|
import torch
|
|
import torch.nn as nn
|
|
import torch.nn.functional as F
|
|
|
|
from .utils import split_feature, merge_splits
|
|
|
|
|
|
def single_head_full_attention(q, k, v):
|
|
|
|
assert q.dim() == k.dim() == v.dim() == 3
|
|
|
|
scores = torch.matmul(q, k.permute(0, 2, 1)) / (q.size(2) ** .5)
|
|
attn = torch.softmax(scores, dim=2)
|
|
out = torch.matmul(attn, v)
|
|
|
|
return out
|
|
|
|
|
|
def generate_shift_window_attn_mask(input_resolution, window_size_h, window_size_w,
|
|
shift_size_h, shift_size_w, device=torch.device('cuda')):
|
|
|
|
|
|
h, w = input_resolution
|
|
img_mask = torch.zeros((1, h, w, 1)).to(device)
|
|
h_slices = (slice(0, -window_size_h),
|
|
slice(-window_size_h, -shift_size_h),
|
|
slice(-shift_size_h, None))
|
|
w_slices = (slice(0, -window_size_w),
|
|
slice(-window_size_w, -shift_size_w),
|
|
slice(-shift_size_w, None))
|
|
cnt = 0
|
|
for h in h_slices:
|
|
for w in w_slices:
|
|
img_mask[:, h, w, :] = cnt
|
|
cnt += 1
|
|
|
|
mask_windows = split_feature(img_mask, num_splits=input_resolution[-1] // window_size_w, channel_last=True)
|
|
|
|
mask_windows = mask_windows.view(-1, window_size_h * window_size_w)
|
|
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
|
|
|
|
|
|
def single_head_split_window_attention(q, k, v,
|
|
num_splits=1,
|
|
with_shift=False,
|
|
h=None,
|
|
w=None,
|
|
attn_mask=None,
|
|
):
|
|
|
|
|
|
assert q.dim() == k.dim() == v.dim() == 3
|
|
|
|
assert h is not None and w is not None
|
|
assert q.size(1) == h * w
|
|
|
|
b, _, c = q.size()
|
|
|
|
b_new = b * num_splits * num_splits
|
|
|
|
window_size_h = h // num_splits
|
|
window_size_w = w // num_splits
|
|
|
|
q = q.view(b, h, w, c)
|
|
k = k.view(b, h, w, c)
|
|
v = v.view(b, h, w, c)
|
|
|
|
scale_factor = c ** 0.5
|
|
|
|
if with_shift:
|
|
assert attn_mask is not None
|
|
shift_size_h = window_size_h // 2
|
|
shift_size_w = window_size_w // 2
|
|
|
|
q = torch.roll(q, shifts=(-shift_size_h, -shift_size_w), dims=(1, 2))
|
|
k = torch.roll(k, shifts=(-shift_size_h, -shift_size_w), dims=(1, 2))
|
|
v = torch.roll(v, shifts=(-shift_size_h, -shift_size_w), dims=(1, 2))
|
|
|
|
q = split_feature(q, num_splits=num_splits, channel_last=True)
|
|
k = split_feature(k, num_splits=num_splits, channel_last=True)
|
|
v = split_feature(v, num_splits=num_splits, channel_last=True)
|
|
|
|
scores = torch.matmul(q.view(b_new, -1, c), k.view(b_new, -1, c).permute(0, 2, 1)
|
|
) / scale_factor
|
|
|
|
if with_shift:
|
|
scores += attn_mask.repeat(b, 1, 1)
|
|
|
|
attn = torch.softmax(scores, dim=-1)
|
|
|
|
out = torch.matmul(attn, v.view(b_new, -1, c))
|
|
|
|
out = merge_splits(out.view(b_new, h // num_splits, w // num_splits, c),
|
|
num_splits=num_splits, channel_last=True)
|
|
|
|
|
|
if with_shift:
|
|
out = torch.roll(out, shifts=(shift_size_h, shift_size_w), dims=(1, 2))
|
|
|
|
out = out.view(b, -1, c)
|
|
|
|
return out
|
|
|
|
|
|
class TransformerLayer(nn.Module):
|
|
def __init__(self,
|
|
d_model=256,
|
|
nhead=1,
|
|
attention_type='swin',
|
|
no_ffn=False,
|
|
ffn_dim_expansion=4,
|
|
with_shift=False,
|
|
**kwargs,
|
|
):
|
|
super(TransformerLayer, self).__init__()
|
|
|
|
self.dim = d_model
|
|
self.nhead = nhead
|
|
self.attention_type = attention_type
|
|
self.no_ffn = no_ffn
|
|
|
|
self.with_shift = with_shift
|
|
|
|
|
|
self.q_proj = nn.Linear(d_model, d_model, bias=False)
|
|
self.k_proj = nn.Linear(d_model, d_model, bias=False)
|
|
self.v_proj = nn.Linear(d_model, d_model, bias=False)
|
|
|
|
self.merge = nn.Linear(d_model, d_model, bias=False)
|
|
|
|
self.norm1 = nn.LayerNorm(d_model)
|
|
|
|
|
|
if not self.no_ffn:
|
|
in_channels = d_model * 2
|
|
self.mlp = nn.Sequential(
|
|
nn.Linear(in_channels, in_channels * ffn_dim_expansion, bias=False),
|
|
nn.GELU(),
|
|
nn.Linear(in_channels * ffn_dim_expansion, d_model, bias=False),
|
|
)
|
|
|
|
self.norm2 = nn.LayerNorm(d_model)
|
|
|
|
def forward(self, source, target,
|
|
height=None,
|
|
width=None,
|
|
shifted_window_attn_mask=None,
|
|
attn_num_splits=None,
|
|
**kwargs,
|
|
):
|
|
|
|
query, key, value = source, target, target
|
|
|
|
|
|
query = self.q_proj(query)
|
|
key = self.k_proj(key)
|
|
value = self.v_proj(value)
|
|
|
|
if self.attention_type == 'swin' and attn_num_splits > 1:
|
|
if self.nhead > 1:
|
|
|
|
|
|
raise NotImplementedError
|
|
else:
|
|
message = single_head_split_window_attention(query, key, value,
|
|
num_splits=attn_num_splits,
|
|
with_shift=self.with_shift,
|
|
h=height,
|
|
w=width,
|
|
attn_mask=shifted_window_attn_mask,
|
|
)
|
|
else:
|
|
message = single_head_full_attention(query, key, value)
|
|
|
|
message = self.merge(message)
|
|
message = self.norm1(message)
|
|
|
|
if not self.no_ffn:
|
|
message = self.mlp(torch.cat([source, message], dim=-1))
|
|
message = self.norm2(message)
|
|
|
|
return source + message
|
|
|
|
|
|
class TransformerBlock(nn.Module):
|
|
"""self attention + cross attention + FFN"""
|
|
|
|
def __init__(self,
|
|
d_model=256,
|
|
nhead=1,
|
|
attention_type='swin',
|
|
ffn_dim_expansion=4,
|
|
with_shift=False,
|
|
**kwargs,
|
|
):
|
|
super(TransformerBlock, self).__init__()
|
|
|
|
self.self_attn = TransformerLayer(d_model=d_model,
|
|
nhead=nhead,
|
|
attention_type=attention_type,
|
|
no_ffn=True,
|
|
ffn_dim_expansion=ffn_dim_expansion,
|
|
with_shift=with_shift,
|
|
)
|
|
|
|
self.cross_attn_ffn = TransformerLayer(d_model=d_model,
|
|
nhead=nhead,
|
|
attention_type=attention_type,
|
|
ffn_dim_expansion=ffn_dim_expansion,
|
|
with_shift=with_shift,
|
|
)
|
|
|
|
def forward(self, source, target,
|
|
height=None,
|
|
width=None,
|
|
shifted_window_attn_mask=None,
|
|
attn_num_splits=None,
|
|
**kwargs,
|
|
):
|
|
|
|
|
|
|
|
source = self.self_attn(source, source,
|
|
height=height,
|
|
width=width,
|
|
shifted_window_attn_mask=shifted_window_attn_mask,
|
|
attn_num_splits=attn_num_splits,
|
|
)
|
|
|
|
|
|
source = self.cross_attn_ffn(source, target,
|
|
height=height,
|
|
width=width,
|
|
shifted_window_attn_mask=shifted_window_attn_mask,
|
|
attn_num_splits=attn_num_splits,
|
|
)
|
|
|
|
return source
|
|
|
|
|
|
class FeatureTransformer(nn.Module):
|
|
def __init__(self,
|
|
num_layers=6,
|
|
d_model=128,
|
|
nhead=1,
|
|
attention_type='swin',
|
|
ffn_dim_expansion=4,
|
|
**kwargs,
|
|
):
|
|
super(FeatureTransformer, self).__init__()
|
|
|
|
self.attention_type = attention_type
|
|
|
|
self.d_model = d_model
|
|
self.nhead = nhead
|
|
|
|
self.layers = nn.ModuleList([
|
|
TransformerBlock(d_model=d_model,
|
|
nhead=nhead,
|
|
attention_type=attention_type,
|
|
ffn_dim_expansion=ffn_dim_expansion,
|
|
with_shift=True if attention_type == 'swin' and i % 2 == 1 else False,
|
|
)
|
|
for i in range(num_layers)])
|
|
|
|
for p in self.parameters():
|
|
if p.dim() > 1:
|
|
nn.init.xavier_uniform_(p)
|
|
|
|
def forward(self, feature0, feature1,
|
|
attn_num_splits=None,
|
|
**kwargs,
|
|
):
|
|
|
|
b, c, h, w = feature0.shape
|
|
assert self.d_model == c
|
|
|
|
feature0 = feature0.flatten(-2).permute(0, 2, 1)
|
|
feature1 = feature1.flatten(-2).permute(0, 2, 1)
|
|
|
|
if self.attention_type == 'swin' and attn_num_splits > 1:
|
|
|
|
window_size_h = h // attn_num_splits
|
|
window_size_w = w // attn_num_splits
|
|
|
|
|
|
shifted_window_attn_mask = generate_shift_window_attn_mask(
|
|
input_resolution=(h, w),
|
|
window_size_h=window_size_h,
|
|
window_size_w=window_size_w,
|
|
shift_size_h=window_size_h // 2,
|
|
shift_size_w=window_size_w // 2,
|
|
device=feature0.device,
|
|
)
|
|
else:
|
|
shifted_window_attn_mask = None
|
|
|
|
|
|
concat0 = torch.cat((feature0, feature1), dim=0)
|
|
concat1 = torch.cat((feature1, feature0), dim=0)
|
|
|
|
for layer in self.layers:
|
|
concat0 = layer(concat0, concat1,
|
|
height=h,
|
|
width=w,
|
|
shifted_window_attn_mask=shifted_window_attn_mask,
|
|
attn_num_splits=attn_num_splits,
|
|
)
|
|
|
|
|
|
concat1 = torch.cat(concat0.chunk(chunks=2, dim=0)[::-1], dim=0)
|
|
|
|
feature0, feature1 = concat0.chunk(chunks=2, dim=0)
|
|
|
|
|
|
feature0 = feature0.view(b, h, w, c).permute(0, 3, 1, 2).contiguous()
|
|
feature1 = feature1.view(b, h, w, c).permute(0, 3, 1, 2).contiguous()
|
|
|
|
return feature0, feature1
|
|
|
|
|
|
class FeatureFlowAttention(nn.Module):
|
|
"""
|
|
flow propagation with self-attention on feature
|
|
query: feature0, key: feature0, value: flow
|
|
"""
|
|
|
|
def __init__(self, in_channels,
|
|
**kwargs,
|
|
):
|
|
super(FeatureFlowAttention, self).__init__()
|
|
|
|
self.q_proj = nn.Linear(in_channels, in_channels)
|
|
self.k_proj = nn.Linear(in_channels, in_channels)
|
|
|
|
for p in self.parameters():
|
|
if p.dim() > 1:
|
|
nn.init.xavier_uniform_(p)
|
|
|
|
def forward(self, feature0, flow,
|
|
local_window_attn=False,
|
|
local_window_radius=1,
|
|
**kwargs,
|
|
):
|
|
|
|
if local_window_attn:
|
|
return self.forward_local_window_attn(feature0, flow,
|
|
local_window_radius=local_window_radius)
|
|
|
|
b, c, h, w = feature0.size()
|
|
|
|
query = feature0.view(b, c, h * w).permute(0, 2, 1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
query = self.q_proj(query)
|
|
key = self.k_proj(query)
|
|
|
|
value = flow.view(b, flow.size(1), h * w).permute(0, 2, 1)
|
|
|
|
scores = torch.matmul(query, key.permute(0, 2, 1)) / (c ** 0.5)
|
|
prob = torch.softmax(scores, dim=-1)
|
|
|
|
out = torch.matmul(prob, value)
|
|
out = out.view(b, h, w, value.size(-1)).permute(0, 3, 1, 2)
|
|
|
|
return out
|
|
|
|
def forward_local_window_attn(self, feature0, flow,
|
|
local_window_radius=1,
|
|
):
|
|
assert flow.size(1) == 2
|
|
assert local_window_radius > 0
|
|
|
|
b, c, h, w = feature0.size()
|
|
|
|
feature0_reshape = self.q_proj(feature0.view(b, c, -1).permute(0, 2, 1)
|
|
).reshape(b * h * w, 1, c)
|
|
|
|
kernel_size = 2 * local_window_radius + 1
|
|
|
|
feature0_proj = self.k_proj(feature0.view(b, c, -1).permute(0, 2, 1)).permute(0, 2, 1).reshape(b, c, h, w)
|
|
|
|
feature0_window = F.unfold(feature0_proj, kernel_size=kernel_size,
|
|
padding=local_window_radius)
|
|
|
|
feature0_window = feature0_window.view(b, c, kernel_size ** 2, h, w).permute(
|
|
0, 3, 4, 1, 2).reshape(b * h * w, c, kernel_size ** 2)
|
|
|
|
flow_window = F.unfold(flow, kernel_size=kernel_size,
|
|
padding=local_window_radius)
|
|
|
|
flow_window = flow_window.view(b, 2, kernel_size ** 2, h, w).permute(
|
|
0, 3, 4, 2, 1).reshape(b * h * w, kernel_size ** 2, 2)
|
|
|
|
scores = torch.matmul(feature0_reshape, feature0_window) / (c ** 0.5)
|
|
|
|
prob = torch.softmax(scores, dim=-1)
|
|
|
|
out = torch.matmul(prob, flow_window).view(b, h, w, 2).permute(0, 3, 1, 2).contiguous()
|
|
|
|
return out
|
|
|