Special test with strict convergence condition

Precision: [tensor(0.6246, device='cuda:0'), tensor(0.6302, device='cuda:0'), tensor(0.6283, device='cuda:0'), tensor(0.6289, device='cuda:0'), tensor(0.6301, device='cuda:0'), tensor(0.6275, device='cuda:0'), tensor(0.6302, device='cuda:0'), tensor(0.6306, device='cuda:0'), tensor(0.6271, device='cuda:0'), tensor(0.6275, device='cuda:0')]

Output distance: [tensor(4.9365, device='cuda:0'), tensor(4.9168, device='cuda:0'), tensor(4.9244, device='cuda:0'), tensor(4.9247, device='cuda:0'), tensor(4.9202, device='cuda:0'), tensor(4.9281, device='cuda:0'), tensor(4.9170, device='cuda:0'), tensor(4.9170, device='cuda:0'), tensor(4.9289, device='cuda:0'), tensor(4.9273, device='cuda:0')]

Prediction loss: [tensor(18499502., device='cuda:0'), tensor(18551236., device='cuda:0'), tensor(17938154., device='cuda:0'), tensor(18169528., device='cuda:0'), tensor(18080116., device='cuda:0'), tensor(18441152., device='cuda:0'), tensor(19530384., device='cuda:0'), tensor(17703790., device='cuda:0'), tensor(17450528., device='cuda:0'), tensor(19179570., device='cuda:0')]

Others: [{'iter_num': 9, 'num_positive': tensor(5652, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(5697, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(5666, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(5637, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(5650, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(5646, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(5692, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(5672, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 11, 'num_positive': tensor(5653, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 9, 'num_positive': tensor(5657, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}]

Compressed training loss: [tensor(40857.1445, device='cuda:0'), tensor(40847.0703, device='cuda:0'), tensor(40705.9336, device='cuda:0'), tensor(40914.1719, device='cuda:0'), tensor(40822.0781, device='cuda:0'), tensor(40790.4102, device='cuda:0'), tensor(40910.7891, device='cuda:0'), tensor(40874.8008, device='cuda:0'), tensor(40790.3828, device='cuda:0'), tensor(40862.0547, device='cuda:0')]

Training loss: 0

Prediction time: [datetime.timedelta(seconds=1, microseconds=83405), datetime.timedelta(seconds=1, microseconds=97346), datetime.timedelta(seconds=1, microseconds=66477), datetime.timedelta(seconds=1, microseconds=67473), datetime.timedelta(seconds=1, microseconds=53532), datetime.timedelta(seconds=1, microseconds=70460), datetime.timedelta(seconds=1, microseconds=49548), datetime.timedelta(seconds=1, microseconds=69463), datetime.timedelta(seconds=1, microseconds=90375), datetime.timedelta(seconds=1, microseconds=65482)]

Phi time: [datetime.timedelta(microseconds=236995), datetime.timedelta(microseconds=248944), datetime.timedelta(microseconds=235004), datetime.timedelta(microseconds=255915), datetime.timedelta(microseconds=237990), datetime.timedelta(microseconds=238987), datetime.timedelta(microseconds=236995), datetime.timedelta(microseconds=238987), datetime.timedelta(microseconds=235003), datetime.timedelta(microseconds=252928)]

