run once to remove initial timing effects

detect.py

@@ -75,6 +75,7 @@ def detect(save_img=False):
 
     # Run inference
     t0 = time.time()
+    _ = model(torch.zeros((1, 3, img_size, img_size), device=device)) if device.type != 'cpu' else None  # run once
     for path, img, im0s, vid_cap in dataset:
         img = torch.from_numpy(img).to(device)
         img = img.half() if half else img.float()  # uint8 to fp16/32

test.py

@@ -70,6 +70,7 @@ def test(cfg,
 
     seen = 0
     model.eval()
+    _ = model(torch.zeros((1, 3, img_size, img_size), device=device)) if device.type != 'cpu' else None  # run once
     coco91class = coco80_to_coco91_class()
     s = ('%20s' + '%10s' * 6) % ('Class', 'Images', 'Targets', 'P', 'R', 'mAP@0.5', 'F1')
     p, r, f1, mp, mr, map, mf1, t0, t1 = 0., 0., 0., 0., 0., 0., 0., 0., 0.