Fix a GPU memory leak in detect. No need to calculate gradients in inference. (#900)

Co-authored-by: JKO095 <juho-pekka.koponen@wartsila.com>

detect.py

@@ -84,7 +84,8 @@ def detect(save_img=False):
 
         # Inference
         t1 = time_synchronized()
-        pred = model(img, augment=opt.augment)[0]
+        with torch.no_grad():   # Calculating gradients would cause a GPU memory leak
+            pred = model(img, augment=opt.augment)[0]
         t2 = time_synchronized()
 
         # Apply NMS