HCCEPose/s3_p2_train_yolo.py at main · WangYuLin-SEU/HCCEPose · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
# Author: Yulin Wang (yulinwang@seu.edu.cn)
# School of Mechanical Engineering, Southeast University, China

'''

Train YOLOv11. After training, the folder structure is:
```
demo-bin-picking
|--- models
|--- train_pbr
|--- yolo11
      |--- train_obj_s
            |--- detection
                |--- obj_s
                    |--- yolo11-detection-obj_s.pt
            |--- images
            |--- labels
            |--- yolo_configs
                |--- data_objs.yaml
            |--- autosplit_train.txt
            |--- autosplit_val.txt
```

------------------------------------------------------

训练 YOLOv11。训练完成后，文件夹结构如下：
```
demo-bin-picking
|--- models
|--- train_pbr
|--- yolo11
      |--- train_obj_s
            |--- detection
                |--- obj_s
                    |--- yolo11-detection-obj_s.pt
            |--- images
            |--- labels
            |--- yolo_configs
                |--- data_objs.yaml
            |--- autosplit_train.txt
            |--- autosplit_val.txt
```
'''

import os
if __name__ == '__main__':

    # Specify the path to the dataset folder.
    # 指定数据集文件夹的路径。
    dataset_path = 'xxx/xxx/demo-bin-picking'

    # Specify the number of GPUs and the number of training epochs.
    # For example, use 8 GPUs to train for 100 epochs.
    # 指定 GPU 的数量以及训练轮数。
    # 例如使用 8 张 GPU 进行 100 轮训练。
    gpu_num = 8
    epochs = 100

    # Train
    # 开始训练
    dataset_name = os.path.basename(dataset_path)
    task_suffix = 'detection'
    dataset_pbr_path = os.path.join(dataset_path, 'train_pbr')
    train_multi_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'yolo_train', 'train.py')
    data_objs_path = os.path.join(os.path.dirname(dataset_pbr_path), 'yolo11', 'train_obj_s', 'yolo_configs', 'data_objs.yaml')
    save_dir = os.path.join(os.path.dirname(os.path.dirname(data_objs_path)), task_suffix, f"obj_s")
    model_name = f"yolo11-{task_suffix}-obj_s.pt"
    final_model_path = os.path.join(os.path.dirname(os.path.dirname(data_objs_path)), save_dir, model_name)
    obj_s_path = os.path.dirname(final_model_path)
    batch_size = 12*gpu_num
    while 1:
        if not os.path.exists('%s'%obj_s_path):
            os.system("python %s --data_path '%s' --epochs %s --imgsz 640 --batch %s --gpu_num %s --task %s"%(train_multi_path, data_objs_path, str(epochs), str(batch_size), str(gpu_num), task_suffix))
        if os.path.exists('%s'%obj_s_path):
            break
    pass