Special test with strict convergence condition

Precision: [tensor(0.5572, device='cuda:0'), tensor(0.5533, device='cuda:0'), tensor(0.5572, device='cuda:0'), tensor(0.5554, device='cuda:0'), tensor(0.5556, device='cuda:0'), tensor(0.5548, device='cuda:0'), tensor(0.5550, device='cuda:0'), tensor(0.5535, device='cuda:0'), tensor(0.5546, device='cuda:0'), tensor(0.5530, device='cuda:0')]

Output distance: [tensor(4.9630, device='cuda:0'), tensor(4.9866, device='cuda:0'), tensor(4.9630, device='cuda:0'), tensor(4.9740, device='cuda:0'), tensor(4.9724, device='cuda:0'), tensor(4.9772, device='cuda:0'), tensor(4.9761, device='cuda:0'), tensor(4.9850, device='cuda:0'), tensor(4.9787, device='cuda:0'), tensor(4.9882, device='cuda:0')]

Prediction loss: [tensor(18836044., device='cuda:0'), tensor(18437694., device='cuda:0'), tensor(17485888., device='cuda:0'), tensor(19013660., device='cuda:0'), tensor(18709642., device='cuda:0'), tensor(19804866., device='cuda:0'), tensor(17571540., device='cuda:0'), tensor(19338684., device='cuda:0'), tensor(18612306., device='cuda:0'), tensor(19705458., device='cuda:0')]

Others: [{'iter_num': 9, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(11427, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}]

Compressed training loss: [tensor(40796.3516, device='cuda:0'), tensor(40914.1602, device='cuda:0'), tensor(40791.8203, device='cuda:0'), tensor(40825.4844, device='cuda:0'), tensor(40887.8828, device='cuda:0'), tensor(40846.0977, device='cuda:0'), tensor(40856.1250, device='cuda:0'), tensor(40920.4961, device='cuda:0'), tensor(40961.8984, device='cuda:0'), tensor(40800.9102, device='cuda:0')]

Training loss: 0

Prediction time: [datetime.timedelta(seconds=1, microseconds=95356), datetime.timedelta(seconds=1, microseconds=111287), datetime.timedelta(seconds=1, microseconds=85397), datetime.timedelta(seconds=1, microseconds=107304), datetime.timedelta(seconds=1, microseconds=97345), datetime.timedelta(seconds=1, microseconds=90376), datetime.timedelta(seconds=1, microseconds=102324), datetime.timedelta(seconds=1, microseconds=93364), datetime.timedelta(seconds=1, microseconds=92367), datetime.timedelta(seconds=1, microseconds=108299)]

Phi time: [datetime.timedelta(microseconds=235999), datetime.timedelta(microseconds=245956), datetime.timedelta(microseconds=255914), datetime.timedelta(microseconds=239982), datetime.timedelta(microseconds=234008), datetime.timedelta(microseconds=256910), datetime.timedelta(microseconds=236000), datetime.timedelta(microseconds=253924), datetime.timedelta(microseconds=238987), datetime.timedelta(microseconds=254920)]

