Search code examples
pythonflaskyoloyolov8

YOLOv8 custom save directory path

I'm currently working in a project in which I'm using Flask and Yolov8 together.

When I run this code

from ultralytics import YOLO

model = YOLO("./yolov8n.pt")

results = model.predict(source="../TEST/doggy.jpg", save=True, save_txt=True)

the output will be saved in this default directory /run/detect/

like

Ultralytics YOLOv8.0.9  Python-3.10.8 torch-2.0.0+cpu CPU
Fusing layers... 
YOLOv8n summary: 168 layers, 3151904 parameters, 0 gradients, 8.7 GFLOPs
Results saved to d:\runs\detect\predict4
1 labels saved to d:\runs\detect\predict4\labels

and what I want is the predict directory number or the entire directory path in a variable.

I tried capturing the path using sys.stdout methods but i want a direct solution.

Solution

  • You can change the directory where the results are saved (save_dir) by modifying two arguments in predict: project and name

    results = model.predict(source=xxx, save_txt = True, project="xxx", name="yyy")

    such that:

    save_dir=project/name