yawc

engine.py

@@ -20,15 +20,15 @@ def train_one_epoch(model, optimizer, data_loader, device, epoch, print_freq):
     if epoch == 0:
         warmup_factor = 1. / 1000
         warmup_iters = min(1000, len(data_loader) - 1)
-
         lr_scheduler = utils.warmup_lr_scheduler(optimizer, warmup_iters, warmup_factor)
 
+        
     for images, targets in metric_logger.log_every(data_loader, print_freq, header):
         images = list(image.to(device) for image in images)
         targets = [{k: v.to(device) for k, v in t.items()} for t in targets]
 
         loss_dict = model(images, targets)
-
+        print('Hey I''m here')
         losses = sum(loss for loss in loss_dict.values())
 
         # reduce losses over all GPUs for logging purposes