Special test with strict convergence condition

Precision: [tensor(0.7356, device='cuda:0'), tensor(0.7305, device='cuda:0'), tensor(0.7355, device='cuda:0'), tensor(0.7273, device='cuda:0'), tensor(0.7255, device='cuda:0'), tensor(0.7386, device='cuda:0'), tensor(0.7349, device='cuda:0'), tensor(0.7273, device='cuda:0'), tensor(0.7366, device='cuda:0'), tensor(0.7373, device='cuda:0')]

Output distance: [tensor(5.0095, device='cuda:0'), tensor(5.0160, device='cuda:0'), tensor(5.0092, device='cuda:0'), tensor(5.0221, device='cuda:0'), tensor(5.0205, device='cuda:0'), tensor(5.0047, device='cuda:0'), tensor(5.0108, device='cuda:0'), tensor(5.0194, device='cuda:0'), tensor(5.0071, device='cuda:0'), tensor(5.0081, device='cuda:0')]

Prediction loss: [tensor(17969938., device='cuda:0'), tensor(18209294., device='cuda:0'), tensor(18532920., device='cuda:0'), tensor(18682116., device='cuda:0'), tensor(18884456., device='cuda:0'), tensor(17609538., device='cuda:0'), tensor(18970572., device='cuda:0'), tensor(19810724., device='cuda:0'), tensor(19394674., device='cuda:0'), tensor(17726852., device='cuda:0')]

Others: [{'iter_num': 9, 'num_positive': tensor(2398, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(2397, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(2401, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 7, 'num_positive': tensor(2380, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(2412, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 7, 'num_positive': tensor(2406, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(2395, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(2402, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(2407, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(2391, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}]

Compressed training loss: [tensor(40897.8594, device='cuda:0'), tensor(40863.6250, device='cuda:0'), tensor(40869.2578, device='cuda:0'), tensor(40813.4141, device='cuda:0'), tensor(40844.7891, device='cuda:0'), tensor(40811.4023, device='cuda:0'), tensor(40852.1602, device='cuda:0'), tensor(40713.8672, device='cuda:0'), tensor(40912.2031, device='cuda:0'), tensor(40817.9648, device='cuda:0')]

Training loss: 0

Prediction time: [datetime.timedelta(seconds=1, microseconds=20671), datetime.timedelta(seconds=1, microseconds=50544), datetime.timedelta(seconds=1, microseconds=14696), datetime.timedelta(microseconds=987811), datetime.timedelta(seconds=1, microseconds=47556), datetime.timedelta(seconds=1, microseconds=25648), datetime.timedelta(seconds=1, microseconds=31625), datetime.timedelta(seconds=1, microseconds=31625), datetime.timedelta(seconds=1, microseconds=19676), datetime.timedelta(microseconds=992790)]

Phi time: [datetime.timedelta(microseconds=237991), datetime.timedelta(microseconds=254919), datetime.timedelta(microseconds=250937), datetime.timedelta(microseconds=257906), datetime.timedelta(microseconds=232017), datetime.timedelta(microseconds=250937), datetime.timedelta(microseconds=233012), datetime.timedelta(microseconds=250936), datetime.timedelta(microseconds=245957), datetime.timedelta(microseconds=250935)]

