Special test with strict convergence condition

Precision: [tensor(0.6941, device='cuda:0'), tensor(0.6915, device='cuda:0'), tensor(0.6889, device='cuda:0'), tensor(0.6857, device='cuda:0'), tensor(0.6855, device='cuda:0'), tensor(0.6876, device='cuda:0'), tensor(0.6844, device='cuda:0'), tensor(0.6918, device='cuda:0'), tensor(0.6868, device='cuda:0'), tensor(0.6894, device='cuda:0')]

Output distance: [tensor(4.9178, device='cuda:0'), tensor(4.9231, device='cuda:0'), tensor(4.9283, device='cuda:0'), tensor(4.9346, device='cuda:0'), tensor(4.9352, device='cuda:0'), tensor(4.9310, device='cuda:0'), tensor(4.9373, device='cuda:0'), tensor(4.9226, device='cuda:0'), tensor(4.9325, device='cuda:0'), tensor(4.9273, device='cuda:0')]

Prediction loss: [tensor(18386944., device='cuda:0'), tensor(19125694., device='cuda:0'), tensor(20261432., device='cuda:0'), tensor(18460390., device='cuda:0'), tensor(18219666., device='cuda:0'), tensor(17157888., device='cuda:0'), tensor(18828286., device='cuda:0'), tensor(19285148., device='cuda:0'), tensor(18333350., device='cuda:0'), tensor(19347198., device='cuda:0')]

Others: [{'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}, {'iter_num': 5, 'num_positive': tensor(3809, device='cuda:0'), 'num_positive_true': tensor(20211, device='cuda:0')}]

Compressed training loss: [tensor(40840.9844, device='cuda:0'), tensor(40858.8203, device='cuda:0'), tensor(40839.9023, device='cuda:0'), tensor(40866.9570, device='cuda:0'), tensor(40843.2539, device='cuda:0'), tensor(40792.3516, device='cuda:0'), tensor(40768.8711, device='cuda:0'), tensor(40810.6641, device='cuda:0'), tensor(40847.2383, device='cuda:0'), tensor(40967.8594, device='cuda:0')]

Training loss: 0

Prediction time: [datetime.timedelta(seconds=6, microseconds=278641), datetime.timedelta(seconds=6, microseconds=256729), datetime.timedelta(seconds=6, microseconds=265695), datetime.timedelta(seconds=6, microseconds=382206), datetime.timedelta(seconds=6, microseconds=267686), datetime.timedelta(seconds=6, microseconds=239804), datetime.timedelta(seconds=6, microseconds=463864), datetime.timedelta(seconds=6, microseconds=203954), datetime.timedelta(seconds=6, microseconds=270675), datetime.timedelta(seconds=6, microseconds=279638)]

Phi time: [datetime.timedelta(microseconds=350529), datetime.timedelta(microseconds=379411), datetime.timedelta(microseconds=353516), datetime.timedelta(microseconds=364470), datetime.timedelta(microseconds=353517), datetime.timedelta(microseconds=421231), datetime.timedelta(microseconds=350527), datetime.timedelta(microseconds=411274), datetime.timedelta(microseconds=376420), datetime.timedelta(microseconds=404301)]

