glenn-jocher
commented
2 weeks ago @CurtisDisc hello,
Thank you for reaching out and providing detailed information about your issue. It looks like you're encountering a problem with maintaining the original frame rate and resolution of your video during pose estimation.
To ensure that all frames are analyzed and the video length remains unaltered, you can adjust the vid_stride parameter to 1, which you have already done. This ensures that every frame is processed. However, the warning about imgsz being updated to [1920, 1088] indicates that the resolution needs to be a multiple of the model's stride (32 in this case).
Here's a refined version of your code to help maintain the original frame rate and resolution:
from ultralytics import YOLO, checks, hub
checks()
hub.login('9e04cb143f43ef901c5251f59fe7bce026278b0054')
model = YOLO('yolov8l-pose.pt')
source = 'https://youtu.be/Ku5sXKIrwL8?si=PTIyvOeCNqg6uh_Z'
results = model.predict(source, stream=True, imgsz=(1920, 1088),
vid_stride=1, max_det=10, save=True, show_boxes=False,
show_labels=False, half=True)
for result in results:
boxes = result.boxes # Boxes object for bounding box outputs
masks = result.masks # Masks object for segmentation masks outputs
keypoints = result.keypoints # Keypoints object for pose outputs
probs = result.probs # Probs object for classification outputs
obb = result.obb # Oriented boxes object for OBB outputsKey Points:
- Resolution Adjustment: The warning indicates that the resolution must be a multiple of the model's stride. Adjusting
imgszto(1920, 1088)ensures compatibility. - Frame Rate: By setting
vid_stride=1, you ensure that every frame is processed, maintaining the original frame rate.
If you continue to experience issues, please ensure you are using the latest version of the Ultralytics packages. You can update them using:
pip install --upgrade ultralyticsFeel free to reach out if you have any further questions or run into any other issues. We're here to help! 😊
CurtisDisc
Search before asking
Question
I am running pose estimation of a video and wanting to have the same resolution and frame rate is the saved result video. How do I get it to analysis all frames so that video length is not altered.
from ultralytics import YOLO, checks, hub checks()
hub.login('9e04cb143f43ef901c5251f59fe7bce026278b0054')
model = YOLO('yolov8l-pose.pt')
source='https://youtu.be/Ku5sXKIrwL8?si=PTIyvOeCNqg6uh_Z'
results = model.predict(source, stream=True, imgsz=(1920,1080), vid_stride=1, max_det=10, save=True, show_boxes=False, show_labels=False, half=True,)
for result in results: boxes = result.boxes # Boxes object for bounding box outputs masks = result.masks # Masks object for segmentation masks outputs keypoints = result.keypoints # Keypoints object for pose outputs probs = result.probs # Probs object for classification outputs obb = result.obb # Oriented boxes object for OBB outputs
Ultralytics YOLOv8.2.81 🚀 Python-3.10.12 torch-2.3.1+cu121 CUDA:0 (Tesla T4, 15102MiB) Setup complete ✅ (2 CPUs, 12.7 GB RAM, 33.9/78.2 GB disk) Ultralytics HUB: New authentication successful ✅
WARNING ⚠️ imgsz=[1920, 1080] must be multiple of max stride 32, updating to [1920, 1088] 1/1: https://youtu.be/Ku5sXKIrwL8?si=PTIyvOeCNqg6uh_Z... Success ✅ (3012 frames of shape 1920x1080 at 60.00 FPS)
Additional
No response