Closed Winkness closed 7 months ago
我使用test_3_dataset.py测试结果,得到如下结果: 172.0 191.0 16 15 tensor(168.5649, device='cuda:0', grad_fn=) tensor(187.3681, device='cuda:0', grad_fn=) tensor(29.5128, device='cuda:0', grad_fn=) tensor(27.5543, device='cuda:0', grad_fn=) tensor(163.6086, device='cuda:0', grad_fn=) tensor(179.3883, device='cuda:0', grad_fn=) tensor(53.1051, device='cuda:0', grad_fn=) tensor(49.2499, device='cuda:0', grad_fn=) tensor(154.1581, device='cuda:0', grad_fn=) tensor(164.2183, device='cuda:0', grad_fn=) tensor(97.9674, device='cuda:0', grad_fn=) tensor(92.4567, device='cuda:0', grad_fn=) tensor(147.2324, device='cuda:0', grad_fn=) tensor(137.6403, device='cuda:0', grad_fn=) tensor(185.2234, device='cuda:0', grad_fn=) tensor(176.8567, device='cuda:0', grad_fn=) tensor(118.0862, device='cuda:0', grad_fn=) 124.0 tensor(353.2392, device='cuda:0', grad_fn=) 248 tensor(139.3141, device='cuda:0', grad_fn=) 124.0 473 248 tensor(230.0239, device='cuda:0', grad_fn=) 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 236.5 124.0 473 248 ... 请问这种是什么原因呢?
每一帧坐标都是一样的,可能是加载了随机权重?先训练好网络再测试吧:)
了解了,感谢!
test_3_dataset.py的create_dataset里面的LaSOTDataset的json文件你有吗?
我使用test_3_dataset.py测试结果,得到如下结果: 172.0 191.0 16 15 tensor(168.5649, device='cuda:0', grad_fn=) tensor(187.3681, device='cuda:0', grad_fn=) tensor(29.5128, device='cuda:0', grad_fn=) tensor(27.5543, device='cuda:0', grad_fn=)
tensor(163.6086, device='cuda:0', grad_fn=) tensor(179.3883, device='cuda:0', grad_fn=) tensor(53.1051, device='cuda:0', grad_fn=) tensor(49.2499, device='cuda:0', grad_fn=)
tensor(154.1581, device='cuda:0', grad_fn=) tensor(164.2183, device='cuda:0', grad_fn=) tensor(97.9674, device='cuda:0', grad_fn=) tensor(92.4567, device='cuda:0', grad_fn=)
tensor(147.2324, device='cuda:0', grad_fn=) tensor(137.6403, device='cuda:0', grad_fn=) tensor(185.2234, device='cuda:0', grad_fn=) tensor(176.8567, device='cuda:0', grad_fn=)
tensor(118.0862, device='cuda:0', grad_fn=) 124.0 tensor(353.2392, device='cuda:0', grad_fn=) 248
tensor(139.3141, device='cuda:0', grad_fn=) 124.0 473 248
tensor(230.0239, device='cuda:0', grad_fn=) 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
236.5 124.0 473 248
...
请问这种是什么原因呢?