try to generate the tracking f1 video but too many points

.gitignore

@@ -0,0 +1,3 @@
+__pycache__/
+.vscode/
+cotracker/__pycache__/

cotracker/predictor.py

@@ -21,6 +21,9 @@ class CoTrackerPredictor(torch.nn.Module):
         self.model = model
         self.model.eval()
 
+        self.device = torch.device("cuda:2" if torch.cuda.is_available() else "cpu")
+        self.model.to(self.device)
+
     @torch.no_grad()
     def forward(
         self,

demo1.py

@@ -10,6 +10,8 @@ video = torch.from_numpy(video).permute(0, 3, 1, 2)[None].float()
 
 from cotracker.predictor import CoTrackerPredictor
 
+device = torch.device('cuda:2' if torch.cuda.is_available() else 'cpu')
+
 model = CoTrackerPredictor(
     checkpoint=os.path.join(
         './checkpoints/cotracker2.pth'
@@ -24,11 +26,12 @@ model = CoTrackerPredictor(
 grid_query_frame=20
 
 import torch.nn.functional as F
-video_interp = F.interpolate(video[0], [200, 360], mode="bilinear")[None]
+# video_interp = F.interpolate(video[0], [200, 360], mode="bilinear")[None].to(device)
 
 import time
 start_time = time.time()
-pred_tracks, pred_visibility = model(video_interp, 
+# pred_tracks, pred_visibility = model(video_interp, 
+pred_tracks, pred_visibility = model(video, 
                                      grid_query_frame=grid_query_frame, backward_tracking=True)
 end_time = time.time() 
 
@@ -41,6 +44,7 @@ vis = Visualizer(
     mode='optical_flow'
 )
 vis.visualize(
+    # video=video_interp,
     video=video,
     tracks=pred_tracks,
     visibility=pred_visibility,