Yolov8 in Azure Machine learning

Yolo V8 in Azure Machine Learning

Pre-requisites

• Azure Machine Learning Service
• Azure Storage Account

Code

• Create a new Azure Machine Learning Service Workspace
• Create a notebook
• Pick the kernel as python with Tensorflow and Pytorch
• Now clone the repo from github
• Change conda environment to azureml_py38_TF_PY

source active azureml_py38_TF_PY

• clone the repo

git clone https://github.com/ultralytics/ultralytics.git

• Open Terminal and run the following command.
• cd to the ultralytics folder
• cd into yolov8

pip install -r requirements.txt

• Now install Ultralytics Yolov8

pip install ultralytics

• Test in terminal

yolo task=detect mode=predict model=yolov8n.pt source="bus.jpg"

• output

yolo task=detect mode=predict model=yolov8n.pt source="bus.jpg"
Ultralytics YOLOv8.0.5 🚀 Python-3.8.5 torch-1.12.1 CPU
Fusing layers...
YOLOv8n summary: 168 layers, 3151904 parameters, 0 gradients, 8.7 GFLOPs
image 1/1 /mnt/batch/tasks/shared/LS_root/mounts/clusters/devbox1/code/Users/babal/yolov8/ultralytics/bus.jpg: 640x480 4 persons, 1 bus, 1 stop sign, 752.6ms
Speed: 1.3ms pre-process, 752.6ms inference, 301.7ms postprocess per image at shape (1, 3, 640, 640)

• Create a notebook with kernel as python with Tensorflow and Pytorch

# Load YOLOv8n, train it on COCO128 for 3 epochs and predict an image with it
from ultralytics import YOLO

model = YOLO('yolov8n.pt')  # load a pretrained YOLOv8n detection model
model.train(data='coco128.yaml', epochs=3)  # train the model
model('https://ultralytics.com/images/bus.jpg')  # predict on an image

• Run cell
• output

Found https://ultralytics.com/images/bus.jpg locally at bus.jpg
Ultralytics YOLOv8.0.5 🚀 Python-3.8.5 torch-1.12.1 CPU
Fusing layers... 
Model summary: 168 layers, 3151904 parameters, 0 gradients, 8.7 GFLOPs
image 1/1 /mnt/batch/tasks/shared/LS_root/mounts/clusters/devbox1/code/Users/babal/yolov8/ultralytics/bus.jpg: 640x480 4 persons, 1 bus, 1 stop sign, 51.3ms
Speed: 2.4ms pre-process, 51.3ms inference, 0.9ms postprocess per image at shape (1, 3, 640, 640)

• Now segmentation

# Load YOLOv8n-seg, train it on COCO128-seg for 3 epochs and predict an image with it
from ultralytics import YOLO

model = YOLO('yolov8n-seg.pt')  # load a pretrained YOLOv8n segmentation model
model.train(data='coco128-seg.yaml', epochs=3)  # train the model
model('https://ultralytics.com/images/bus.jpg')  # predict on an image

• Output

Found https://ultralytics.com/images/bus.jpg locally at bus.jpg
Ultralytics YOLOv8.0.5 🚀 Python-3.8.5 torch-1.12.1 CPU
Fusing layers... 
YOLOv8n-seg summary: 195 layers, 3404320 parameters, 0 gradients, 12.6 GFLOPs
image 1/1 /mnt/batch/tasks/shared/LS_root/mounts/clusters/devbox1/code/Users/babal/yolov8/ultralytics/bus.jpg: 640x480 4 persons, 1 bus, 1 skateboard, 75.6ms
Speed: 0.6ms pre-process, 75.6ms inference, 2.4ms postprocess per image at shape (1, 3, 640, 640)

Originally published at https://github.com.

Mlearning.ai Submission Suggestions