init repo

README.md

@@ -1,3 +1,40 @@
 ---
 license: agpl-3.0
+library: ultralytics
+tags:
+  - object-detection
+  - pytorch
+  - coco
+  - p2-p6-layer
+  - yolov8
 ---
+
+# YOLOv8-p2 COCO Pretrained Model
+
+This model is a YOLOv8-p26 model trained on the COCO dataset, **with P2-P6 output layers.** 
+
+## Example Usage
+
+```angular2html
+from huggingface_hub import hf_hub_download
+from ultralytics import YOLO
+from PIL import Image
+
+model_path = hf_hub_download("soyeollee/yolov8x-p26-coco", "model.pt")
+model = YOLO(model_path)
+
+image_path = "/path/to/image"
+output = model(Image.open(image_path))
+```
+
+
+## Performance
+
+| metric (maxDets=100)                            | yolov8x <br/>(official) | [yolov8x-p2](https://huggingface.co/soyeollee/yolov8x-p2-coco) | yolov8x-p26<br/>(this repo) | yolov11x <br/>(official) | [yolo11x-p2](https://huggingface.co/soyeollee/yolo11x-p2-coco) |
+|-------------------------------------------------|--------------------------|--------------------------------|-----------------------------|--------------------------|--------------------------------|
+| AP @[ IoU=0.50:0.95 / area=   all / ]           | **0.540**                | **0.541**                      | **0.544**                   | **0.546**                | **0.553**                      |
+| AP @[ IoU=0.50      / area=   all / ]           | 0.710                    | 0.712                          | 0.713                       | 0.716                    | 0.722                          |
+| AP @[ IoU=0.75      / area=   all / ]           | 0.588                    | 0.590                          | 0.593                       | 0.595                    | 0.606                          |
+| AP @[ IoU=0.50:0.95 / area= small / ]           | 0.360                    | 0.386                          | 0.381                       | 0.377                    | 0.386                          |
+| AP @[ IoU=0.50:0.95 /area=medium  / ]           | 0.594                    | 0.586                          | 0.586                       | 0.597                    | 0.595                          |
+| AP @[ IoU=0.50:0.95 / area= large / ]           | 0.707                    | 0.686                          | 0.691                       | 0.702                    | 0.702                          |

events.out.tfevents.1717087574.05dad19e7779.6477.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0002608f81ab2d7c5ada7f959e2d0b9129909778e3c4afa4ad860e7ff7473323
+size 779594

model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25b20fe3155491f3b289ae42de854cd0552b27d3d2fdb7351a4f3bf1acaf93fd
+size 195340380

train.py

@@ -0,0 +1,36 @@
+import yaml
+import os
+
+from ultralytics import YOLO
+
+
+config = {
+    'name': '01_v8x-p26_coco_pretrained',
+    'epochs': 500,
+    'model': 'yolov8x-p26-c2f.yaml',
+    'pretrained': None,
+    'patience': 50,
+    'seed': 0,
+    'deterministic': True,
+    'data': 'coco.yaml',
+    'imgsz': 640,
+    'batch': 64,  # 16
+    'optimizer': 'SGD',
+    'close_mosaic': 10,
+    'mixup': 0.15,
+    'mosaic': 1.0,
+    'copy_paste': 0.3,
+    'scale': 0.9,
+    'cos_lr': False,
+    'lr0': 0.01,
+    'lrf': 0.01,
+    'warmup_epochs': 3.0,
+    'cache': 'disk',
+    'device': [0, 1, 2, 3],
+    'workers': 16,
+    'project': 'runs/train',
+    'val': True,
+}
+
+model = YOLO(config['model'])
+model.train(**config)

yolov8-p26-c2f.yaml

@@ -0,0 +1,66 @@
+# Ultralytics YOLO 🚀, AGPL-3.0 license
+# YOLOv8 object detection model with P3-P6 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-p6.yaml' will call yolov8-p6.yaml with scale 'n'
+  # [depth, width, max_channels]
+  n: [0.33, 0.25, 1024]
+  s: [0.33, 0.50, 1024]
+  m: [0.67, 0.75, 768]
+  l: [1.00, 1.00, 512]
+  x: [1.00, 1.25, 512]
+  # YOLOv8x-p26-c2f summary: 557 layers, 98725296 parameters, 98725280 gradients, 335.2 GFLOPs
+  #
+
+# YOLOv8.0x6 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, 3, C2f, [128, True]]
+  - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8
+  - [-1, 6, C2f, [256, True]]
+  - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16
+  - [-1, 6, C2f, [512, True]]
+  - [-1, 1, Conv, [768, 3, 2]] # 7-P5/32
+  - [-1, 3, C2f, [768, True]]
+  - [-1, 1, Conv, [1024, 3, 2]] # 9-P6/64
+  - [-1, 3, C2f, [1024, True]]
+  - [-1, 1, SPPF, [1024, 5]] # 11
+
+# YOLOv8.0x6 head
+head:
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 8], 1, Concat, [1]] # cat backbone P5
+  - [-1, 3, C2f, [768, False]] # 14
+
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 6], 1, Concat, [1]] # cat backbone P4
+  - [-1, 3, C2f, [512, False]] # 17
+
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 4], 1, Concat, [1]] # cat backbone P3
+  - [-1, 3, C2f, [256, False]] # 20
+
+  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
+  - [[-1, 2], 1, Concat, [1]] # cat backbone P2
+  - [-1, 3, C2f, [128, False]] # 23 (P2/4-xsmall)
+
+  - [-1, 1, Conv, [128, 3, 2]]
+  - [[-1, 20], 1, Concat, [1]] # cat head P3
+  - [-1, 3, C2f, [256, False]] # 26 (P3/8-small)
+
+  - [-1, 1, Conv, [256, 3, 2]]
+  - [[-1, 17], 1, Concat, [1]] # cat head P4
+  - [-1, 3, C2f, [512, False]] # 29 (P4/32-medium)
+
+  - [-1, 1, Conv, [512, 3, 2]]
+  - [[-1, 14], 1, Concat, [1]] # cat head P5
+  - [-1, 3, C2f, [768, False]] # 32 (P5/64-large)
+
+  - [-1, 1, Conv, [768, 3, 2]]
+  - [[-1, 11], 1, Concat, [1]] # cat head P6
+  - [-1, 3, C2f, [1024, False]] # 35 (P6/64-xlarge)
+
+  - [[23, 26, 29, 32, 35], 1, Detect, [nc]] # Detect(P2, P3, P4, P5, P6)