Repository navigation
How can i use deep oc sort? #2239
Unanswered
barisatayy
asked this question in
Q&A
Replies: 2 comments 1 reply
|
Now workflow such as: from boxmot import Boxmot
model = Boxmot(detector="yolov8l.pt", reid="lmbn_n_duke.pt", tracker="occluboost")
results = model.track(source=0)
for frame_result in results:
tracks = frame_result.tracks
ids = frame_result.tracks.id
confs = frame_result.tracks.conf
boxes = frame_result.tracks.xyxy
dets = frame_result.detections
embs = frame_result.embeddings
print(f"Frame {frame_result.frame_idx}: {len(ids)} tracks")
frame_result.save_csv("tracks.csv")
frame_result.save_vid("output.mp4")
if not frame_result.show():
break
frame_result.close_vid() # finalize the video fileare also available 😄 |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hi, im working with my own trained model which is yolo26x-seg. for tracking i was using default botsort in yolo. but this isnt enough for me. i need to use that deep oc sort algorithm. i cant couldnt make architecture. i need track a video and save it like use standart yolo. here is my codes:
i trained my own yolo26x-seg model with my custom data. how can i implement ocsort like this?
new_track = model_new.track(
source=problem1_new,
save=True,
project="results",
name="segmentation",
line_width=1,
tracker=new_tracker,
device="cuda",
stream=True,
iou=0.7,
conf=0.1,
half=True,
verbose=True
)
for r in new_track:
pass
All reactions