Apply Transformer in the backbone
1. Put the attention structure code in the common.py file
2. Take Yolov5 (v6.1) to add attention mechanism (1) (and attach more than 30 kinds of Attention schematic diagrams)
3. Take you hand in hand with Yolov5 (v6.1) to add attention mechanism (2) (add attention mechanism to C3 module)
4. YOLOv5 adds attention mechanism
1、 put it as the last part of the backbone instead of a C3 block.
class TransformerLayer(nn.Module):
# Transformer layer https://arxiv.org/abs/2010.11929 (LayerNorm layers removed for better performance)
def __init__(self, c, num_heads):
super().__init__()
self.q = nn.Linear(c, c, bias=False)
self.k = nn.Linear(c, c, bias=False)
self.v = nn.Linear(c, c, bias=False)
self.ma = nn.MultiheadAttention(embed_dim=c, num_heads=num_heads)
self.fc1 = nn.Linear(c, c, bias=False)
self.fc2 = nn.Linear(c, c, bias=False)
def forward(self, x):
x = self.ma(self.q(x), self.k(x), self.v(x))[0] + x
x = self.fc2(self.fc1(x)) + x
return x
class TransformerBlock(nn.Module):
# Vision Transformer https://arxiv.org/abs/2010.11929
def __init__(self, c1, c2, num_heads, num_layers):
super().__init__()
self.conv = None
if c1 != c2:
self.conv = Conv(c1, c2)
self.linear = nn.Linear(c2, c2) # learnable position embedding
self.tr = nn.Sequential(*(TransformerLayer(c2, num_heads) for _ in range(num_layers)))
self.c2 = c2
def forward(self, x):
if self.conv is not None:
x = self.conv(x)
b, _, w, h = x.shape
p = x.flatten(2).permute(2, 0, 1)
return self.tr(p + self.linear(p)).permute(1, 2, 0).reshape(b, self.c2, w, h)
class C3TR(C3):
# C3 module with TransformerBlock()
def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
super().__init__(c1, c2, n, shortcut, g, e)
c_ = int(c2 * e)
self.m = TransformerBlock(c_, c_, 4, n)
# Parameters
nc: 80 # number of classes
depth_multiple: 0.33 # model depth multiple
width_multiple: 0.50 # layer channel multiple
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
# YOLOv5 v6.0 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, C3, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 6, C3, [256]],
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, C3, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 3, C3TR, [1024]], # 9 <--- C3TR() Transformer module
[-1, 1, SPPF, [1024, 5]], # 9
]
# YOLOv5 v6.0 head
head:
[[-1, 1, Conv, [512, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, C3, [512, False]], # 13
[-1, 1, Conv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, C3, [256, False]], # 17 (P3/8-small)
[-1, 1, Conv, [256, 3, 2]],
[[-1, 14], 1, Concat, [1]], # cat head P4
[-1, 3, C3, [512, False]], # 20 (P4/16-medium)
[-1, 1, Conv, [512, 3, 2]],
[[-1, 10], 1, Concat, [1]], # cat head P5
[-1, 3, C3, [1024, False]], # 23 (P5/32-large)
[[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
class SE(nn.Module):
def __init__(self, c1, c2, ratio=16):
super(SE, self).__init__()
#c*1*1
self.avgpool = nn.AdaptiveAvgPool2d(1)
self.l1 = nn.Linear(c1, c1 // ratio, bias=False)
self.relu = nn.ReLU(inplace=True)
self.l2 = nn.Linear(c1 // ratio, c1, bias=False)
self.sig = nn.Sigmoid()
def forward(self, x):
b, c, _, _ = x.size()
y = self.avgpool(x).view(b, c)
y = self.l1(y)
y = self.relu(y)
y = self.l2(y)
y = self.sig(y)
y = y.view(b, c, 1, 1)
return x * y.expand_as(x)
2. The second step; find the parse_model function in the yolo.py file and add the class name to it
def parse_model(d, ch): # model_dict, input_channels(3)
LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}")
anchors, nc, gd, gw = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple']
na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors
no = na * (nc + 5) # number of outputs = anchors * (classes + 5)
layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out
for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']): # from, number, module, args
m = eval(m) if isinstance(m, str) else m # eval strings
for j, a in enumerate(args):
try:
args[j] = eval(a) if isinstance(a, str) else a # eval strings
except NameError:
pass
n = n_ = max(round(n * gd), 1) if n > 1 else n # depth gain
#此处添加了C3TR
if m in [Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
BottleneckCSP, C3, C3TR, C3SPP, C3Ghost]:
c1, c2 = ch[f], args[0]
if c2 != no: # if not output
c2 = make_divisible(c2 * gw, 8)
args = [c1, c2, *args[1:]]
#此处添加了C3TR
if m in [BottleneckCSP, C3, C3TR, C3Ghost]:
args.insert(2, n) # number of repeats
n = 1
elif m is nn.BatchNorm2d:
args = [ch[f]]
elif m is Concat:
c2 = sum(ch[x] for x in f)
elif m is Detect:
args.append([ch[x] for x in f])
if isinstance(args[1], int): # number of anchors
args[1] = [list(range(args[1] * 2))] * len(f)
elif m is Contract:
c2 = ch[f] * args[0] ** 2
elif m is Expand:
c2 = ch[f] // args[0] ** 2
else:
c2 = ch[f]
m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module
t = str(m)[8:-2].replace('__main__.', '') # module type
np = sum(x.numel() for x in m_.parameters()) # number params
m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params
LOGGER.info(f'{i:>3}{str(f):>18}{n_:>3}{np:10.0f} {t:<40}{str(args):<30}') # print
save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist
layers.append(m_)
if i == 0:
ch = []
ch.append(c2)
return nn.Sequential(*layers), sorted(save)
3. The third step: modify the configuration file (I’ll take yolov5s.yaml as an example here), and add the attention layer to the position you want to add; commonly used is to add it to the last layer of the backbone, or to C3, here is added to the last layer
4. When a new layer is added to the network, the numbers of the subsequent layers of the network will change, and the corresponding positions will be modified.
Look at the picture below, the original Detect specified [17, 20, 23] [17,20,23][17,20,23] layer, so after we added the SE attention layer, we also need to modify it here, That is, the original 17 1717 layer has become 18 1818 layers; the original 20 2020 layers have become 21 2121 layers; the original 23 2323 layers have become 24 2424 layers; so the from coefficient of Detecet should be changed to [ 18 , 21 , 24 ] [18,21,24][18,21,24]
Similarly, the from coefficient of Concat should also be modified, so as to keep the original network structure from changing significantly. We just added the SE layer to the 9th layer , so the numbers after the 9th layer will be increased by 1. Here we need to change the from coefficients of the next two Concats from [ − 1 , 14 ] , [ − 1 , 10 ] [-1,14],[-1,10 ][−1,14],[−1,10] is changed to [ − 1 , 15 ] , [ − 1 , 11 ] [-1,15],[-1,11][−1,15],[ −1,11]
Here is the complete configuration file SE.yaml after I added the SE attention layer
# Parameters
nc: 80 # number of classes
depth_multiple: 0.33 # model depth multiple
width_multiple: 0.50 # layer channel multiple
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
# YOLOv5 v6.0 backbone+SE
backbone:
# [from, number, module, args]
[[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, C3, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 6, C3, [256]],
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, C3, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 3, C3, [1024]],
[-1,1,SE,[1024]], #SE
[-1, 1, SPPF, [1024, 5]], # 10
]
# YOLOv5+SE v6.0 head
head:
[[-1, 1, Conv, [512, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, C3, [512, False]], # 14
[-1, 1, Conv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, C3, [256, False]], # 18 (P3/8-small)
[-1, 1, Conv, [256, 3, 2]],
[[-1, 15], 1, Concat, [1]], # cat head P4
[-1, 3, C3, [512, False]], # 21 (P4/16-medium)
[-1, 1, Conv, [512, 3, 2]],
[[-1, 11], 1, Concat, [1]], # cat head P5
[-1, 3, C3, [1024, False]], # 24 (P5/32-large)
[[18, 21, 24], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
]
After adding it, you can train, and you will see the position of our attention layer when running:
