Precision: [tensor(0.9376, device='cuda:0'), tensor(0.9289, device='cuda:0'), tensor(0.9246, device='cuda:0'), tensor(0.9522, device='cuda:0'), tensor(0.9476, device='cuda:0'), tensor(0.9115, device='cuda:0'), tensor(0.9506, device='cuda:0'), tensor(0.9538, device='cuda:0'), tensor(0.9454, device='cuda:0'), tensor(0.9441, device='cuda:0')]

Output distance: [tensor(732.4822, device='cuda:0'), tensor(1935.3638, device='cuda:0'), tensor(205.2725, device='cuda:0'), tensor(111.0255, device='cuda:0'), tensor(151.5340, device='cuda:0'), tensor(5104.8418, device='cuda:0'), tensor(139.5765, device='cuda:0'), tensor(115.0672, device='cuda:0'), tensor(149.0624, device='cuda:0'), tensor(144.4626, device='cuda:0')]

Prediction loss: [tensor(959.9467, device='cuda:0'), tensor(2860.8010, device='cuda:0'), tensor(392.9657, device='cuda:0'), tensor(374.3687, device='cuda:0'), tensor(391.5212, device='cuda:0'), tensor(4873.4927, device='cuda:0'), tensor(389.4265, device='cuda:0'), tensor(342.5225, device='cuda:0'), tensor(388.0382, device='cuda:0'), tensor(371.5194, device='cuda:0')]

Others: [{'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 30, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}]

Compressed training loss: [tensor(3522299.2500, device='cuda:0'), tensor(3691194., device='cuda:0'), tensor(3842283.7500, device='cuda:0'), tensor(3570499.5000, device='cuda:0'), tensor(3726879.7500, device='cuda:0'), tensor(3294702.7500, device='cuda:0'), tensor(3720403.7500, device='cuda:0'), tensor(3258494., device='cuda:0'), tensor(3665136.7500, device='cuda:0'), tensor(3510606., device='cuda:0')]

Training loss: 3591119.0

Prediction time: [datetime.timedelta(seconds=1, microseconds=68469), datetime.timedelta(seconds=1, microseconds=101332), datetime.timedelta(seconds=1, microseconds=93363), datetime.timedelta(seconds=1, microseconds=101330), datetime.timedelta(seconds=1, microseconds=121245), datetime.timedelta(seconds=1, microseconds=85397), datetime.timedelta(seconds=1, microseconds=91372), datetime.timedelta(seconds=1, microseconds=107306), datetime.timedelta(seconds=1, microseconds=100336), datetime.timedelta(seconds=1, microseconds=85398)]

Phi time: [datetime.timedelta(seconds=1, microseconds=254683), datetime.timedelta(microseconds=764914), datetime.timedelta(microseconds=670411), datetime.timedelta(microseconds=666707), datetime.timedelta(microseconds=668550), datetime.timedelta(microseconds=673252), datetime.timedelta(microseconds=671824), datetime.timedelta(microseconds=672082), datetime.timedelta(microseconds=671506), datetime.timedelta(microseconds=675109)]

