Skip to main content

7-11 MobileSAM Development

PANDORA DEVELOPER GUIDE · CHAPTER 07

Learning Objectives​

SAM (Segment Anything Model) is a powerful image segmentation model developed by Meta AI that supports segmentation based on various prompts such as text and coordinate points.

However, due to its large model size, it is challenging to run smoothly on the Jetson Orin platform. Therefore, a lightweight version called MobileSAM is used to achieve similar segmentation capabilities while balancing performance and efficiency.

7-11 MobileSAM Development

Initial environment setup​

sudo apt update && sudo apt install vim
pip3 install ultralytics
pip3 install numpy==1.26.4

Run MobileSAM​

1. Create an empty folder and download test images

mkdir -p MobileSAM && cd MobileSAM
wget https://github.com/ultralytics/ultralytics/blob/main/ultralytics/assets/zidane.jpg?raw=true -O zidane.jpg

2. Create a test Python file

touch demo.py && vim demo.py

3. Paste the following code

from ultralytics import SAM
# Load the model
model = SAM("mobile_sam.pt")
# Predict a segment based on a single point prompt
model.predict("zidane.jpg", points=[900, 370], labels=[1], save=Ture)

4. Run and verify the results

// The results will be automatically saved to runs/segment/predict
python3 demo.py

7-11 MobileSAM Development

Adapted from the original YUAN Developer Hub article. Source ID: 266.

On this page