Add Conv2() module (#2820)

docs/usage/engine.md

@@ -48,16 +48,17 @@ trainer.train()
 
 You now realize that you need to customize the trainer further to:
 
-* * Customize the `loss function`. 
+* Customize the `loss function`.
 * Add `callback` that uploads model to your Google Drive after every 10 `epochs`
   Here's how you can do it:
 
 ```python
 from ultralytics.yolo.v8.detect import DetectionTrainer
-from ultralytcs.nn.tasks import DetectionModel
+from ultralytics.nn.tasks import DetectionModel
+
 
 class MyCustomModel(DetectionModel):
-  def init_criterion():
+    def init_criterion(self):
         ...
 
 
@@ -65,6 +66,7 @@ class CustomTrainer(DetectionTrainer):
     def get_model(self, cfg, weights):
         return MyCustomModel(...)
 
+
 # callback to upload model weights
 def log_model(trainer):
     last_weight_path = trainer.last

ultralytics/models/rt-detr/rt-detr-l.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# RT-DETR-l object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/rtdetr
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/rt-detr/rt-detr-x.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# RT-DETR-x object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/rtdetr
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/v3/yolov3-spp.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv3-SPP object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/yolov3
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/v3/yolov3-tiny.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv3-tiny object detection model with P4-P5 outputs. For details see https://docs.ultralytics.com/models/yolov3
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/v3/yolov3.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv3 object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/yolov3
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/v5/yolov5-p6.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv5 object detection model with P3-P6 outputs. For details see https://docs.ultralytics.com/models/yolov5
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/v5/yolov5.yaml

@@ -1,4 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv5 object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/yolov5
 
 # Parameters
 nc: 80  # number of classes

ultralytics/models/v6/yolov6.yaml

@@ -1,51 +1,53 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
-# YOLOv6 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect
+# YOLOv6 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/models/yolov6
 
 # Parameters
 act: nn.ReLU()
 nc: 80  # number of classes
 scales: # model compound scaling constants, i.e. 'model=yolov6n.yaml' will call yolov8.yaml with scale 'n'
   # [depth, width, max_channels]
-  n: [ 0.33, 0.25, 1024 ]
+  n: [0.33, 0.25, 1024]
-  s: [ 0.33, 0.50, 1024 ]
+  s: [0.33, 0.50, 1024]
-  m: [ 0.67, 0.75, 768 ]
+  m: [0.67, 0.75, 768]
-  l: [ 1.00, 1.00, 512 ]
+  l: [1.00, 1.00, 512]
-  x: [ 1.00, 1.25, 512 ]
+  x: [1.00, 1.25, 512]
 
 # YOLOv6-3.0s backbone
 backbone:
   # [from, repeats, module, args]
-  - [ -1, 1, Conv, [ 64, 3, 2 ] ]  # 0-P1/2
+  - [-1, 1, Conv, [64, 3, 2]]  # 0-P1/2
-  - [ -1, 1, Conv, [ 128, 3, 2 ] ]  # 1-P2/4
+  - [-1, 1, Conv, [128, 3, 2]]  # 1-P2/4
-  - [ -1, 6, Conv, [ 128, 3, 1 ] ]
+  - [-1, 6, Conv, [128, 3, 1]]
-  - [ -1, 1, Conv, [ 256, 3, 2 ] ]  # 3-P3/8
+  - [-1, 1, Conv, [256, 3, 2]]  # 3-P3/8
-  - [ -1, 12, Conv, [ 256, 3, 1 ] ]
+  - [-1, 12, Conv, [256, 3, 1]]
-  - [ -1, 1, Conv, [ 512, 3, 2 ] ]  # 5-P4/16
+  - [-1, 1, Conv, [512, 3, 2]]  # 5-P4/16
-  - [ -1, 18, Conv, [ 512, 3, 1 ] ]
+  - [-1, 18, Conv, [512, 3, 1]]
-  - [ -1, 1, Conv, [ 1024, 3, 2 ] ]  # 7-P5/32
+  - [-1, 1, Conv, [1024, 3, 2]]  # 7-P5/32
-  - [ -1, 9, Conv, [ 1024, 3, 1 ] ]
+  - [-1, 6, Conv, [1024, 3, 1]]
-  - [ -1, 1, SPPF, [ 1024, 5 ] ]  # 9
+  - [-1, 1, SPPF, [1024, 5]]  # 9
 
 # YOLOv6-3.0s head
 head:
-  - [ -1, 1, nn.ConvTranspose2d, [ 256, 2, 2, 0 ] ]
+  - [-1, 1, Conv, [256, 1, 1]]
-  - [ [ -1, 6 ], 1, Concat, [ 1 ] ]  # cat backbone P4
+  - [-1, 1, nn.ConvTranspose2d, [256, 2, 2, 0]]
-  - [ -1, 1, Conv, [ 256, 3, 1 ] ]
+  - [[-1, 6], 1, Concat, [1]]  # cat backbone P4
-  - [ -1, 9, Conv, [ 256, 3, 1 ] ]  # 13
+  - [-1, 1, Conv, [256, 3, 1]]
+  - [-1, 9, Conv, [256, 3, 1]]  # 14
 
-  - [ -1, 1, nn.ConvTranspose2d, [ 128, 2, 2, 0 ] ]
+  - [-1, 1, Conv, [128, 1, 1]]
-  - [ [ -1, 4 ], 1, Concat, [ 1 ] ]  # cat backbone P3
+  - [-1, 1, nn.ConvTranspose2d, [128, 2, 2, 0]]
-  - [ -1, 1, Conv, [ 128, 3, 1 ] ]
+  - [[-1, 4], 1, Concat, [1]]  # cat backbone P3
-  - [ -1, 9, Conv, [ 128, 3, 1 ] ]  # 17
+  - [-1, 1, Conv, [128, 3, 1]]
+  - [-1, 9, Conv, [128, 3, 1]]  # 19
 
-  - [ -1, 1, Conv, [ 128, 3, 2 ] ]
+  - [-1, 1, Conv, [128, 3, 2]]
-  - [ [ -1, 12 ], 1, Concat, [ 1 ] ]  # cat head P4
+  - [[-1, 15], 1, Concat, [1]]  # cat head P4
-  - [ -1, 1, Conv, [ 256, 3, 1 ] ]
+  - [-1, 1, Conv, [256, 3, 1]]
-  - [ -1, 9, Conv, [ 256, 3, 1 ] ]  # 21
+  - [-1, 9, Conv, [256, 3, 1]]  # 23
 
-  - [ -1, 1, Conv, [ 256, 3, 2 ] ]
+  - [-1, 1, Conv, [256, 3, 2]]
-  - [ [ -1, 9 ], 1, Concat, [ 1 ] ]  # cat head P5
+  - [[-1, 10], 1, Concat, [1]]  # cat head P5
-  - [ -1, 1, Conv, [ 512, 3, 1 ] ]
+  - [-1, 1, Conv, [512, 3, 1]]
-  - [ -1, 9, Conv, [ 512, 3, 1 ] ]  # 25
+  - [-1, 9, Conv, [512, 3, 1]]  # 27
 
-  - [ [ 17, 21, 25 ], 1, Detect, [ nc ] ]  # Detect(P3, P4, P5)
+  - [[19, 23, 27], 1, Detect, [nc]]  # Detect(P3, P4, P5)

ultralytics/models/v8/yolov8-pose-p6.yaml

@@ -1,5 +1,5 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
-# YOLOv8 object detection model with P3-P6 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect
+# YOLOv8-pose keypoints/pose estimation model. For Usage examples see https://docs.ultralytics.com/tasks/pose
 
 # Parameters
 nc: 1  # number of classes

ultralytics/nn/modules/__init__.py

@@ -1,15 +1,28 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
+"""
+Ultralytics modules. Visualize with:
+
+from ultralytics.nn.modules import *
+import torch
+import os
+
+x = torch.ones(1, 128, 40, 40)
+m = Conv(128, 128)
+f = f'{m._get_name()}.onnx'
+torch.onnx.export(m, x, f)
+os.system(f'onnxsim {f} {f} && open {f}')
+"""
 
 from .block import (C1, C2, C3, C3TR, DFL, SPP, SPPF, Bottleneck, BottleneckCSP, C2f, C3Ghost, C3x, GhostBottleneck,
                     HGBlock, HGStem, Proto, RepC3)
-from .conv import (CBAM, ChannelAttention, Concat, Conv, ConvTranspose, DWConv, DWConvTranspose2d, Focus, GhostConv,
+from .conv import (CBAM, ChannelAttention, Concat, Conv, Conv2, ConvTranspose, DWConv, DWConvTranspose2d, Focus,
-                   LightConv, RepConv, SpatialAttention)
+                   GhostConv, LightConv, RepConv, SpatialAttention)
 from .head import Classify, Detect, Pose, RTDETRDecoder, Segment
 from .transformer import (AIFI, MLP, DeformableTransformerDecoder, DeformableTransformerDecoderLayer, LayerNorm2d,
                           MLPBlock, MSDeformAttn, TransformerBlock, TransformerEncoderLayer, TransformerLayer)
 
 __all__ = [
-    'Conv', 'LightConv', 'RepConv', 'DWConv', 'DWConvTranspose2d', 'ConvTranspose', 'Focus', 'GhostConv',
+    'Conv', 'Conv2', 'LightConv', 'RepConv', 'DWConv', 'DWConvTranspose2d', 'ConvTranspose', 'Focus', 'GhostConv',
     'ChannelAttention', 'SpatialAttention', 'CBAM', 'Concat', 'TransformerLayer', 'TransformerBlock', 'MLPBlock',
     'LayerNorm2d', 'DFL', 'HGBlock', 'HGStem', 'SPP', 'SPPF', 'C1', 'C2', 'C3', 'C2f', 'C3x', 'C3TR', 'C3Ghost',
     'GhostBottleneck', 'Bottleneck', 'BottleneckCSP', 'Proto', 'Detect', 'Segment', 'Pose', 'Classify',

ultralytics/nn/modules/conv.py

@@ -43,6 +43,27 @@ class Conv(nn.Module):
         return self.act(self.conv(x))
 
 
+class Conv2(Conv):
+    """Simplified RepConv module with Conv fusing."""
+
+    def __init__(self, c1, c2, k=3, s=1, p=None, g=1, d=1, act=True):
+        """Initialize Conv layer with given arguments including activation."""
+        super().__init__(c1, c2, k, s, p, g=g, d=d, act=act)
+        self.cv2 = nn.Conv2d(c1, c2, 1, s, autopad(1, p, d), groups=g, dilation=d, bias=False)  # add 1x1 conv
+
+    def forward(self, x):
+        """Apply convolution, batch normalization and activation to input tensor."""
+        return self.act(self.bn(self.conv(x) + self.cv2(x)))
+
+    def fuse_convs(self):
+        """Fuse parallel convolutions."""
+        w = torch.zeros_like(self.conv.weight.data)
+        i = [x // 2 for x in w.shape[2:]]
+        w[:, :, i[0] - 1:i[0], i[1] - 1:i[1]] = self.cv2.weight.data.clone()
+        self.conv.weight.data += w
+        self.__delattr__('cv2')
+
+
 class LightConv(nn.Module):
     """Light convolution with args(ch_in, ch_out, kernel).
     https://github.com/PaddlePaddle/PaddleDetection/blob/develop/ppdet/modeling/backbones/hgnet_v2.py

ultralytics/nn/tasks.py

@@ -8,9 +8,9 @@ import torch
 import torch.nn as nn
 
 from ultralytics.nn.modules import (AIFI, C1, C2, C3, C3TR, SPP, SPPF, Bottleneck, BottleneckCSP, C2f, C3Ghost, C3x,
-                                    Classify, Concat, Conv, ConvTranspose, Detect, DWConv, DWConvTranspose2d, Focus,
+                                    Classify, Concat, Conv, Conv2, ConvTranspose, Detect, DWConv, DWConvTranspose2d,
-                                    GhostBottleneck, GhostConv, HGBlock, HGStem, Pose, RepC3, RepConv, RTDETRDecoder,
+                                    Focus, GhostBottleneck, GhostConv, HGBlock, HGStem, Pose, RepC3, RepConv,
-                                    Segment)
+                                    RTDETRDecoder, Segment)
 from ultralytics.yolo.utils import DEFAULT_CFG_DICT, DEFAULT_CFG_KEYS, LOGGER, colorstr, emojis, yaml_load
 from ultralytics.yolo.utils.checks import check_requirements, check_suffix, check_yaml
 from ultralytics.yolo.utils.loss import v8ClassificationLoss, v8DetectionLoss, v8PoseLoss, v8SegmentationLoss
@@ -103,7 +103,9 @@ class BaseModel(nn.Module):
         """
         if not self.is_fused():
             for m in self.model.modules():
-                if isinstance(m, (Conv, DWConv)) and hasattr(m, 'bn'):
+                if isinstance(m, (Conv, Conv2, DWConv)) and hasattr(m, 'bn'):
+                    if isinstance(m, Conv2):
+                        m.fuse_convs()
                     m.conv = fuse_conv_and_bn(m.conv, m.bn)  # update conv
                     delattr(m, 'bn')  # remove batchnorm
                     m.forward = m.forward_fuse  # update forward