Skip to content

Conversation

@cerebrai
Copy link
Contributor

Description

The current implementation supports sending one image at a time. In some applications, sending a batch of images through the model and leveraging the GPU capabilities to speed up the processing would be useful. I have made slight modifications to the "forward" function in "factory.py" in the "models" folder. Now it accepts a single image or a list of images and do batch processing.

Tests

I tested it using the following code and it checks out
import cv2
import onepose

if __name__ == "__main__":
   
    model = onepose.create_model("ViTPose+_base_coco_wholebody")
    img1 = cv2.imread('/path/to/image1')
    img2 = cv2.imread('/path/to/image2')
    
    batch_img = [img1, img2]
    keypoints = model(img1)
    keypoints_batch = model(batch_img)

@developer0hye
Copy link
Owner

developer0hye commented Nov 24, 2024

Hello @cerebrai,

Great suggestion, I will check it.

@developer0hye developer0hye merged commit 7f78d0f into developer0hye:main Nov 24, 2024
2 checks passed
@developer0hye
Copy link
Owner

@cerebrai It is merged! Thanks for your contribution!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants