물체 κ°μ§
κ°μ²΄ κ°μ§λ μ΄λ―Έμ§ λλ λμμ μ€νΈλ¦Όμμ κ°μ²΄μ μμΉμ ν΄λμ€λ₯Ό μλ³νλ μμ μ λλ€.
κ°μ²΄ κ°μ§κΈ°μ μΆλ ₯μ μ΄λ―Έμ§μμ κ°μ²΄λ₯Ό λλ¬μΈλ κ²½κ³ μμ μ§ν©κ³Ό κ° μμμ λν ν΄λμ€ λ μ΄λΈ λ° μ λ’°λ μ μλ‘ κ΅¬μ±λ©λλ€. κ°μ²΄ κ°μ§λ μ₯λ©΄μμ κ΄μ¬ μλ κ°μ²΄λ₯Ό μλ³ν΄μΌ νμ§λ§ κ°μ²΄μ μ νν μμΉλ λͺ¨μμ μ νν μ νμλ μλ κ²½μ°μ μ ν©ν μ νμ λλ€.
Watch: μ¬μ νμ΅λ Ultralytics YOLOv8 λͺ¨λΈμ μ¬μ©ν κ°μ²΄ κ°μ§.
ν
YOLOv8 κ°μ§ λͺ¨λΈμ κΈ°λ³Έκ°μΈ YOLOv8 λͺ¨λΈμ
λλ€. yolov8n.pt
μ λν΄ μ¬μ κ΅μ‘μ λ°μμΌλ©° COCO.
λͺ¨λΈ
YOLOv8 μ¬μ νμ΅λ κ°μ§ λͺ¨λΈμ΄ μ¬κΈ°μ λμ μμ΅λλ€. κ°μ§, μΈκ·Έλ¨ΌνΈ λ° ν¬μ¦ λͺ¨λΈμ COCO λ°μ΄ν° μΈνΈμμ μ¬μ νμ΅λ λͺ¨λΈμ΄λ©°, λΆλ₯ λͺ¨λΈμ ImageNet λ°μ΄ν° μΈνΈμμ μ¬μ νμ΅λ λͺ¨λΈμ λλ€.
λͺ¨λΈμ μ²μ μ¬μ©ν λ μ΅μ Ultralytics 릴리μ€μμ μλμΌλ‘ λ€μ΄λ‘λλ©λλ€.
λͺ¨λΈ | ν¬κΈ° (ν½μ ) |
mAPval 50-95 |
μλ CPU ONNX (ms) |
μλ A100 TensorRT (ms) |
맀κ°λ³μ (M) |
FLOPs (B) |
---|---|---|---|---|---|---|
YOLOv8n | 640 | 37.3 | 80.4 | 0.99 | 3.2 | 8.7 |
YOLOv8s | 640 | 44.9 | 128.4 | 1.20 | 11.2 | 28.6 |
YOLOv8m | 640 | 50.2 | 234.7 | 1.83 | 25.9 | 78.9 |
YOLOv8l | 640 | 52.9 | 375.2 | 2.39 | 43.7 | 165.2 |
YOLOv8x | 640 | 53.9 | 479.1 | 3.53 | 68.2 | 257.8 |
- mAPval κ°μ λ¨μΌ λͺ¨λΈ λ¨μΌ μ€μΌμΌμ λν κ²μ
λλ€. COCO val2017 λ°μ΄ν° μΈνΈ.
볡μ λμyolo val detect data=coco.yaml device=0
- μλ λ₯Ό μ¬μ©νμ¬ COCO κ° μ΄λ―Έμ§μ λν νκ· μ ꡬν©λλ€. Amazon EC2 P4d μΈμ€ν΄μ€.
볡μ λμyolo val detect data=coco8.yaml batch=1 device=0|cpu
κΈ°μ°¨
YOLOv8n μ μ΄λ―Έμ§ ν¬κΈ° 640μΌλ‘ 100κ°μ μν¬ν¬μ λν΄ COCO8 λ°μ΄ν°μ μΌλ‘ νλ ¨ν©λλ€. μ¬μ© κ°λ₯ν μΈμμ μ 체 λͺ©λ‘μ κ΅¬μ± νμ΄μ§λ₯Ό μ°Έμ‘°νμΈμ.
μ
from ultralytics import YOLO
# Load a model
model = YOLO('yolov8n.yaml') # build a new model from YAML
model = YOLO('yolov8n.pt') # load a pretrained model (recommended for training)
model = YOLO('yolov8n.yaml').load('yolov8n.pt') # build from YAML and transfer weights
# Train the model
results = model.train(data='coco8.yaml', epochs=100, imgsz=640)
# Build a new model from YAML and start training from scratch
yolo detect train data=coco8.yaml model=yolov8n.yaml epochs=100 imgsz=640
# Start training from a pretrained *.pt model
yolo detect train data=coco8.yaml model=yolov8n.pt epochs=100 imgsz=640
# Build a new model from YAML, transfer pretrained weights to it and start training
yolo detect train data=coco8.yaml model=yolov8n.yaml pretrained=yolov8n.pt epochs=100 imgsz=640
λ°μ΄ν° μ§ν© νμ
YOLO νμ§ λ°μ΄ν° μΈνΈ νμμ λ°μ΄ν° μΈνΈ κ°μ΄λμμ μμΈν νμΈν μ μμ΅λλ€. κΈ°μ‘΄ λ°μ΄ν°μ μ λ€λ₯Έ νμ(μ: COCO λ±)μμ YOLO νμμΌλ‘ λ³ννλ €λ©΄ JSON2YOLO λꡬ( Ultralytics)λ₯Ό μ¬μ©νμΈμ.
Val
COCO8 λ°μ΄ν° μΈνΈμμ νμ΅λ YOLOv8n λͺ¨λΈ μ νλλ₯Ό κ²μ¦ν©λλ€. μΈμλ₯Ό μ λ¬ν νμκ° μμ΅λλ€. model
κ΅μ‘ μ μ§ data
λ° μΈμλ₯Ό λͺ¨λΈ μμ±μΌλ‘ μ¬μ©ν©λλ€.
μ
from ultralytics import YOLO
# Load a model
model = YOLO('yolov8n.pt') # load an official model
model = YOLO('path/to/best.pt') # load a custom model
# Validate the model
metrics = model.val() # no arguments needed, dataset and settings remembered
metrics.box.map # map50-95
metrics.box.map50 # map50
metrics.box.map75 # map75
metrics.box.maps # a list contains map50-95 of each category
μμΈ‘
νμ΅λ YOLOv8n λͺ¨λΈμ μ¬μ©νμ¬ μ΄λ―Έμ§μ λν μμΈ‘μ μ€νν©λλ€.
μ
μ 체 보기 predict
λͺ¨λ μΈλΆ μ 보μμ μμΈ‘ νμ΄μ§λ‘ μ΄λν©λλ€.
λ΄λ³΄λ΄κΈ°
YOLOv8n λͺ¨λΈμ ONNX, CoreML λ±κ³Ό κ°μ λ€λ₯Έ νμμΌλ‘ λ΄λ³΄λ λλ€.
μ
μ¬μ© κ°λ₯ν YOLOv8 λ΄λ³΄λ΄κΈ° νμμ μλ νμ λμ μμ΅λλ€. λ΄λ³΄λ΄λ νμμ format
μΈμ, μ¦ format='onnx'
λλ format='engine'
. λ΄λ³΄λΈ λͺ¨λΈμμ μ§μ μμΈ‘νκ±°λ κ²μ¦ν μ μμ΅λλ€. yolo predict model=yolov8n.onnx
. λ΄λ³΄λ΄κΈ°κ° μλ£λ ν λͺ¨λΈμ λν μ¬μ© μκ° νμλ©λλ€.
νμ | format μΈμ |
λͺ¨λΈ | λ©νλ°μ΄ν° | μΈμ |
---|---|---|---|---|
PyTorch | - | yolov8n.pt |
β | - |
TorchScript | torchscript |
yolov8n.torchscript |
β | imgsz , optimize , batch |
ONNX | onnx |
yolov8n.onnx |
β | imgsz , half , dynamic , simplify , opset , batch |
OpenVINO | openvino |
yolov8n_openvino_model/ |
β | imgsz , half , int8 , batch |
TensorRT | engine |
yolov8n.engine |
β | imgsz , half , dynamic , simplify , workspace , batch |
CoreML | coreml |
yolov8n.mlpackage |
β | imgsz , half , int8 , nms , batch |
TF SavedModel | saved_model |
yolov8n_saved_model/ |
β | imgsz , keras , int8 , batch |
TF GraphDef | pb |
yolov8n.pb |
β | imgsz , batch |
TF Lite | tflite |
yolov8n.tflite |
β | imgsz , half , int8 , batch |
TF Edge TPU | edgetpu |
yolov8n_edgetpu.tflite |
β | imgsz , batch |
TF.js | tfjs |
yolov8n_web_model/ |
β | imgsz , half , int8 , batch |
PaddlePaddle | paddle |
yolov8n_paddle_model/ |
β | imgsz , batch |
NCNN | ncnn |
yolov8n_ncnn_model/ |
β | imgsz , half , batch |
μ 체 보기 export
μΈλΆ μ 보μμ λ΄λ³΄λ΄κΈ° νμ΄μ§λ‘ μ΄λν©λλ€.
μμ± 2023-11-12, μ λ°μ΄νΈ 2024-04-27
μμ±μ: glenn-jocher (14), Burhan-Q (1), Laughing-q (1), AyushExel (1)