add deconv pre

regainOWO · Aug 22, 2022 · ff73dad · ff73dad
1 parent a3826aa
commit ff73dad
Show file tree

Hide file tree

Showing 8 changed files with 320 additions and 6 deletions.
diff --git a/README.md b/README.md
@@ -188,7 +188,7 @@ $ python wbf.py
 ```
 ______________________________________________________________________
 
-### Benchmark
+### Performance
 | Model                                                                                                | size<br><sup>(pixels) | mAP<sup>val<br>0.5:0.95 | mAP<sup>val<br>0.5 | Speed<br><sup>CPU b1<br>(ms) | Speed<br><sup>V100 b1<br>(ms) | Speed<br><sup>V100 b32<br>(ms) | params<br><sup>(M) | FLOPs<br><sup>@640 (B) |
 |------------------------------------------------------------------------------------------------------|-----------------------|-------------------------|--------------------|------------------------------|-------------------------------|--------------------------------|--------------------|------------------------|
 | YOLOv5n                   | 640                   | 28.0                    | 45.7               | **45**                       | **6.3**                       | **0.6**                        | **1.9**            | **4.5**                |
@@ -203,8 +203,46 @@ ______________________________________________________________________
 | YOLOv5l6                 | 1280                  | 53.7                    | 71.3               | 1784                         | 15.8                          | 10.5                           | 76.8               | 111.4                  |
 | YOLOv5x6<br>+ TTA | 1280<br>1536          | 55.0<br>**55.8**        | 72.7<br>**72.7**   | 3136<br>-                    | 26.2<br>-                     | 19.4<br>-                      | 140.7<br>-         | 209.8<br>-             |
 
+<details><summary> <b>Expand</b> </summary>
+
+* The original version was created based on YOLOv5
+
+</details>
+
+______________________________________________________________________
+
+| Model                                                                                                | size<br><sup>(pixels) | mAP<sup>val<br>0.5:0.95 | mAP<sup>val<br>0.5 | deconv kernel size<br><sup> | Speed<br><sup>V100 b1<br>(ms) | Speed<br><sup>V100 b32<br>(ms) | params<br><sup>(M) | FLOPs<br><sup>@640 (B) |
+|------------------------------------------------------------------------------------------------------|-----------------------|-------------------------|--------------------|------------------------------|-------------------------------|--------------------------------|--------------------|------------------------|
+| YOLOv5s                   | 640                   | 33.7                    | 52.9               | -                       | **5.6**                       | **2.2**                        | **7.23**            | **16.5**                |
+| YOLOv5s-deconv-exp1                   | 640                   | 33.4                    | 52.5               | 2                       | **5.6**                       | 2.4                        | 7.55            | 18.2                |
+| YOLOv5s-deconv-exp2                   | 640                   | **34.7**                    | **54.2**               | 4                           | 5.8                           | 2.5                            | 8.54                | 23.2                   |
+<details><summary> <b>Expand</b> </summary>
+
+* The training process depends on 4xV100 GPU
+```
+# train
+python -m torch.distributed.run --nproc_per_node 4 train.py --device 0,1,2,3 --data data/coco.yaml --hyp data/hyps/hyp.scratch-low.yaml  --cfg path/to/model.yaml --batch 512 --epochs 300 --weights ''
+# val
+python val.py --verbose --data data/coco.yaml --conf 0.001 --iou 0.65 --batch 1 --weights path/to/model.pt
+```
+</details>
+
 ______________________________________________________________________
 
+| Model                                                                                                | size<br><sup>(pixels) | mAP<sup>val<br>0.5:0.95 | mAP<sup>val<br>0.5 | params<br><sup>(M) | FLOPs<br><sup>@640 (B) |
+|------------------------------------------------------------------------------------------------------|-----------------------|-------------------------|--------------------------------|--------------------|------------------------|
+| YOLOv5s                   | 640                   | 37.4                       | 56.6                        | **7.226**            | **16.5**                |
+| YOLOv5s-decon             | 640                   | **37.8**                       | **57.1**                        | 7.232            | **16.5**                |
+
+<details><summary> <b>Expand</b> </summary>
+
+* tested the 4x4 depthwise-separable deconv by setting the groups as input_channel
+* their params number and FLOPS are nearly the same while the new model's mAP is about 0.4 higher than the origin.
+</details>
+
+______________________________________________________________________
+
+
 ### YOLO网络模型具体改进方式教程及原理参考
 
 - 11.[改进YOLOv5系列：11.ConvNeXt结合YOLO | CVPR2022 多种搭配，即插即用 | Backbone主干CNN模型](https://blog.csdn.net/qq_38668236/article/details/126454548)

diff --git a/README_EN.md b/README_EN.md
@@ -28,6 +28,8 @@ The YOLOAir algorithm library is a combination toolbox of a series of YOLO detec
 
 ✨Based on the YOLOv5 code framework, and synchronously adapt to the **stable YOLOv5_v6.1 update**, synchronize the v6.1 deployment ecosystem. Before using this project, you can learn about the YOLOv5 library.
 
+______________________________________________________________________
+
 [Features🚀](#Mainfeatures) • [Usage🍉](#Usage) • [Documentation📒](https://github.com/iscyy/yoloair) • [issues🌟](https://github.com/iscyy/yoloair/issues/new)
 
 ![](https://img.shields.io/badge/News-2022-red)  ![](https://img.shields.io/badge/Update-YOLOAir-orange) ![](https://visitor-badge.glitch.me/badge?page_id=iscyy.yoloair)  
@@ -38,7 +40,8 @@ The YOLOAir algorithm library is a combination toolbox of a series of YOLO detec
 
 Project address: https://github.com/iscyy/yoloair
 
-Part of the improved tutorial: [CSDN芒果汁没有芒果](https://blog.csdn.net/qq_38668236?type=blog)
+
+______________________________________________________________________
 
 ### Main features🚀
 
@@ -57,6 +60,8 @@ The above multiple detection algorithms use a unified model code framework, whic
 
 🔥🔥🔥 Heavy! ! ! As a supplement to the open source project of the attention mechanism, a 6300+ 🌟Star attention mechanism algorithm code base is strongly recommended 👉[External-Attention-pytorch](https://github.com/xmu-xiaoma666/External-Attention-pytorch)，The summary is very comprehensive, including various Attention, Self-Attention and other codes. The code is concise and easy to read, and one line of code implements the Attention mechanism. Welcome everyone to play!
 
+______________________________________________________________________
+
 🚀Support loading the official pre-training weights of YOLOv3, YOLOv4, YOLOv5, YOLOv7, YOLOR and other networks for transfer learning
 
 🚀 Support more Backbone
@@ -140,7 +145,7 @@ Support SPD-Conv✅
 Support HorNet Network ✅
 Support ConvNext module ✅
 Updating...
-
+______________________________________________________________________
 ## Usage🍉
 
 **About the code.** Follow the design principle of [YOLOv5](https://github.com/ultralytics/yolov5).  
@@ -179,9 +184,60 @@ You just need to set img path and txt path in wbf.py file.
 ```bash
 $ python wbf.py
 ```
+______________________________________________________________________
+### Performance
+| Model                                                                                                | size<br><sup>(pixels) | mAP<sup>val<br>0.5:0.95 | mAP<sup>val<br>0.5 | Speed<br><sup>CPU b1<br>(ms) | Speed<br><sup>V100 b1<br>(ms) | Speed<br><sup>V100 b32<br>(ms) | params<br><sup>(M) | FLOPs<br><sup>@640 (B) |
+|------------------------------------------------------------------------------------------------------|-----------------------|-------------------------|--------------------|------------------------------|-------------------------------|--------------------------------|--------------------|------------------------|
+| YOLOv5n                   | 640                   | 28.0                    | 45.7               | **45**                       | **6.3**                       | **0.6**                        | **1.9**            | **4.5**                |
+| YOLOv5s                   | 640                   | 37.4                    | 56.8               | 98                           | 6.4                           | 0.9                            | 7.2                | 16.5                   |
+| YOLOv5m                   | 640                   | 45.4                    | 64.1               | 224                          | 8.2                           | 1.7                            | 21.2               | 49.0                   |
+| YOLOv5l                   | 640                   | 49.0                    | 67.3               | 430                          | 10.1                          | 2.7                            | 46.5               | 109.1                  |
+| YOLOv5x                   | 640                   | 50.7                    | 68.9               | 766                          | 12.1                          | 4.8                            | 86.7               | 205.7                  |
+|                                                                                                      |                       |                         |                    |                              |                               |                                |                    |                        |
+| YOLOv5n6                 | 1280                  | 36.0                    | 54.4               | 153                          | 8.1                           | 2.1                            | 3.2                | 4.6                    |
+| YOLOv5s6                 | 1280                  | 44.8                    | 63.7               | 385                          | 8.2                           | 3.6                            | 12.6               | 16.8                   |
+| YOLOv5m6                 | 1280                  | 51.3                    | 69.3               | 887                          | 11.1                          | 6.8                            | 35.7               | 50.0                   |
+| YOLOv5l6                 | 1280                  | 53.7                    | 71.3               | 1784                         | 15.8                          | 10.5                           | 76.8               | 111.4                  |
+| YOLOv5x6<br>+ TTA | 1280<br>1536          | 55.0<br>**55.8**        | 72.7<br>**72.7**   | 3136<br>-                    | 26.2<br>-                     | 19.4<br>-                      | 140.7<br>-         | 209.8<br>-             |
 
-### Benchmark
-Updating...
+<details><summary> <b>Expand</b> </summary>
+
+* The original version was created based on YOLOv5
+
+</details>
+
+______________________________________________________________________
+
+| Model                                                                                                | size<br><sup>(pixels) | mAP<sup>val<br>0.5:0.95 | mAP<sup>val<br>0.5 | deconv kernel size<br><sup> | Speed<br><sup>V100 b1<br>(ms) | Speed<br><sup>V100 b32<br>(ms) | params<br><sup>(M) | FLOPs<br><sup>@640 (B) |
+|------------------------------------------------------------------------------------------------------|-----------------------|-------------------------|--------------------|------------------------------|-------------------------------|--------------------------------|--------------------|------------------------|
+| YOLOv5s                   | 640                   | 33.7                    | 52.9               | -                       | **5.6**                       | **2.2**                        | **7.23**            | **16.5**                |
+| YOLOv5s-deconv-exp1                   | 640                   | 33.4                    | 52.5               | 2                       | **5.6**                       | 2.4                        | 7.55            | 18.2                |
+| YOLOv5s-deconv-exp2                   | 640                   | **34.7**                    | **54.2**               | 4                           | 5.8                           | 2.5                            | 8.54                | 23.2                   |
+<details><summary> <b>Expand</b> </summary>
+
+* The training process depends on 4xV100 GPU
+```
+# train
+python -m torch.distributed.run --nproc_per_node 4 train.py --device 0,1,2,3 --data data/coco.yaml --hyp data/hyps/hyp.scratch-low.yaml  --cfg path/to/model.yaml --batch 512 --epochs 300 --weights ''
+# val
+python val.py --verbose --data data/coco.yaml --conf 0.001 --iou 0.65 --batch 1 --weights path/to/model.pt
+```
+</details>
+
+______________________________________________________________________
+
+| Model                                                                                                | size<br><sup>(pixels) | mAP<sup>val<br>0.5:0.95 | mAP<sup>val<br>0.5 | params<br><sup>(M) | FLOPs<br><sup>@640 (B) |
+|------------------------------------------------------------------------------------------------------|-----------------------|-------------------------|--------------------------------|--------------------|------------------------|
+| YOLOv5s                   | 640                   | 37.4                       | 56.6                        | **7.226**            | **16.5**                |
+| YOLOv5s-decon             | 640                   | **37.8**                       | **57.1**                        | 7.232            | **16.5**                |
+
+<details><summary> <b>Expand</b> </summary>
+
+* tested the 4x4 depthwise-separable deconv by setting the groups as input_channel
+* their params number and FLOPS are nearly the same while the new model's mAP is about 0.4 higher than the origin.
+</details>
+
+______________________________________________________________________
 
 ### YOLO network model specific improvement method tutorial and principle reference
 

diff --git a/configs/yolov5-Improved/yolov5s_deconv.yaml b/configs/yolov5-Improved/yolov5s_deconv.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# 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, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head with Deconvolution Upsample
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.ConvTranspose2d, [512, 4, 2, 1, 0, 512]],
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.ConvTranspose2d, [256, 4, 2, 1, 0, 256]],
+   [[-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)
+  ]
diff --git a/configs/yolov5-Improved/yolov5s_deconv_exp.yaml b/configs/yolov5-Improved/yolov5s_deconv_exp.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# 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, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head with Deconvolution Upsample
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.ConvTranspose2d, [512, 2, 2]], # nn.ConvTransposed2d with 2x2 deconv kernel
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.ConvTranspose2d, [256, 2, 2]], # nn.ConvTransposed2d with 2x2 deconv kernel
+   [[-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)
+  ]
diff --git a/configs/yolov5-Improved/yolov5s_dwconv_exp2.yaml b/configs/yolov5-Improved/yolov5s_dwconv_exp2.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# 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, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head with Deconvolution Upsample
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.ConvTranspose2d, [512, 4, 2, 1]], # nn.ConvTransposed2d with 4x4 deconv kernel
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.ConvTranspose2d, [256, 4, 2, 1]], # nn.ConvTransposed2d with 4x4 deconv kernel
+   [[-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)
+  ]