add C3f2

ultralytics/cfg/models/exp/yolov8-c3f2-2222-0.25-True.yaml

@@ -0,0 +1,46 @@
+# Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect
+
+# Parameters
+nc: 80 # number of classes
+scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n'
+  # [depth, width, max_channels]
+  n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers,  3157200 parameters,  3157184 gradients,   8.9 GFLOPs
+  s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients,  28.8 GFLOPs
+  m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients,  79.3 GFLOPs
+  l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs
+  x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs
+
+# YOLOv8.0n backbone
+backbone:
+  # [from, repeats, module, args]
+  - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2
+  - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4
+  - [-1, 2, C3f2, [256, True, 1, 0.25]]
+  - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8
+  - [-1, 2, C3f2, [512, True, 1, 0.25]]
+  - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16
+  - [-1, 2, C3f2, [512, True]]
+  - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32
+  - [-1, 2, C3f2, [1024, True]]
+  - [-1, 1, SPPF, [1024, 5]] # 9
+
+# YOLOv8.0n head
+head:
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 6], 1, Concat, [1]] # cat backbone P4
+  - [-1, 2, C3f2, [512, True]] # 12
+
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 4], 1, Concat, [1]] # cat backbone P3
+  - [-1, 2, C3f2, [256, True]] # 15 (P3/8-small)
+
+  - [-1, 1, Conv, [256, 3, 2]]
+  - [[-1, 12], 1, Concat, [1]] # cat head P4
+  - [-1, 2, C3f2, [512, True]] # 18 (P4/16-medium)
+
+  - [-1, 1, Conv, [512, 3, 2]]
+  - [[-1, 9], 1, Concat, [1]] # cat head P5
+  - [-1, 2, C3f2, [1024, True]] # 21 (P5/32-large)
+
+  - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5)

ultralytics/cfg/models/exp/yolov8-c3f2-2222-True.yaml

@@ -0,0 +1,46 @@
+# Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect
+
+# Parameters
+nc: 80 # number of classes
+scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n'
+  # [depth, width, max_channels]
+  n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers,  3157200 parameters,  3157184 gradients,   8.9 GFLOPs
+  s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients,  28.8 GFLOPs
+  m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients,  79.3 GFLOPs
+  l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs
+  x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs
+
+# YOLOv8.0n backbone
+backbone:
+  # [from, repeats, module, args]
+  - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2
+  - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4
+  - [-1, 2, C3f2, [128, True]]
+  - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8
+  - [-1, 2, C3f2, [256, True]]
+  - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16
+  - [-1, 2, C3f2, [512, True]]
+  - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32
+  - [-1, 2, C3f2, [1024, True]]
+  - [-1, 1, SPPF, [1024, 5]] # 9
+
+# YOLOv8.0n head
+head:
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 6], 1, Concat, [1]] # cat backbone P4
+  - [-1, 2, C3f2, [512, True]] # 12
+
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 4], 1, Concat, [1]] # cat backbone P3
+  - [-1, 2, C3f2, [256, True]] # 15 (P3/8-small)
+
+  - [-1, 1, Conv, [256, 3, 2]]
+  - [[-1, 12], 1, Concat, [1]] # cat head P4
+  - [-1, 2, C3f2, [512, True]] # 18 (P4/16-medium)
+
+  - [-1, 1, Conv, [512, 3, 2]]
+  - [[-1, 9], 1, Concat, [1]] # cat head P5
+  - [-1, 2, C3f2, [1024, True]] # 21 (P5/32-large)
+
+  - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5)

ultralytics/nn/modules/__init__.py

@@ -47,6 +47,7 @@ from .block import (
     ResNetLayer,
     Silence,
     C2f2,
+    C3f2,
     C2k2,
     C3k2,
     C3K2,
@@ -112,6 +113,7 @@ __all__ = (
     "C3",
     "C2f",
     "C2f2",
+    "C3f2",
     "C2k2",
     "C3k2",
     "C3K2",

ultralytics/nn/modules/block.py

@@ -38,6 +38,7 @@ __all__ = (
     "CBLinear",
     "Silence",
     "C2f2",
+    "C3f2",
     "C2k2",
     "C3k2",
     "C3K2",
@@ -271,6 +272,27 @@ class C2f2(nn.Module):
         return self.cv2(torch.cat(y, 1))
 
 
+class C3f2(nn.Module):
+    """Faster Implementation of CSP Bottleneck with 2 convolutions."""
+
+    def __init__(self, c1, c2, n=1, shortcut=False, g=1, e=0.5):
+        """Initialize CSP bottleneck layer with two convolutions with arguments ch_in, ch_out, number, shortcut, groups,
+        expansion.
+        """
+        super().__init__()
+        c_ = int(c2 * e)  # hidden channels
+        self.cv1 = Conv(c1, c_, 1, 1)
+        self.cv2 = Conv(c1, c_, 1, 1)
+        self.cv3 = Conv((2 + n) * c_, c2, 1)  # optional act=FReLU(c2)
+        self.m = nn.ModuleList(C3k(c_, c_, 2, shortcut, g) for _ in range(n))
+
+    def forward(self, x):
+        """Forward pass through C2f layer."""
+        y = [self.cv2(x), self.cv1(x)]
+        y.extend(m(y[-1]) for m in self.m)
+        return self.cv3(torch.cat(y, 1))
+
+
 class C2k2(C2f2):
     """Faster Implementation of CSP Bottleneck with 2 convolutions."""
 

ultralytics/nn/tasks.py

@@ -22,6 +22,7 @@ from ultralytics.nn.modules import (
     BottleneckCSP,
     C2f,
     C2f2,
+    C3f2,
     C2k2,
     C3k2,
     C3s2,
@@ -883,6 +884,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
             C2,
             C2f,
             C2f2,
+            C3f2,
             C2k2,
             C3k2,
             C3n2,
@@ -912,7 +914,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
                 )  # num heads
 
             args = [c1, c2, *args[1:]]
-            if m in (BottleneckCSP, C1, C2, C2f, C2f2, C2k2, C3s2, C3n2, C3k2, C3K2, C3m1, C3k3, C2fAttn, C3, C3TR, C3Ghost, C3x, RepC3):
+            if m in (BottleneckCSP, C1, C2, C2f, C3f2, C2k2, C3s2, C3n2, C3k2, C3K2, C3m1, C3k3, C2fAttn, C3, C3TR, C3Ghost, C3x, RepC3):
                 args.insert(2, n)  # number of repeats
                 n = 1
         elif m is AIFI: