Open 18845108835 opened 11 months ago
请补充一下错误信息。特别是尝试把number_works设置为0之后,单线程采集数据收集更干净的错误信息。
谢谢您的建议,昨天经过核实后,发现是自制数据集的问题,P2的最后一个数多了一个空格,已经解决闪退的问题啦,再次感谢您的建议~~~
现在出现了新的问题,描述如下,麻烦您啦:
Backend TkAgg is interactive backend. Turning interactive mode on.
clean up the recorder directory of C:\Users\FangGZ\Desktop\yolo3D_first\workdirs\Stereo3D\log\firstconfig=config/Stereo3D.py
-1
number of trained parameters of the model: 107466592
Found evaluate function evaluate_kitti_obj
Num training images: 2969
PSM Cos Volume takes 0.009999275207519531 seconds at call time 1
PSM Cos Volume takes 0.011998891830444336 seconds at call time 2
PSM Cos Volume takes 0.004999637603759766 seconds at call time 3
Cost Volume takes 0.006999969482421875 seconds at call time 1
PSM Cos Volume takes 0.012000083923339844 seconds at call time 4
PSM Cos Volume takes 0.006999969482421875 seconds at call time 5
Cost Volume takes 0.008991718292236328 seconds at call time 2
PSM Cos Volume takes 0.010965347290039062 seconds at call time 6
PSM Cos Volume takes 0.005000114440917969 seconds at call time 7
Cost Volume takes 0.004999876022338867 seconds at call time 3
PSM Cos Volume takes 0.015001296997070312 seconds at call time 8
PSM Cos Volume takes 0.007007122039794922 seconds at call time 9
Cost Volume takes 0.007995367050170898 seconds at call time 4
PSM Cos Volume takes 0.01299905776977539 seconds at call time 10
PSM Cos Volume takes 0.007002115249633789 seconds at call time 11
Cost Volume takes 0.010006427764892578 seconds at call time 5
PSM Cos Volume takes 0.011986494064331055 seconds at call time 12
PSM Cos Volume takes 0.008991479873657227 seconds at call time 13
Cost Volume takes 0.009996414184570312 seconds at call time 6
PSM Cos Volume takes 0.01297307014465332 seconds at call time 14
PSM Cos Volume takes 0.006989955902099609 seconds at call time 15
Cost Volume takes 0.00999903678894043 seconds at call time 7
PSM Cos Volume takes 0.013988256454467773 seconds at call time 16
PSM Cos Volume takes 0.007997989654541016 seconds at call time 17
Cost Volume takes 0.007001399993896484 seconds at call time 8
PSM Cos Volume takes 0.011969327926635742 seconds at call time 18
PSM Cos Volume takes 0.006997585296630859 seconds at call time 19
Cost Volume takes 0.009976387023925781 seconds at call time 9
/ start testing after training epoch 9 **/
clean up the recorder directory of C:\Users\FangGZ\Desktop\yolo3D_first\workdirs\Stereo3D\output\validation\data
rebuild C:\Users\FangGZ\Desktop\yolo3D_first\workdirs\Stereo3D\output\validation\data
0%| | 0/31 [00:00<?, ?it/s]C:\Users\FangGZ\Desktop\yolo3D_first\data\stereo_dataset.py:166: UserWarning: Creating a tensor from a list of numpy.ndarrays is extremely slow. Please consider converting the list to a single numpy.ndarray with numpy.array() before converting to a tensor. (Triggered internally at C:\cb\pytorch_1000000000000\work\torch\csrc\utils\tensor_new.cpp:248.)
return torch.from_numpy(left_images).float(), torch.from_numpy(right_images).float(), torch.tensor(
C:\Users\FangGZ\Desktop\yolo3D_first\networks\pipelines\testers.py:39: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requiresgrad(True), rather than torch.tensor(sourceTensor).
scores, bbox, obj_index = module([left_images.cuda().float().contiguous(), right_images.cuda().float().contiguous(), torch.tensor(P2).cuda().float(), torch.tensor(P3).cuda().float()])
C:\Users\FangGZ\Desktop\yolo3D_first\networks\lib\PSM_cost_volume.py:82: UserWarning: volatile was removed and now has no effect. Use with torch.no_grad():
instead.
cost = Variable(
C:\Users\FangGZ\Desktop\yolo3D_first\networks\lib\PSM_cost_volume.py:49: UserWarning: volatile was removed and now has no effect. Use with torch.no_grad():
instead.
cost = Variable(
100%|██████████| 31/31 [00:06<00:00, 4.47it/s]
Traceback (most recent call last):
File "C:\Program Files\JetBrains\PyCharm 2022.3.3\plugins\python\helpers\pydev\pydevconsole.py", line 364, in runcode
coro = func()
File "", line 1, in
感觉评估函数出了一些问题,正在改动,不知道您有什么建议嘞,期待您的回复>.<
作者您好,我在ue4中自制了KITTI数据集(照片像素为(1242,375,3)),但是在训练的时候在train.py文件的 for iter_num, data in enumerate(dataloader_train):闪退,但是换成kitti3D数据集,便可以正常运行,想请教您一下是什么原因?