Yolov8_커스텀 데이터 학습_2

a.py 코드 추가

경로 확인 (왼쪽 구조 보면서 경로 알맞게 코드 집어넣기)

 

내 경로=> C:\Univercity\2024\Capstone\YOLOV8_0409\ultralytics-main\ultralytics-main\ultralytics\data\a.py

from glob import glob

img_list = glob('C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\data\\images\\*.jpg')

print(len(img_list))

from sklearn.model_selection import train_test_split
train_img_list, val_img_list= train_test_split(img_list, test_size=0.2,random_state=2000)
print(len(train_img_list), len(val_img_list))

with open('C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\data\\train.txt','w') as f:
    f.write('\n'.join(train_img_list) +'\n')

with open('C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\data\\val.txt','w') as f:
    f.write('\n'.join(val_img_list) +'\n')

 

 

실행하면 train.txt ,val.txt 파일 생성됨. 안에 경로 확인해보면 본인 라벨링, 이미지 경로들이 들어가있음. 

---

data.yaml 코드 추가

경로 확인 (왼쪽 구조 보면서 경로 알맞게 코드 집어넣기)

 

내 경로 => C:\Univercity\2024\Capstone\YOLOV8_0409\ultralytics-main\ultralytics-main\ultralytics\cfg\datasets\data.yaml

# train and val data as 1) directory: path/images/, 2) file: path/images.txt, or 3) list: [path1/images/, path2/images/]
train: C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\data\\train.txt  # 118287 images
val: C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\data\\val.txt  # 5000 images

# number of classes
nc: 1

# class names
names: [ 'r' ]

 

 

 

---

 

 

학습 코드

yolo task=detect mode=train data=C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\cfg\\datasets\\data.yaml model=yolov8n.pt epochs=100 imgsz=640  

 

이제 학습이 끝나면 runs/detect 폴더 안에 best.pt 인 학습 파일이 존재.

 

 

 

확인 코드

yolo task=detect mode=train data=C:\\Univercity\\2024\\Capstone\\YOLOV8_0405\\ultralytics-main\\ultralytics-main\\ultralytics\\cfg\\datasets\\data.yaml model=yolov8n.pt epochs=100 imgsz=640