num_heads, window_size,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.depth = depth
self.use_checkpoint = use_checkpoint
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock_WMSA(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[0] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
BLA_Block(dim=dim,input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=window_size // 2,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[1] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer)])
# patch merging layer
if downsample is not None:
self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def forward(self, x):
for blk in self.blocks:
if self.use_checkpoint:
x = checkpoint.checkpoint(blk, x)
else:
x = blk(x)
if self.downsample is not None:
x = self.downsample(x)
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
def flops(self):
flops = 0
for blk in self.blocks:
flops += blk.flops()
if self.downsample is not None:
flops += self.downsample.flops()
return flops
class BasicLayer2(nn.Module):
def __init__(self, dim, input_resolution, depth, num_heads, window_size,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.depth = depth
self.use_checkpoint = use_checkpoint
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock_WMSA(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[0] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
BLA_Block(dim=dim,input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=window_size // 2,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[1] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer)])
# patch merging layer
if downsample is not None:
self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def forward(self, x):
for blk in self.blocks:
if self.use_checkpoint:
x = checkpoint.checkpoint(blk, x)
else:
x = blk(x)
if self.downsample is not None:
x = self.downsample(x)
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
def flops(self):
flops = 0
for blk in self.blocks:
flops += blk.flops()
if self.downsample is not None:
flops += self.downsample.flops()
return flops
class BasicLayer3(nn.Module):
def __init__(self, dim, input_resolution, depth, num_heads, window_size,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.depth = depth
self.use_checkpoint = use_checkpoint
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock_WMSA(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[0] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
BLA_Block(dim=dim,input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=window_size // 2,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[1] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
SwinTransformerBlock_WMSA(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[2] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
BLA_Block(dim=dim,input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=window_size // 2,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[3] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
SwinTransformerBlock_WMSA(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[4] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
BLA_Block(dim=dim,input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=window_size // 2,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[5] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),])
# patch merging layer
if downsample is not None:
self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def forward(self, x):
for blk in self.blocks:
if self.use_checkpoint:
x = checkpoint.checkpoint(blk, x)
else:
x = blk(x)
if self.downsample is not None:
x = self.downsample(x)
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
def flops(self):
flops = 0
for blk in self.blocks:
flops += blk.flops()
if self.downsample is not None:
flops += self.downsample.flops()
return flops
class BasicLayer4(nn.Module):
def __init__(self, dim, input_resolution, depth, num_heads, window_size,
mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
super().__init__()
self.dim = dim
self.input_resolution = input_resolution
self.depth = depth
self.use_checkpoint = use_checkpoint
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock_WMSA(dim=dim, input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[0] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer),
SwinTransformerBlock_WMSA(dim=dim,input_resolution=input_resolution,
num_heads=num_heads, window_size=window_size,
shift_size=0,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop, attn_drop=attn_drop,
drop_path=drop_path[1] if isinstance(drop_path, list) else drop_path,
norm_layer=norm_layer)])
# patch merging layer
if downsample is not None:
self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def forward(self, x):
for blk in self.blocks:
if self.use_checkpoint:
x = checkpoint.checkpoint(blk, x)
else:
x = blk(x)
if self.downsample is not None:
x = self.downsample(x)
return x
def extra_repr(self) -> str:
return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
def flops(self):
flops = 0
for blk in self.blocks:
flops += blk.flops()
if self.downsample is not None:
flops += self.downsample.flops()
return flops
class PatchEmbed(nn.Module):
r""" Image to Patch Embedding
Args:
img_size (int): Image size. Default: 224.
patch_size (int): Patch token size. Default: 4.
in_chans (int): Number of input image channels. Default: 3.
embed_dim (int): Number of linear projection output channels. Default: 96.
norm_layer (nn.Module, optional): Normalization layer. Default: None
"""
def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
super().__init__()
img_size = to_2tuple(img_size)
patch_size = to_2tuple(patch_size)
patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
self.img_size = img_size
self.patch_size = patch_size
self.patches_resolution = patches_resolution
self.num_patches = patches_resolution[0] * patches_resolution[1]
self.in_chans = in_chans
self.embed_dim = embed_dim
self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
if norm_layer is not None:
self.norm = norm_layer(embed_dim)
else:
self.norm = None
def forward(self, x):
B, C, H, W = x.shape
# FIXME look at relaxing size constraints
assert H == self.img_size[0] and W == self.img_size[1], \
f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
x = self.proj(x).flatten(2).transpose(1, 2) # B Ph*Pw C
if self.norm is not None:
x = self.norm(x)
return x
def flops(self):
Ho, Wo = self.patches_resolution
flops = Ho * Wo * self.embed_dim * self.in_chans * (self.patch_size[0] * self.patch_size[1])
if self.norm is not None:
flops += Ho * Wo * self.embed_dim
return flops
class SwinTransformer(nn.Module):
def __init__(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000,
embed_dim=96, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24],
window_size=7, mlp_ratio=4., qkv_bias=True, qk_scale=None,
drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
use_checkpoint=False, **kwargs):
super().__init__()
self.num_classes = num_classes
self.num_layers = len(depths)
self.embed_dim = embed_dim
self.ape = ape
self.patch_norm = patch_norm
self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
self.mlp_ratio = mlp_ratio
# split image into non-overlapping patches
self.patch_embed = PatchEmbed(
img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
norm_layer=norm_layer if self.patch_norm else None)
num_patches = self.patch_embed.num_patches
patches_resolution = self.patch_embed.patches_resolution
self.patches_resolution = patches_resolution
# absolute position embedding
if self.ape:
self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
trunc_normal_(self.absolute_pos_embed, std=.02)
self.pos_drop = nn.Dropout(p=drop_rate)
# stochastic depth
dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] # stochastic depth decay rule
# build layers
self.layers = nn.ModuleList()
layer1 = BasicLayer1(dim=int(embed_dim * 2 ** 0),
input_resolution=(patches_resolution[0] // (2 ** 0),
patches_resolution[1] // (2 ** 0)),
depth=depths[0],
num_heads=num_heads[0],
window_size=window_size,
mlp_ratio=self.mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop_rate, attn_drop=attn_drop_rate,
drop_path=dpr[0:2],
norm_layer=norm_layer,
downsample=PatchMerging,
use_checkpoint=use_checkpoint)
self.layers.append(layer1)
layer2 = BasicLayer2(dim=int(embed_dim * 2 ** 1),
input_resolution=(patches_resolution[0] // (2 ** 1),
patches_resolution[1] // (2 ** 1)),
depth=depths[1],
num_heads=num_heads[1],
window_size=window_size,
mlp_ratio=self.mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop_rate, attn_drop=attn_drop_rate,
drop_path=dpr[2:4],
norm_layer=norm_layer,
downsample=PatchMerging,
use_checkpoint=use_checkpoint)
self.layers.append(layer2)
layer3 = BasicLayer3(dim=int(embed_dim * 2 ** 2),
input_resolution=(patches_resolution[0] // (2 ** 2),
patches_resolution[1] // (2 ** 2)),
depth=depths[2],
num_heads=num_heads[2],
window_size=window_size,
mlp_ratio=self.mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop_rate, attn_drop=attn_drop_rate,
drop_path=dpr[4:10],
norm_layer=norm_layer,
downsample=PatchMerging,
use_checkpoint=use_checkpoint)
self.layers.append(layer3)
layer4 = BasicLayer4(dim=int(embed_dim * 2 ** 3),
input_resolution=(patches_resolution[0] // (2 ** 3),
patches_resolution[1] // (2 ** 3)),
depth=depths[3],
num_heads=num_heads[3],
window_size=window_size,
mlp_ratio=self.mlp_ratio,
qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop_rate, attn_drop=attn_drop_rate,
drop_path=dpr[10:12],
norm_layer=norm_layer,
downsample=None,
use_checkpoint=use_checkpoint)
self.layers.append(layer4)
self.norm = norm_layer(self.num_features)
self.avgpool = nn.AdaptiveAvgPool1d(1)
self.head = nn.Linear(self.num_features, num_classes) if num_classes > 0 else nn.Identity()
self.apply(self._init_weights)
def _init_weights(self, m):
if isinstance(m, nn.Linear):
trunc_normal_(m.weight, std=.02)
if isinstance(m, nn.Linear) and m.bias is not None:
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.LayerNorm):
nn.init.constant_(m.bias, 0)
nn.init.constant_(m.weight, 1.0)
@torch.jit.ignore
def no_weight_decay(self):
return {'absolute_pos_embed'}
@torch.jit.ignore
def no_weight_decay_keywords(self):
return {'relative_position_bias_table'}
def forward_features(self, x):
x = self.patch_embed(x)
if self.ape:
x = x + self.absolute_pos_embed
x = self.pos_drop(x)
for layer in self.layers:
x = layer(x)
x = self.norm(x) # B L C
x = self.avgpool(x.transpose(1, 2)) # B C 1
x = torch.flatten(x, 1)
return x
def forward(self, x):
x = self.forward_features(x)
x = self.head(x)
return x
def flops(self):
flops = 0
flops += self.patch_embed.flops()
for i, layer in enumerate(self.layers):
flops += layer.flops()
flops += self.num_features * self.patches_resolution[0] * self.patches_resolution[1] // (2 ** self.num_layers)
flops += self.num_features * self.num_classes
return flops
