Loading pretrained DanceTrack model from pth file
#1
by
irith-k - opened
If I wanted to use one of the CenterNet model weights in a Python script, say as such:
from dancetrack.model import DanceTrack # ??
model = DanceTrack() # ??
model.load_state_dict(torch.load('dancetrack_bbox.pth', weights_only=True))
model.eval()
for image in video: bbox_preds = model(image) ...
What model architecture/class should I use? Where can I import the appropriate model class that is used for DanceTrack? Specifically, what would the first two lines above look like? It wasn't clear from the GitHub repository how I could implement this.
Thank you for your help!