Precision: [tensor(0.8218, device='cuda:0'), tensor(0.8236, device='cuda:0'), tensor(0.8239, device='cuda:0'), tensor(0.8250, device='cuda:0'), tensor(0.8243, device='cuda:0'), tensor(0.8238, device='cuda:0'), tensor(0.8248, device='cuda:0'), tensor(0.8215, device='cuda:0'), tensor(0.8242, device='cuda:0'), tensor(0.8246, device='cuda:0')]

Output distance: [tensor(13890.3828, device='cuda:0'), tensor(13826.0449, device='cuda:0'), tensor(13822.6318, device='cuda:0'), tensor(13766.9355, device='cuda:0'), tensor(13701.7236, device='cuda:0'), tensor(13772.4775, device='cuda:0'), tensor(13754.6514, device='cuda:0'), tensor(14003.2256, device='cuda:0'), tensor(13737.0322, device='cuda:0'), tensor(13674.0117, device='cuda:0')]

Prediction loss: [tensor(10504.6719, device='cuda:0'), tensor(10672.2842, device='cuda:0'), tensor(10605.8994, device='cuda:0'), tensor(10576.1787, device='cuda:0'), tensor(10640.8877, device='cuda:0'), tensor(10265.2529, device='cuda:0'), tensor(10404.7871, device='cuda:0'), tensor(10532.7441, device='cuda:0'), tensor(10611.5127, device='cuda:0'), tensor(10459.5508, device='cuda:0')]

Others: [{'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 17, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 17, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 17, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}, {'iter_num': 15, 'num_positive': tensor(18000, device='cuda:0'), 'num_positive_true': tensor(18000, device='cuda:0')}]

Compressed training loss: [tensor(1.9155e+08, device='cuda:0'), tensor(1.9366e+08, device='cuda:0'), tensor(1.9356e+08, device='cuda:0'), tensor(1.9389e+08, device='cuda:0'), tensor(1.9317e+08, device='cuda:0'), tensor(1.8745e+08, device='cuda:0'), tensor(1.9083e+08, device='cuda:0'), tensor(1.9239e+08, device='cuda:0'), tensor(1.9267e+08, device='cuda:0'), tensor(1.8958e+08, device='cuda:0')]

Training loss: 191994016.0

Prediction time: [datetime.timedelta(microseconds=680117), datetime.timedelta(microseconds=708995), datetime.timedelta(microseconds=711978), datetime.timedelta(microseconds=748825), datetime.timedelta(microseconds=712979), datetime.timedelta(microseconds=742850), datetime.timedelta(microseconds=760769), datetime.timedelta(microseconds=711979), datetime.timedelta(microseconds=718951), datetime.timedelta(microseconds=719946)]

Phi time: [datetime.timedelta(seconds=1, microseconds=184899), datetime.timedelta(microseconds=704233), datetime.timedelta(microseconds=665170), datetime.timedelta(microseconds=644684), datetime.timedelta(microseconds=648833), datetime.timedelta(microseconds=653570), datetime.timedelta(microseconds=647545), datetime.timedelta(microseconds=647657), datetime.timedelta(microseconds=645866), datetime.timedelta(microseconds=649634)]

