accuracy for nw-ucla data is below 97.2%

[yanghhx]

I have run your code according your settings, but I have got the accuracy for NW-UCLA 96.12%.

[MartinXM]

Hi, can you share your training log and other configurations? What the batchsize and gpu number do you use for training?

[yanghhx]

The configurations are set the same as your code in your Github, the batchsize and the test batchsize are both 64. The GPU number is 2. The attachments are logs. At 2023-04-06 17:38:07, "Martin" ***@***.***> wrote: Hi, can you share your training log and other configurations? What the batchsize and gpu number do you use for training? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***> [ Fri Apr 14 14:28:13 2023 ] using warm up, epoch: 5 [ Fri Apr 14 14:28:18 2023 ] Parameters: {'work_dir': 'work_dir/ucla/lst_bone', 'model_saved_name': 'work_dir/ucla/lst_bone/runs', 'config': 'config/ucla/lst_bone.yaml', 'phase': 'train', 'save_score': False, 'seed': 1, 'log_interval': 100, 'save_interval': 1, 'save_epoch': 30, 'eval_interval': 5, 'print_log': True, 'show_topk': [1, 5], 'feeder': 'feeders.feeder_ucla.Feeder', 'num_worker': 32, 'train_feeder_args': {'data_path': 'joint', 'label_path': 'train', 'debug': False, 'random_choose': True, 'random_shift': False, 'random_move': False, 'window_size': 52, 'normalization': False, 'repeat': 5}, 'test_feeder_args': {'data_path': 'joint', 'label_path': 'val', 'debug': False}, 'model': 'model.ctrgcn.Model_lst_4part_ucla', 'model_args': {'num_class': 10, 'num_point': 20, 'num_person': 1, 'graph': 'graph.ucla.Graph', 'graph_args': {'labeling_mode': 'spatial'}, 'head': ['ViT-B/32'], 'k': 1}, 'weights': None, 'ignore_weights': [], 'base_lr': 0.2, 'step': [90, 100], 'device': 0, 'optimizer': 'SGD', 'nesterov': True, 'batch_size': 64, 'test_batch_size': 64, 'start_epoch': 0, 'num_epoch': 110, 'weight_decay': 0.0004, 'lr_decay_rate': 0.1, 'warm_up_epoch': 5, 'loss_alpha': 1.0, 'te_lr_ratio': 1} [ Fri Apr 14 14:28:18 2023 ] # Parameters: 2084817 [ Fri Apr 14 14:28:18 2023 ] Training epoch: 1 [ Fri Apr 14 14:29:13 2023 ] using warm up, epoch: 5 [ Fri Apr 14 14:29:19 2023 ] Parameters: {'work_dir': 'work_dir/ucla/lst_bone', 'model_saved_name': 'work_dir/ucla/lst_bone/runs', 'config': 'config/ucla/lst_bone.yaml', 'phase': 'train', 'save_score': False, 'seed': 1, 'log_interval': 100, 'save_interval': 1, 'save_epoch': 30, 'eval_interval': 5, 'print_log': True, 'show_topk': [1, 5], 'feeder': 'feeders.feeder_ucla.Feeder', 'num_worker': 32, 'train_feeder_args': {'data_path': 'joint', 'label_path': 'train', 'debug': False, 'random_choose': True, 'random_shift': False, 'random_move': False, 'window_size': 52, 'normalization': False, 'repeat': 5}, 'test_feeder_args': {'data_path': 'joint', 'label_path': 'val', 'debug': False}, 'model': 'model.ctrgcn.Model_lst_4part_ucla', 'model_args': {'num_class': 10, 'num_point': 20, 'num_person': 1, 'graph': 'graph.ucla.Graph', 'graph_args': {'labeling_mode': 'spatial'}, 'head': ['ViT-B/32'], 'k': 1}, 'weights': None, 'ignore_weights': [], 'base_lr': 0.2, 'step': [90, 100], 'device': 0, 'optimizer': 'SGD', 'nesterov': True, 'batch_size': 64, 'test_batch_size': 64, 'start_epoch': 0, 'num_epoch': 110, 'weight_decay': 0.0004, 'lr_decay_rate': 0.1, 'warm_up_epoch': 5, 'loss_alpha': 1.0, 'te_lr_ratio': 1} [ Fri Apr 14 14:29:19 2023 ] # Parameters: 2084817 [ Fri Apr 14 14:29:19 2023 ] Training epoch: 1 [ Fri Apr 14 14:31:00 2023 ] using warm up, epoch: 5 [ Fri Apr 14 14:31:04 2023 ] Parameters: {'work_dir': 'work_dir/ucla/lst_bone', 'model_saved_name': 'work_dir/ucla/lst_bone/runs', 'config': 'config/ucla/lst_bone.yaml', 'phase': 'train', 'save_score': False, 'seed': 1, 'log_interval': 100, 'save_interval': 1, 'save_epoch': 30, 'eval_interval': 5, 'print_log': True, 'show_topk': [1, 5], 'feeder': 'feeders.feeder_ucla.Feeder', 'num_worker': 32, 'train_feeder_args': {'data_path': 'joint', 'label_path': 'train', 'debug': False, 'random_choose': True, 'random_shift': False, 'random_move': False, 'window_size': 52, 'normalization': False, 'repeat': 5}, 'test_feeder_args': {'data_path': 'joint', 'label_path': 'val', 'debug': False}, 'model': 'model.ctrgcn.Model_lst_4part_ucla', 'model_args': {'num_class': 10, 'num_point': 20, 'num_person': 1, 'graph': 'graph.ucla.Graph', 'graph_args': {'labeling_mode': 'spatial'}, 'head': ['ViT-B/32'], 'k': 1}, 'weights': None, 'ignore_weights': [], 'base_lr': 0.2, 'step': [90, 100], 'device': 0, 'optimizer': 'SGD', 'nesterov': True, 'batch_size': 64, 'test_batch_size': 64, 'start_epoch': 0, 'num_epoch': 110, 'weight_decay': 0.0004, 'lr_decay_rate': 0.1, 'warm_up_epoch': 5, 'loss_alpha': 1.0, 'te_lr_ratio': 1} [ Fri Apr 14 14:31:04 2023 ] # Parameters: 2084817 [ Fri Apr 14 14:31:04 2023 ] Training epoch: 1 [ Fri Apr 14 14:33:01 2023 ] Mean training loss: 4.1335. Mean training acc: 49.80%. [ Fri Apr 14 14:33:01 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 14:33:01 2023 ] Eval epoch: 1 [ Fri Apr 14 14:33:09 2023 ] Mean test loss of 8 batches: 0.8324917778372765. [ Fri Apr 14 14:33:09 2023 ] Top1: 67.24% [ Fri Apr 14 14:33:09 2023 ] Top5: 99.14% [ Fri Apr 14 14:33:09 2023 ] Training epoch: 2 [ Fri Apr 14 14:35:04 2023 ] Mean training loss: 2.1379. Mean training acc: 68.77%. [ Fri Apr 14 14:35:04 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 14:35:04 2023 ] Eval epoch: 2 [ Fri Apr 14 14:35:11 2023 ] Mean test loss of 8 batches: 0.6256089396774769. [ Fri Apr 14 14:35:11 2023 ] Top1: 78.88% [ Fri Apr 14 14:35:11 2023 ] Top5: 99.14% [ Fri Apr 14 14:35:11 2023 ] Training epoch: 3 [ Fri Apr 14 14:37:07 2023 ] Mean training loss: 1.6068. Mean training acc: 72.41%. [ Fri Apr 14 14:37:07 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 14:37:07 2023 ] Eval epoch: 3 [ Fri Apr 14 14:37:14 2023 ] Mean test loss of 8 batches: 0.7112999930977821. [ Fri Apr 14 14:37:14 2023 ] Top1: 71.55% [ Fri Apr 14 14:37:14 2023 ] Top5: 99.78% [ Fri Apr 14 14:37:14 2023 ] Training epoch: 4 [ Fri Apr 14 14:39:07 2023 ] Mean training loss: 1.6813. Mean training acc: 75.40%. [ Fri Apr 14 14:39:07 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 14:39:07 2023 ] Eval epoch: 4 [ Fri Apr 14 14:39:14 2023 ] Mean test loss of 8 batches: 0.6535387821495533. [ Fri Apr 14 14:39:14 2023 ] Top1: 72.41% [ Fri Apr 14 14:39:14 2023 ] Top5: 99.57% [ Fri Apr 14 14:39:14 2023 ] Training epoch: 5 [ Fri Apr 14 14:41:03 2023 ] Mean training loss: 1.2787. Mean training acc: 78.66%. [ Fri Apr 14 14:41:03 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 14:41:03 2023 ] Eval epoch: 5 [ Fri Apr 14 14:41:10 2023 ] Mean test loss of 8 batches: 0.6742357425391674. [ Fri Apr 14 14:41:10 2023 ] Top1: 72.84% [ Fri Apr 14 14:41:10 2023 ] Top5: 99.14% [ Fri Apr 14 14:41:10 2023 ] Training epoch: 6 [ Fri Apr 14 14:43:07 2023 ] Mean training loss: 1.0441. Mean training acc: 82.34%. [ Fri Apr 14 14:43:07 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 14:43:07 2023 ] Eval epoch: 6 [ Fri Apr 14 14:43:14 2023 ] Mean test loss of 8 batches: 0.6236459240317345. [ Fri Apr 14 14:43:14 2023 ] Top1: 79.09% [ Fri Apr 14 14:43:14 2023 ] Top5: 97.84% [ Fri Apr 14 14:43:14 2023 ] Training epoch: 7 [ Fri Apr 14 14:45:09 2023 ] Mean training loss: 0.9222. Mean training acc: 84.75%. [ Fri Apr 14 14:45:09 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 14:45:09 2023 ] Eval epoch: 7 [ Fri Apr 14 14:45:17 2023 ] Mean test loss of 8 batches: 0.47836102545261383. [ Fri Apr 14 14:45:17 2023 ] Top1: 82.54% [ Fri Apr 14 14:45:17 2023 ] Top5: 98.92% [ Fri Apr 14 14:45:17 2023 ] Training epoch: 8 [ Fri Apr 14 14:47:05 2023 ] Mean training loss: 0.8507. Mean training acc: 86.51%. [ Fri Apr 14 14:47:05 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 14:47:05 2023 ] Eval epoch: 8 [ Fri Apr 14 14:47:14 2023 ] Mean test loss of 8 batches: 0.6567780002951622. [ Fri Apr 14 14:47:14 2023 ] Top1: 73.71% [ Fri Apr 14 14:47:14 2023 ] Top5: 99.57% [ Fri Apr 14 14:47:14 2023 ] Training epoch: 9 [ Fri Apr 14 14:49:06 2023 ] Mean training loss: 0.7795. Mean training acc: 87.16%. [ Fri Apr 14 14:49:06 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 14:49:06 2023 ] Eval epoch: 9 [ Fri Apr 14 14:49:14 2023 ] Mean test loss of 8 batches: 0.4856749903410673. [ Fri Apr 14 14:49:14 2023 ] Top1: 82.97% [ Fri Apr 14 14:49:14 2023 ] Top5: 98.71% [ Fri Apr 14 14:49:14 2023 ] Training epoch: 10 [ Fri Apr 14 14:51:11 2023 ] Mean training loss: 0.6753. Mean training acc: 88.79%. [ Fri Apr 14 14:51:11 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 14:51:11 2023 ] Eval epoch: 10 [ Fri Apr 14 14:51:19 2023 ] Mean test loss of 8 batches: 0.3985324651002884. [ Fri Apr 14 14:51:19 2023 ] Top1: 85.56% [ Fri Apr 14 14:51:19 2023 ] Top5: 99.57% [ Fri Apr 14 14:51:19 2023 ] Training epoch: 11 [ Fri Apr 14 14:53:16 2023 ] Mean training loss: 0.6805. Mean training acc: 88.84%. [ Fri Apr 14 14:53:16 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 14:53:16 2023 ] Eval epoch: 11 [ Fri Apr 14 14:53:23 2023 ] Mean test loss of 8 batches: 1.6088696867227554. [ Fri Apr 14 14:53:23 2023 ] Top1: 62.50% [ Fri Apr 14 14:53:23 2023 ] Top5: 90.09% [ Fri Apr 14 14:53:23 2023 ] Training epoch: 12 [ Fri Apr 14 14:55:16 2023 ] Mean training loss: 0.6907. Mean training acc: 88.83%. [ Fri Apr 14 14:55:16 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 14:55:16 2023 ] Eval epoch: 12 [ Fri Apr 14 14:55:24 2023 ] Mean test loss of 8 batches: 0.5189987495541573. [ Fri Apr 14 14:55:24 2023 ] Top1: 82.76% [ Fri Apr 14 14:55:24 2023 ] Top5: 99.78% [ Fri Apr 14 14:55:24 2023 ] Training epoch: 13 [ Fri Apr 14 14:57:16 2023 ] Mean training loss: 0.5772. Mean training acc: 90.90%. [ Fri Apr 14 14:57:16 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 14:57:16 2023 ] Eval epoch: 13 [ Fri Apr 14 14:57:24 2023 ] Mean test loss of 8 batches: 0.35401466209441423. [ Fri Apr 14 14:57:24 2023 ] Top1: 86.85% [ Fri Apr 14 14:57:24 2023 ] Top5: 99.57% [ Fri Apr 14 14:57:24 2023 ] Training epoch: 14 [ Fri Apr 14 14:59:19 2023 ] Mean training loss: 0.5476. Mean training acc: 91.52%. [ Fri Apr 14 14:59:19 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 14:59:19 2023 ] Eval epoch: 14 [ Fri Apr 14 14:59:26 2023 ] Mean test loss of 8 batches: 0.4853922240436077. [ Fri Apr 14 14:59:26 2023 ] Top1: 84.91% [ Fri Apr 14 14:59:26 2023 ] Top5: 99.35% [ Fri Apr 14 14:59:26 2023 ] Training epoch: 15 [ Fri Apr 14 15:01:22 2023 ] Mean training loss: 0.5007. Mean training acc: 91.75%. [ Fri Apr 14 15:01:22 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:01:22 2023 ] Eval epoch: 15 [ Fri Apr 14 15:01:29 2023 ] Mean test loss of 8 batches: 0.35810780338943005. [ Fri Apr 14 15:01:29 2023 ] Top1: 90.09% [ Fri Apr 14 15:01:29 2023 ] Top5: 98.71% [ Fri Apr 14 15:01:29 2023 ] Training epoch: 16 [ Fri Apr 14 15:03:22 2023 ] Mean training loss: 0.5049. Mean training acc: 92.15%. [ Fri Apr 14 15:03:22 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:03:22 2023 ] Eval epoch: 16 [ Fri Apr 14 15:03:30 2023 ] Mean test loss of 8 batches: 0.6098749041557312. [ Fri Apr 14 15:03:30 2023 ] Top1: 81.03% [ Fri Apr 14 15:03:30 2023 ] Top5: 99.35% [ Fri Apr 14 15:03:30 2023 ] Training epoch: 17 [ Fri Apr 14 15:05:22 2023 ] Mean training loss: 0.4314. Mean training acc: 93.47%. [ Fri Apr 14 15:05:22 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:05:22 2023 ] Eval epoch: 17 [ Fri Apr 14 15:05:29 2023 ] Mean test loss of 8 batches: 0.8456620313227177. [ Fri Apr 14 15:05:29 2023 ] Top1: 64.87% [ Fri Apr 14 15:05:29 2023 ] Top5: 98.06% [ Fri Apr 14 15:05:29 2023 ] Training epoch: 18 [ Fri Apr 14 15:07:24 2023 ] Mean training loss: 0.4161. Mean training acc: 93.83%. [ Fri Apr 14 15:07:24 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:07:24 2023 ] Eval epoch: 18 [ Fri Apr 14 15:07:31 2023 ] Mean test loss of 8 batches: 0.42832107841968536. [ Fri Apr 14 15:07:31 2023 ] Top1: 84.27% [ Fri Apr 14 15:07:31 2023 ] Top5: 99.57% [ Fri Apr 14 15:07:31 2023 ] Training epoch: 19 [ Fri Apr 14 15:09:28 2023 ] Mean training loss: 0.4713. Mean training acc: 92.27%. [ Fri Apr 14 15:09:28 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:09:28 2023 ] Eval epoch: 19 [ Fri Apr 14 15:09:35 2023 ] Mean test loss of 8 batches: 0.483685702085495. [ Fri Apr 14 15:09:35 2023 ] Top1: 84.27% [ Fri Apr 14 15:09:35 2023 ] Top5: 99.57% [ Fri Apr 14 15:09:35 2023 ] Training epoch: 20 [ Fri Apr 14 15:11:29 2023 ] Mean training loss: 0.4161. Mean training acc: 93.95%. [ Fri Apr 14 15:11:29 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:11:29 2023 ] Eval epoch: 20 [ Fri Apr 14 15:11:36 2023 ] Mean test loss of 8 batches: 0.4027852453291416. [ Fri Apr 14 15:11:36 2023 ] Top1: 87.72% [ Fri Apr 14 15:11:36 2023 ] Top5: 99.78% [ Fri Apr 14 15:11:36 2023 ] Training epoch: 21 [ Fri Apr 14 15:13:28 2023 ] Mean training loss: 0.3797. Mean training acc: 94.22%. [ Fri Apr 14 15:13:28 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 15:13:28 2023 ] Eval epoch: 21 [ Fri Apr 14 15:13:35 2023 ] Mean test loss of 8 batches: 0.4636695086956024. [ Fri Apr 14 15:13:35 2023 ] Top1: 85.56% [ Fri Apr 14 15:13:35 2023 ] Top5: 99.78% [ Fri Apr 14 15:13:35 2023 ] Training epoch: 22 [ Fri Apr 14 15:15:30 2023 ] Mean training loss: 0.5362. Mean training acc: 91.79%. [ Fri Apr 14 15:15:30 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:15:30 2023 ] Eval epoch: 22 [ Fri Apr 14 15:15:37 2023 ] Mean test loss of 8 batches: 0.7674081437289715. [ Fri Apr 14 15:15:37 2023 ] Top1: 70.91% [ Fri Apr 14 15:15:37 2023 ] Top5: 99.35% [ Fri Apr 14 15:15:37 2023 ] Training epoch: 23 [ Fri Apr 14 15:17:33 2023 ] Mean training loss: 0.4081. Mean training acc: 93.75%. [ Fri Apr 14 15:17:33 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:17:33 2023 ] Eval epoch: 23 [ Fri Apr 14 15:17:40 2023 ] Mean test loss of 8 batches: 0.9160239174962044. [ Fri Apr 14 15:17:40 2023 ] Top1: 74.35% [ Fri Apr 14 15:17:40 2023 ] Top5: 98.28% [ Fri Apr 14 15:17:40 2023 ] Training epoch: 24 [ Fri Apr 14 15:19:38 2023 ] Mean training loss: 0.4006. Mean training acc: 94.11%. [ Fri Apr 14 15:19:38 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:19:38 2023 ] Eval epoch: 24 [ Fri Apr 14 15:19:46 2023 ] Mean test loss of 8 batches: 0.2703039990738034. [ Fri Apr 14 15:19:46 2023 ] Top1: 91.16% [ Fri Apr 14 15:19:46 2023 ] Top5: 98.92% [ Fri Apr 14 15:19:46 2023 ] Training epoch: 25 [ Fri Apr 14 15:21:40 2023 ] Mean training loss: 0.4054. Mean training acc: 93.95%. [ Fri Apr 14 15:21:40 2023 ] Time consumption: [Data]09%, [Network]90% [ Fri Apr 14 15:21:40 2023 ] Eval epoch: 25 [ Fri Apr 14 15:21:49 2023 ] Mean test loss of 8 batches: 0.5331466756761074. [ Fri Apr 14 15:21:49 2023 ] Top1: 83.84% [ Fri Apr 14 15:21:49 2023 ] Top5: 98.92% [ Fri Apr 14 15:21:49 2023 ] Training epoch: 26 [ Fri Apr 14 15:23:47 2023 ] Mean training loss: 0.3672. Mean training acc: 94.68%. [ Fri Apr 14 15:23:47 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:23:47 2023 ] Eval epoch: 26 [ Fri Apr 14 15:23:54 2023 ] Mean test loss of 8 batches: 0.4169522598385811. [ Fri Apr 14 15:23:54 2023 ] Top1: 85.99% [ Fri Apr 14 15:23:54 2023 ] Top5: 98.71% [ Fri Apr 14 15:23:54 2023 ] Training epoch: 27 [ Fri Apr 14 15:25:52 2023 ] Mean training loss: 0.3739. Mean training acc: 94.24%. [ Fri Apr 14 15:25:52 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:25:52 2023 ] Eval epoch: 27 [ Fri Apr 14 15:25:59 2023 ] Mean test loss of 8 batches: 0.3448748290538788. [ Fri Apr 14 15:25:59 2023 ] Top1: 87.72% [ Fri Apr 14 15:25:59 2023 ] Top5: 99.14% [ Fri Apr 14 15:25:59 2023 ] Training epoch: 28 [ Fri Apr 14 15:27:53 2023 ] Mean training loss: 0.3220. Mean training acc: 95.49%. [ Fri Apr 14 15:27:53 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:27:53 2023 ] Eval epoch: 28 [ Fri Apr 14 15:27:59 2023 ] Mean test loss of 8 batches: 0.389761483296752. [ Fri Apr 14 15:27:59 2023 ] Top1: 86.21% [ Fri Apr 14 15:27:59 2023 ] Top5: 99.14% [ Fri Apr 14 15:27:59 2023 ] Training epoch: 29 [ Fri Apr 14 15:29:52 2023 ] Mean training loss: 0.3265. Mean training acc: 95.49%. [ Fri Apr 14 15:29:52 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 15:29:52 2023 ] Eval epoch: 29 [ Fri Apr 14 15:29:58 2023 ] Mean test loss of 8 batches: 0.6759087610989809. [ Fri Apr 14 15:29:58 2023 ] Top1: 77.37% [ Fri Apr 14 15:29:58 2023 ] Top5: 99.14% [ Fri Apr 14 15:29:58 2023 ] Training epoch: 30 [ Fri Apr 14 15:31:54 2023 ] Mean training loss: 0.4943. Mean training acc: 95.31%. [ Fri Apr 14 15:31:54 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:31:54 2023 ] Eval epoch: 30 [ Fri Apr 14 15:32:01 2023 ] Mean test loss of 8 batches: 0.4952161554247141. [ Fri Apr 14 15:32:01 2023 ] Top1: 84.91% [ Fri Apr 14 15:32:01 2023 ] Top5: 98.92% [ Fri Apr 14 15:32:01 2023 ] Training epoch: 31 [ Fri Apr 14 15:33:57 2023 ] Mean training loss: 0.3665. Mean training acc: 94.98%. [ Fri Apr 14 15:33:57 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:33:57 2023 ] Eval epoch: 31 [ Fri Apr 14 15:34:03 2023 ] Mean test loss of 8 batches: 0.318013628013432. [ Fri Apr 14 15:34:03 2023 ] Top1: 88.58% [ Fri Apr 14 15:34:03 2023 ] Top5: 100.00% [ Fri Apr 14 15:34:03 2023 ] Training epoch: 32 [ Fri Apr 14 15:35:58 2023 ] Mean training loss: 0.2939. Mean training acc: 96.04%. [ Fri Apr 14 15:35:58 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 15:35:58 2023 ] Eval epoch: 32 [ Fri Apr 14 15:36:06 2023 ] Mean test loss of 8 batches: 0.35687373392283916. [ Fri Apr 14 15:36:06 2023 ] Top1: 87.50% [ Fri Apr 14 15:36:06 2023 ] Top5: 99.57% [ Fri Apr 14 15:36:06 2023 ] Training epoch: 33 [ Fri Apr 14 15:37:59 2023 ] Mean training loss: 0.3754. Mean training acc: 94.26%. [ Fri Apr 14 15:37:59 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 15:37:59 2023 ] Eval epoch: 33 [ Fri Apr 14 15:38:06 2023 ] Mean test loss of 8 batches: 0.4109928719699383. [ Fri Apr 14 15:38:06 2023 ] Top1: 86.85% [ Fri Apr 14 15:38:06 2023 ] Top5: 99.35% [ Fri Apr 14 15:38:06 2023 ] Training epoch: 34 [ Fri Apr 14 15:40:02 2023 ] Mean training loss: 0.3953. Mean training acc: 93.67%. [ Fri Apr 14 15:40:02 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 15:40:02 2023 ] Eval epoch: 34 [ Fri Apr 14 15:40:09 2023 ] Mean test loss of 8 batches: 0.2770312773063779. [ Fri Apr 14 15:40:09 2023 ] Top1: 90.73% [ Fri Apr 14 15:40:09 2023 ] Top5: 99.78% [ Fri Apr 14 15:40:09 2023 ] Training epoch: 35 [ Fri Apr 14 15:42:05 2023 ] Mean training loss: 0.3199. Mean training acc: 95.67%. [ Fri Apr 14 15:42:05 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:42:06 2023 ] Eval epoch: 35 [ Fri Apr 14 15:42:12 2023 ] Mean test loss of 8 batches: 0.5161713464185596. [ Fri Apr 14 15:42:12 2023 ] Top1: 83.19% [ Fri Apr 14 15:42:12 2023 ] Top5: 99.78% [ Fri Apr 14 15:42:12 2023 ] Training epoch: 36 [ Fri Apr 14 15:44:08 2023 ] Mean training loss: 0.3374. Mean training acc: 95.41%. [ Fri Apr 14 15:44:08 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:44:08 2023 ] Eval epoch: 36 [ Fri Apr 14 15:44:15 2023 ] Mean test loss of 8 batches: 0.2840273641049862. [ Fri Apr 14 15:44:15 2023 ] Top1: 89.22% [ Fri Apr 14 15:44:15 2023 ] Top5: 99.57% [ Fri Apr 14 15:44:15 2023 ] Training epoch: 37 [ Fri Apr 14 15:46:09 2023 ] Mean training loss: 0.2906. Mean training acc: 95.77%. [ Fri Apr 14 15:46:09 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:46:10 2023 ] Eval epoch: 37 [ Fri Apr 14 15:46:17 2023 ] Mean test loss of 8 batches: 0.42925952398218215. [ Fri Apr 14 15:46:17 2023 ] Top1: 85.13% [ Fri Apr 14 15:46:17 2023 ] Top5: 99.78% [ Fri Apr 14 15:46:17 2023 ] Training epoch: 38 [ Fri Apr 14 15:48:16 2023 ] Mean training loss: 0.2785. Mean training acc: 96.00%. [ Fri Apr 14 15:48:16 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 15:48:16 2023 ] Eval epoch: 38 [ Fri Apr 14 15:48:23 2023 ] Mean test loss of 8 batches: 0.48768186569213867. [ Fri Apr 14 15:48:23 2023 ] Top1: 84.91% [ Fri Apr 14 15:48:23 2023 ] Top5: 100.00% [ Fri Apr 14 15:48:23 2023 ] Training epoch: 39 [ Fri Apr 14 15:50:22 2023 ] Mean training loss: 0.2739. Mean training acc: 96.46%. [ Fri Apr 14 15:50:22 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:50:22 2023 ] Eval epoch: 39 [ Fri Apr 14 15:50:28 2023 ] Mean test loss of 8 batches: 0.4094662172719836. [ Fri Apr 14 15:50:28 2023 ] Top1: 85.13% [ Fri Apr 14 15:50:29 2023 ] Top5: 99.78% [ Fri Apr 14 15:50:29 2023 ] Training epoch: 40 [ Fri Apr 14 15:52:21 2023 ] Mean training loss: 0.3015. Mean training acc: 96.02%. [ Fri Apr 14 15:52:21 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 15:52:21 2023 ] Eval epoch: 40 [ Fri Apr 14 15:52:28 2023 ] Mean test loss of 8 batches: 0.31678924337029457. [ Fri Apr 14 15:52:28 2023 ] Top1: 90.52% [ Fri Apr 14 15:52:28 2023 ] Top5: 99.57% [ Fri Apr 14 15:52:28 2023 ] Training epoch: 41 [ Fri Apr 14 15:54:20 2023 ] Mean training loss: 0.2930. Mean training acc: 96.20%. [ Fri Apr 14 15:54:20 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 15:54:20 2023 ] Eval epoch: 41 [ Fri Apr 14 15:54:27 2023 ] Mean test loss of 8 batches: 0.45982869155704975. [ Fri Apr 14 15:54:27 2023 ] Top1: 84.05% [ Fri Apr 14 15:54:27 2023 ] Top5: 98.92% [ Fri Apr 14 15:54:27 2023 ] Training epoch: 42 [ Fri Apr 14 15:56:23 2023 ] Mean training loss: 0.2529. Mean training acc: 96.62%. [ Fri Apr 14 15:56:23 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 15:56:23 2023 ] Eval epoch: 42 [ Fri Apr 14 15:56:30 2023 ] Mean test loss of 8 batches: 0.926629900932312. [ Fri Apr 14 15:56:30 2023 ] Top1: 74.57% [ Fri Apr 14 15:56:30 2023 ] Top5: 98.92% [ Fri Apr 14 15:56:30 2023 ] Training epoch: 43 [ Fri Apr 14 15:58:27 2023 ] Mean training loss: 0.2623. Mean training acc: 96.70%. [ Fri Apr 14 15:58:27 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 15:58:27 2023 ] Eval epoch: 43 [ Fri Apr 14 15:58:34 2023 ] Mean test loss of 8 batches: 0.8122075647115707. [ Fri Apr 14 15:58:34 2023 ] Top1: 71.34% [ Fri Apr 14 15:58:34 2023 ] Top5: 99.14% [ Fri Apr 14 15:58:34 2023 ] Training epoch: 44 [ Fri Apr 14 16:00:26 2023 ] Mean training loss: 0.3460. Mean training acc: 95.73%. [ Fri Apr 14 16:00:26 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 16:00:26 2023 ] Eval epoch: 44 [ Fri Apr 14 16:00:33 2023 ] Mean test loss of 8 batches: 0.5207112981006503. [ Fri Apr 14 16:00:33 2023 ] Top1: 82.33% [ Fri Apr 14 16:00:33 2023 ] Top5: 98.49% [ Fri Apr 14 16:00:33 2023 ] Training epoch: 45 [ Fri Apr 14 16:02:26 2023 ] Mean training loss: 0.3003. Mean training acc: 96.28%. [ Fri Apr 14 16:02:26 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 16:02:26 2023 ] Eval epoch: 45 [ Fri Apr 14 16:02:34 2023 ] Mean test loss of 8 batches: 0.3768977830186486. [ Fri Apr 14 16:02:34 2023 ] Top1: 85.78% [ Fri Apr 14 16:02:34 2023 ] Top5: 99.35% [ Fri Apr 14 16:02:34 2023 ] Training epoch: 46 [ Fri Apr 14 16:04:32 2023 ] Mean training loss: 0.2279. Mean training acc: 97.15%. [ Fri Apr 14 16:04:32 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:04:32 2023 ] Eval epoch: 46 [ Fri Apr 14 16:04:38 2023 ] Mean test loss of 8 batches: 0.6003256477415562. [ Fri Apr 14 16:04:38 2023 ] Top1: 81.47% [ Fri Apr 14 16:04:38 2023 ] Top5: 99.57% [ Fri Apr 14 16:04:38 2023 ] Training epoch: 47 [ Fri Apr 14 16:06:34 2023 ] Mean training loss: 0.3075. Mean training acc: 95.85%. [ Fri Apr 14 16:06:34 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:06:34 2023 ] Eval epoch: 47 [ Fri Apr 14 16:06:39 2023 ] Mean test loss of 8 batches: 0.5130227495683357. [ Fri Apr 14 16:06:39 2023 ] Top1: 83.62% [ Fri Apr 14 16:06:39 2023 ] Top5: 99.78% [ Fri Apr 14 16:06:39 2023 ] Training epoch: 48 [ Fri Apr 14 16:08:32 2023 ] Mean training loss: 0.2577. Mean training acc: 96.56%. [ Fri Apr 14 16:08:32 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:08:32 2023 ] Eval epoch: 48 [ Fri Apr 14 16:08:38 2023 ] Mean test loss of 8 batches: 0.3509411420673132. [ Fri Apr 14 16:08:38 2023 ] Top1: 90.09% [ Fri Apr 14 16:08:38 2023 ] Top5: 99.14% [ Fri Apr 14 16:08:38 2023 ] Training epoch: 49 [ Fri Apr 14 16:10:32 2023 ] Mean training loss: 0.2826. Mean training acc: 95.97%. [ Fri Apr 14 16:10:32 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:10:32 2023 ] Eval epoch: 49 [ Fri Apr 14 16:10:38 2023 ] Mean test loss of 8 batches: 0.5099343257024884. [ Fri Apr 14 16:10:38 2023 ] Top1: 81.25% [ Fri Apr 14 16:10:38 2023 ] Top5: 98.71% [ Fri Apr 14 16:10:38 2023 ] Training epoch: 50 [ Fri Apr 14 16:12:36 2023 ] Mean training loss: 0.2365. Mean training acc: 97.07%. [ Fri Apr 14 16:12:36 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:12:36 2023 ] Eval epoch: 50 [ Fri Apr 14 16:12:42 2023 ] Mean test loss of 8 batches: 0.7676894329488277. [ Fri Apr 14 16:12:42 2023 ] Top1: 79.53% [ Fri Apr 14 16:12:42 2023 ] Top5: 99.14% [ Fri Apr 14 16:12:42 2023 ] Training epoch: 51 [ Fri Apr 14 16:14:38 2023 ] Mean training loss: 0.3256. Mean training acc: 95.37%. [ Fri Apr 14 16:14:38 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:14:38 2023 ] Eval epoch: 51 [ Fri Apr 14 16:14:43 2023 ] Mean test loss of 8 batches: 0.9849775768816471. [ Fri Apr 14 16:14:44 2023 ] Top1: 69.18% [ Fri Apr 14 16:14:44 2023 ] Top5: 98.92% [ Fri Apr 14 16:14:44 2023 ] Training epoch: 52 [ Fri Apr 14 16:16:35 2023 ] Mean training loss: 0.2408. Mean training acc: 96.88%. [ Fri Apr 14 16:16:35 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:16:35 2023 ] Eval epoch: 52 [ Fri Apr 14 16:16:41 2023 ] Mean test loss of 8 batches: 0.3778057973831892. [ Fri Apr 14 16:16:41 2023 ] Top1: 89.66% [ Fri Apr 14 16:16:41 2023 ] Top5: 99.35% [ Fri Apr 14 16:16:41 2023 ] Training epoch: 53 [ Fri Apr 14 16:18:35 2023 ] Mean training loss: 0.2612. Mean training acc: 96.70%. [ Fri Apr 14 16:18:35 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:18:35 2023 ] Eval epoch: 53 [ Fri Apr 14 16:18:40 2023 ] Mean test loss of 8 batches: 0.2884333999827504. [ Fri Apr 14 16:18:40 2023 ] Top1: 90.73% [ Fri Apr 14 16:18:40 2023 ] Top5: 100.00% [ Fri Apr 14 16:18:40 2023 ] Training epoch: 54 [ Fri Apr 14 16:20:37 2023 ] Mean training loss: 0.2181. Mean training acc: 97.45%. [ Fri Apr 14 16:20:37 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:20:37 2023 ] Eval epoch: 54 [ Fri Apr 14 16:20:44 2023 ] Mean test loss of 8 batches: 0.4845354678109288. [ Fri Apr 14 16:20:44 2023 ] Top1: 84.70% [ Fri Apr 14 16:20:44 2023 ] Top5: 99.57% [ Fri Apr 14 16:20:44 2023 ] Training epoch: 55 [ Fri Apr 14 16:22:43 2023 ] Mean training loss: 0.2406. Mean training acc: 96.93%. [ Fri Apr 14 16:22:43 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 16:22:43 2023 ] Eval epoch: 55 [ Fri Apr 14 16:22:49 2023 ] Mean test loss of 8 batches: 0.4304638300091028. [ Fri Apr 14 16:22:49 2023 ] Top1: 88.15% [ Fri Apr 14 16:22:49 2023 ] Top5: 98.28% [ Fri Apr 14 16:22:49 2023 ] Training epoch: 56 [ Fri Apr 14 16:24:43 2023 ] Mean training loss: 0.2889. Mean training acc: 96.12%. [ Fri Apr 14 16:24:43 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:24:43 2023 ] Eval epoch: 56 [ Fri Apr 14 16:24:48 2023 ] Mean test loss of 8 batches: 0.5165979489684105. [ Fri Apr 14 16:24:48 2023 ] Top1: 81.68% [ Fri Apr 14 16:24:48 2023 ] Top5: 99.14% [ Fri Apr 14 16:24:48 2023 ] Training epoch: 57 [ Fri Apr 14 16:26:40 2023 ] Mean training loss: 0.2218. Mean training acc: 97.21%. [ Fri Apr 14 16:26:40 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:26:41 2023 ] Eval epoch: 57 [ Fri Apr 14 16:26:46 2023 ] Mean test loss of 8 batches: 0.5670568151399493. [ Fri Apr 14 16:26:46 2023 ] Top1: 83.62% [ Fri Apr 14 16:26:46 2023 ] Top5: 99.14% [ Fri Apr 14 16:26:46 2023 ] Training epoch: 58 [ Fri Apr 14 16:28:44 2023 ] Mean training loss: 0.2547. Mean training acc: 96.76%. [ Fri Apr 14 16:28:44 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:28:44 2023 ] Eval epoch: 58 [ Fri Apr 14 16:28:50 2023 ] Mean test loss of 8 batches: 0.3453879076987505. [ Fri Apr 14 16:28:50 2023 ] Top1: 89.66% [ Fri Apr 14 16:28:50 2023 ] Top5: 98.49% [ Fri Apr 14 16:28:50 2023 ] Training epoch: 59 [ Fri Apr 14 16:30:47 2023 ] Mean training loss: 0.2661. Mean training acc: 96.34%. [ Fri Apr 14 16:30:47 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:30:47 2023 ] Eval epoch: 59 [ Fri Apr 14 16:30:53 2023 ] Mean test loss of 8 batches: 0.46271123364567757. [ Fri Apr 14 16:30:53 2023 ] Top1: 84.70% [ Fri Apr 14 16:30:53 2023 ] Top5: 98.28% [ Fri Apr 14 16:30:53 2023 ] Training epoch: 60 [ Fri Apr 14 16:32:48 2023 ] Mean training loss: 0.2651. Mean training acc: 96.40%. [ Fri Apr 14 16:32:48 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:32:48 2023 ] Eval epoch: 60 [ Fri Apr 14 16:32:54 2023 ] Mean test loss of 8 batches: 0.5672187507152557. [ Fri Apr 14 16:32:54 2023 ] Top1: 85.13% [ Fri Apr 14 16:32:54 2023 ] Top5: 98.49% [ Fri Apr 14 16:32:54 2023 ] Training epoch: 61 [ Fri Apr 14 16:34:47 2023 ] Mean training loss: 0.2331. Mean training acc: 97.15%. [ Fri Apr 14 16:34:47 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:34:47 2023 ] Eval epoch: 61 [ Fri Apr 14 16:34:53 2023 ] Mean test loss of 8 batches: 0.3904128894209862. [ Fri Apr 14 16:34:53 2023 ] Top1: 87.93% [ Fri Apr 14 16:34:53 2023 ] Top5: 100.00% [ Fri Apr 14 16:34:53 2023 ] Training epoch: 62 [ Fri Apr 14 16:36:50 2023 ] Mean training loss: 0.2204. Mean training acc: 97.29%. [ Fri Apr 14 16:36:50 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:36:50 2023 ] Eval epoch: 62 [ Fri Apr 14 16:36:56 2023 ] Mean test loss of 8 batches: 0.6052924543619156. [ Fri Apr 14 16:36:56 2023 ] Top1: 82.97% [ Fri Apr 14 16:36:56 2023 ] Top5: 98.49% [ Fri Apr 14 16:36:56 2023 ] Training epoch: 63 [ Fri Apr 14 16:38:49 2023 ] Mean training loss: 0.2256. Mean training acc: 97.43%. [ Fri Apr 14 16:38:49 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:38:49 2023 ] Eval epoch: 63 [ Fri Apr 14 16:38:54 2023 ] Mean test loss of 8 batches: 0.42666138149797916. [ Fri Apr 14 16:38:54 2023 ] Top1: 88.15% [ Fri Apr 14 16:38:54 2023 ] Top5: 99.57% [ Fri Apr 14 16:38:54 2023 ] Training epoch: 64 [ Fri Apr 14 16:40:52 2023 ] Mean training loss: 0.2490. Mean training acc: 96.93%. [ Fri Apr 14 16:40:52 2023 ] Time consumption: [Data]05%, [Network]95% [ Fri Apr 14 16:40:52 2023 ] Eval epoch: 64 [ Fri Apr 14 16:40:57 2023 ] Mean test loss of 8 batches: 0.2828368656337261. [ Fri Apr 14 16:40:57 2023 ] Top1: 90.95% [ Fri Apr 14 16:40:57 2023 ] Top5: 99.57% [ Fri Apr 14 16:40:57 2023 ] Training epoch: 65 [ Fri Apr 14 16:42:49 2023 ] Mean training loss: 0.2658. Mean training acc: 96.60%. [ Fri Apr 14 16:42:49 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:42:49 2023 ] Eval epoch: 65 [ Fri Apr 14 16:42:55 2023 ] Mean test loss of 8 batches: 0.4150556782260537. [ Fri Apr 14 16:42:55 2023 ] Top1: 87.07% [ Fri Apr 14 16:42:55 2023 ] Top5: 99.35% [ Fri Apr 14 16:42:55 2023 ] Training epoch: 66 [ Fri Apr 14 16:44:55 2023 ] Mean training loss: 0.2204. Mean training acc: 97.27%. [ Fri Apr 14 16:44:55 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 16:44:55 2023 ] Eval epoch: 66 [ Fri Apr 14 16:45:03 2023 ] Mean test loss of 8 batches: 0.4066222351975739. [ Fri Apr 14 16:45:03 2023 ] Top1: 88.36% [ Fri Apr 14 16:45:03 2023 ] Top5: 99.78% [ Fri Apr 14 16:45:03 2023 ] Training epoch: 67 [ Fri Apr 14 16:46:58 2023 ] Mean training loss: 0.2407. Mean training acc: 96.70%. [ Fri Apr 14 16:46:58 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:46:58 2023 ] Eval epoch: 67 [ Fri Apr 14 16:47:04 2023 ] Mean test loss of 8 batches: 0.40801322646439075. [ Fri Apr 14 16:47:04 2023 ] Top1: 87.93% [ Fri Apr 14 16:47:04 2023 ] Top5: 100.00% [ Fri Apr 14 16:47:04 2023 ] Training epoch: 68 [ Fri Apr 14 16:49:01 2023 ] Mean training loss: 0.2155. Mean training acc: 97.27%. [ Fri Apr 14 16:49:01 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:49:01 2023 ] Eval epoch: 68 [ Fri Apr 14 16:49:06 2023 ] Mean test loss of 8 batches: 0.3511704676784575. [ Fri Apr 14 16:49:06 2023 ] Top1: 88.58% [ Fri Apr 14 16:49:06 2023 ] Top5: 99.35% [ Fri Apr 14 16:49:06 2023 ] Training epoch: 69 [ Fri Apr 14 16:51:00 2023 ] Mean training loss: 0.2381. Mean training acc: 96.89%. [ Fri Apr 14 16:51:00 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:51:00 2023 ] Eval epoch: 69 [ Fri Apr 14 16:51:06 2023 ] Mean test loss of 8 batches: 0.7570857424288988. [ Fri Apr 14 16:51:06 2023 ] Top1: 79.09% [ Fri Apr 14 16:51:06 2023 ] Top5: 98.49% [ Fri Apr 14 16:51:06 2023 ] Training epoch: 70 [ Fri Apr 14 16:53:05 2023 ] Mean training loss: 0.2534. Mean training acc: 97.17%. [ Fri Apr 14 16:53:05 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 16:53:05 2023 ] Eval epoch: 70 [ Fri Apr 14 16:53:13 2023 ] Mean test loss of 8 batches: 1.1075509525835514. [ Fri Apr 14 16:53:13 2023 ] Top1: 74.57% [ Fri Apr 14 16:53:13 2023 ] Top5: 98.71% [ Fri Apr 14 16:53:13 2023 ] Training epoch: 71 [ Fri Apr 14 16:55:09 2023 ] Mean training loss: 0.5741. Mean training acc: 97.88%. [ Fri Apr 14 16:55:09 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 16:55:09 2023 ] Eval epoch: 71 [ Fri Apr 14 16:55:16 2023 ] Mean test loss of 8 batches: 0.5100827999413013. [ Fri Apr 14 16:55:16 2023 ] Top1: 82.33% [ Fri Apr 14 16:55:16 2023 ] Top5: 99.57% [ Fri Apr 14 16:55:16 2023 ] Training epoch: 72 [ Fri Apr 14 16:57:12 2023 ] Mean training loss: 0.2948. Mean training acc: 96.62%. [ Fri Apr 14 16:57:12 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 16:57:12 2023 ] Eval epoch: 72 [ Fri Apr 14 16:57:18 2023 ] Mean test loss of 8 batches: 0.3594448193907738. [ Fri Apr 14 16:57:18 2023 ] Top1: 87.72% [ Fri Apr 14 16:57:18 2023 ] Top5: 98.71% [ Fri Apr 14 16:57:18 2023 ] Training epoch: 73 [ Fri Apr 14 16:59:17 2023 ] Mean training loss: 0.2628. Mean training acc: 96.89%. [ Fri Apr 14 16:59:17 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 16:59:17 2023 ] Eval epoch: 73 [ Fri Apr 14 16:59:25 2023 ] Mean test loss of 8 batches: 0.4629015224054456. [ Fri Apr 14 16:59:25 2023 ] Top1: 83.84% [ Fri Apr 14 16:59:25 2023 ] Top5: 99.14% [ Fri Apr 14 16:59:25 2023 ] Training epoch: 74 [ Fri Apr 14 17:01:25 2023 ] Mean training loss: 0.2042. Mean training acc: 97.55%. [ Fri Apr 14 17:01:25 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 17:01:25 2023 ] Eval epoch: 74 [ Fri Apr 14 17:01:31 2023 ] Mean test loss of 8 batches: 0.4135208409279585. [ Fri Apr 14 17:01:31 2023 ] Top1: 89.22% [ Fri Apr 14 17:01:31 2023 ] Top5: 98.71% [ Fri Apr 14 17:01:31 2023 ] Training epoch: 75 [ Fri Apr 14 17:03:28 2023 ] Mean training loss: 0.2335. Mean training acc: 97.01%. [ Fri Apr 14 17:03:28 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:03:28 2023 ] Eval epoch: 75 [ Fri Apr 14 17:03:35 2023 ] Mean test loss of 8 batches: 0.5563358834479004. [ Fri Apr 14 17:03:35 2023 ] Top1: 81.68% [ Fri Apr 14 17:03:35 2023 ] Top5: 99.14% [ Fri Apr 14 17:03:35 2023 ] Training epoch: 76 [ Fri Apr 14 17:05:30 2023 ] Mean training loss: 0.2321. Mean training acc: 97.13%. [ Fri Apr 14 17:05:30 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:05:30 2023 ] Eval epoch: 76 [ Fri Apr 14 17:05:39 2023 ] Mean test loss of 8 batches: 0.3749515376985073. [ Fri Apr 14 17:05:39 2023 ] Top1: 89.22% [ Fri Apr 14 17:05:39 2023 ] Top5: 99.57% [ Fri Apr 14 17:05:39 2023 ] Training epoch: 77 [ Fri Apr 14 17:07:39 2023 ] Mean training loss: 0.2554. Mean training acc: 96.72%. [ Fri Apr 14 17:07:39 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:07:39 2023 ] Eval epoch: 77 [ Fri Apr 14 17:07:45 2023 ] Mean test loss of 8 batches: 0.6634449735283852. [ Fri Apr 14 17:07:45 2023 ] Top1: 77.59% [ Fri Apr 14 17:07:45 2023 ] Top5: 99.57% [ Fri Apr 14 17:07:45 2023 ] Training epoch: 78 [ Fri Apr 14 17:09:40 2023 ] Mean training loss: 0.1991. Mean training acc: 97.69%. [ Fri Apr 14 17:09:40 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:09:40 2023 ] Eval epoch: 78 [ Fri Apr 14 17:09:48 2023 ] Mean test loss of 8 batches: 0.2975238896906376. [ Fri Apr 14 17:09:48 2023 ] Top1: 91.16% [ Fri Apr 14 17:09:48 2023 ] Top5: 99.57% [ Fri Apr 14 17:09:48 2023 ] Training epoch: 79 [ Fri Apr 14 17:11:48 2023 ] Mean training loss: 0.2152. Mean training acc: 97.31%. [ Fri Apr 14 17:11:48 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 17:11:48 2023 ] Eval epoch: 79 [ Fri Apr 14 17:11:55 2023 ] Mean test loss of 8 batches: 1.269447237253189. [ Fri Apr 14 17:11:55 2023 ] Top1: 72.63% [ Fri Apr 14 17:11:55 2023 ] Top5: 97.84% [ Fri Apr 14 17:11:55 2023 ] Training epoch: 80 [ Fri Apr 14 17:13:50 2023 ] Mean training loss: 0.2314. Mean training acc: 96.89%. [ Fri Apr 14 17:13:50 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:13:50 2023 ] Eval epoch: 80 [ Fri Apr 14 17:13:58 2023 ] Mean test loss of 8 batches: 0.4147953391075134. [ Fri Apr 14 17:13:58 2023 ] Top1: 84.48% [ Fri Apr 14 17:13:58 2023 ] Top5: 99.57% [ Fri Apr 14 17:13:58 2023 ] Training epoch: 81 [ Fri Apr 14 17:15:58 2023 ] Mean training loss: 0.2456. Mean training acc: 96.95%. [ Fri Apr 14 17:15:58 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 17:15:58 2023 ] Eval epoch: 81 [ Fri Apr 14 17:16:03 2023 ] Mean test loss of 8 batches: 1.4877559300512075. [ Fri Apr 14 17:16:03 2023 ] Top1: 71.55% [ Fri Apr 14 17:16:03 2023 ] Top5: 99.14% [ Fri Apr 14 17:16:03 2023 ] Training epoch: 82 [ Fri Apr 14 17:17:58 2023 ] Mean training loss: 0.2359. Mean training acc: 97.05%. [ Fri Apr 14 17:17:58 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:17:59 2023 ] Eval epoch: 82 [ Fri Apr 14 17:18:07 2023 ] Mean test loss of 8 batches: 0.3670224826782942. [ Fri Apr 14 17:18:07 2023 ] Top1: 85.78% [ Fri Apr 14 17:18:07 2023 ] Top5: 99.35% [ Fri Apr 14 17:18:07 2023 ] Training epoch: 83 [ Fri Apr 14 17:20:04 2023 ] Mean training loss: 0.2423. Mean training acc: 96.56%. [ Fri Apr 14 17:20:04 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:20:04 2023 ] Eval epoch: 83 [ Fri Apr 14 17:20:12 2023 ] Mean test loss of 8 batches: 0.7229135828092694. [ Fri Apr 14 17:20:12 2023 ] Top1: 81.03% [ Fri Apr 14 17:20:12 2023 ] Top5: 99.35% [ Fri Apr 14 17:20:12 2023 ] Training epoch: 84 [ Fri Apr 14 17:22:12 2023 ] Mean training loss: 0.2301. Mean training acc: 97.03%. [ Fri Apr 14 17:22:12 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:22:12 2023 ] Eval epoch: 84 [ Fri Apr 14 17:22:18 2023 ] Mean test loss of 8 batches: 0.33520559780299664. [ Fri Apr 14 17:22:18 2023 ] Top1: 90.73% [ Fri Apr 14 17:22:18 2023 ] Top5: 100.00% [ Fri Apr 14 17:22:18 2023 ] Training epoch: 85 [ Fri Apr 14 17:24:18 2023 ] Mean training loss: 0.1837. Mean training acc: 97.78%. [ Fri Apr 14 17:24:18 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 17:24:18 2023 ] Eval epoch: 85 [ Fri Apr 14 17:24:25 2023 ] Mean test loss of 8 batches: 0.3012602720409632. [ Fri Apr 14 17:24:25 2023 ] Top1: 92.46% [ Fri Apr 14 17:24:25 2023 ] Top5: 99.57% [ Fri Apr 14 17:24:25 2023 ] Training epoch: 86 [ Fri Apr 14 17:26:22 2023 ] Mean training loss: 0.2300. Mean training acc: 97.31%. [ Fri Apr 14 17:26:22 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:26:22 2023 ] Eval epoch: 86 [ Fri Apr 14 17:26:30 2023 ] Mean test loss of 8 batches: 0.2894723905483261. [ Fri Apr 14 17:26:30 2023 ] Top1: 89.66% [ Fri Apr 14 17:26:30 2023 ] Top5: 99.14% [ Fri Apr 14 17:26:30 2023 ] Training epoch: 87 [ Fri Apr 14 17:28:29 2023 ] Mean training loss: 0.2351. Mean training acc: 97.29%. [ Fri Apr 14 17:28:29 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 17:28:29 2023 ] Eval epoch: 87 [ Fri Apr 14 17:28:36 2023 ] Mean test loss of 8 batches: 0.5256123784929514. [ Fri Apr 14 17:28:36 2023 ] Top1: 83.84% [ Fri Apr 14 17:28:36 2023 ] Top5: 99.35% [ Fri Apr 14 17:28:36 2023 ] Training epoch: 88 [ Fri Apr 14 17:30:36 2023 ] Mean training loss: 0.2715. Mean training acc: 96.52%. [ Fri Apr 14 17:30:36 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:30:36 2023 ] Eval epoch: 88 [ Fri Apr 14 17:30:42 2023 ] Mean test loss of 8 batches: 0.4516562847420573. [ Fri Apr 14 17:30:42 2023 ] Top1: 88.15% [ Fri Apr 14 17:30:42 2023 ] Top5: 98.49% [ Fri Apr 14 17:30:42 2023 ] Training epoch: 89 [ Fri Apr 14 17:32:37 2023 ] Mean training loss: 0.2110. Mean training acc: 97.53%. [ Fri Apr 14 17:32:37 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 17:32:38 2023 ] Eval epoch: 89 [ Fri Apr 14 17:32:46 2023 ] Mean test loss of 8 batches: 0.439712205901742. [ Fri Apr 14 17:32:46 2023 ] Top1: 86.85% [ Fri Apr 14 17:32:46 2023 ] Top5: 99.57% [ Fri Apr 14 17:32:46 2023 ] Training epoch: 90 [ Fri Apr 14 17:34:46 2023 ] Mean training loss: 0.2257. Mean training acc: 97.13%. [ Fri Apr 14 17:34:46 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:34:46 2023 ] Eval epoch: 90 [ Fri Apr 14 17:34:53 2023 ] Mean test loss of 8 batches: 0.40959030017256737. [ Fri Apr 14 17:34:53 2023 ] Top1: 87.50% [ Fri Apr 14 17:34:53 2023 ] Top5: 98.28% [ Fri Apr 14 17:34:53 2023 ] Training epoch: 91 [ Fri Apr 14 17:36:47 2023 ] Mean training loss: 0.1235. Mean training acc: 98.91%. [ Fri Apr 14 17:36:47 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:36:47 2023 ] Eval epoch: 91 [ Fri Apr 14 17:36:54 2023 ] Mean test loss of 8 batches: 0.26408413518220186. [ Fri Apr 14 17:36:54 2023 ] Top1: 92.67% [ Fri Apr 14 17:36:54 2023 ] Top5: 99.35% [ Fri Apr 14 17:36:54 2023 ] Training epoch: 92 [ Fri Apr 14 17:38:53 2023 ] Mean training loss: 0.0719. Mean training acc: 99.56%. [ Fri Apr 14 17:38:53 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:38:54 2023 ] Eval epoch: 92 [ Fri Apr 14 17:39:00 2023 ] Mean test loss of 8 batches: 0.243209108710289. [ Fri Apr 14 17:39:00 2023 ] Top1: 93.10% [ Fri Apr 14 17:39:00 2023 ] Top5: 99.35% [ Fri Apr 14 17:39:00 2023 ] Training epoch: 93 [ Fri Apr 14 17:40:53 2023 ] Mean training loss: 0.0563. Mean training acc: 99.80%. [ Fri Apr 14 17:40:53 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 17:40:54 2023 ] Eval epoch: 93 [ Fri Apr 14 17:41:00 2023 ] Mean test loss of 8 batches: 0.23074489552527666. [ Fri Apr 14 17:41:00 2023 ] Top1: 93.32% [ Fri Apr 14 17:41:00 2023 ] Top5: 99.35% [ Fri Apr 14 17:41:00 2023 ] Training epoch: 94 [ Fri Apr 14 17:42:59 2023 ] Mean training loss: 0.0538. Mean training acc: 99.86%. [ Fri Apr 14 17:42:59 2023 ] Time consumption: [Data]05%, [Network]94% [ Fri Apr 14 17:42:59 2023 ] Eval epoch: 94 [ Fri Apr 14 17:43:07 2023 ] Mean test loss of 8 batches: 0.23704099655151367. [ Fri Apr 14 17:43:07 2023 ] Top1: 93.53% [ Fri Apr 14 17:43:07 2023 ] Top5: 99.14% [ Fri Apr 14 17:43:07 2023 ] Training epoch: 95 [ Fri Apr 14 17:45:07 2023 ] Mean training loss: 0.0487. Mean training acc: 99.92%. [ Fri Apr 14 17:45:07 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 17:45:07 2023 ] Eval epoch: 95 [ Fri Apr 14 17:45:16 2023 ] Mean test loss of 8 batches: 0.22716999240219593. [ Fri Apr 14 17:45:16 2023 ] Top1: 93.75% [ Fri Apr 14 17:45:16 2023 ] Top5: 99.35% [ Fri Apr 14 17:45:16 2023 ] Training epoch: 96 [ Fri Apr 14 17:47:17 2023 ] Mean training loss: 0.0514. Mean training acc: 99.82%. [ Fri Apr 14 17:47:17 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:47:17 2023 ] Eval epoch: 96 [ Fri Apr 14 17:47:25 2023 ] Mean test loss of 8 batches: 0.23398507945239544. [ Fri Apr 14 17:47:25 2023 ] Top1: 93.97% [ Fri Apr 14 17:47:25 2023 ] Top5: 99.35% [ Fri Apr 14 17:47:25 2023 ] Training epoch: 97 [ Fri Apr 14 17:49:21 2023 ] Mean training loss: 0.0455. Mean training acc: 99.90%. [ Fri Apr 14 17:49:21 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 17:49:21 2023 ] Eval epoch: 97 [ Fri Apr 14 17:49:29 2023 ] Mean test loss of 8 batches: 0.22355580888688564. [ Fri Apr 14 17:49:29 2023 ] Top1: 94.18% [ Fri Apr 14 17:49:29 2023 ] Top5: 99.35% [ Fri Apr 14 17:49:29 2023 ] Training epoch: 98 [ Fri Apr 14 17:51:26 2023 ] Mean training loss: 0.0475. Mean training acc: 99.88%. [ Fri Apr 14 17:51:26 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 17:51:26 2023 ] Eval epoch: 98 [ Fri Apr 14 17:51:34 2023 ] Mean test loss of 8 batches: 0.2377201751805842. [ Fri Apr 14 17:51:34 2023 ] Top1: 93.53% [ Fri Apr 14 17:51:34 2023 ] Top5: 99.35% [ Fri Apr 14 17:51:34 2023 ] Training epoch: 99 [ Fri Apr 14 17:53:34 2023 ] Mean training loss: 0.0435. Mean training acc: 99.94%. [ Fri Apr 14 17:53:34 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:53:34 2023 ] Eval epoch: 99 [ Fri Apr 14 17:53:42 2023 ] Mean test loss of 8 batches: 0.240583386272192. [ Fri Apr 14 17:53:42 2023 ] Top1: 93.53% [ Fri Apr 14 17:53:42 2023 ] Top5: 99.35% [ Fri Apr 14 17:53:42 2023 ] Training epoch: 100 [ Fri Apr 14 17:55:38 2023 ] Mean training loss: 0.0416. Mean training acc: 99.96%. [ Fri Apr 14 17:55:38 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 17:55:38 2023 ] Eval epoch: 100 [ Fri Apr 14 17:55:46 2023 ] Mean test loss of 8 batches: 0.22863355558365583. [ Fri Apr 14 17:55:46 2023 ] Top1: 93.32% [ Fri Apr 14 17:55:46 2023 ] Top5: 99.35% [ Fri Apr 14 17:55:46 2023 ] Training epoch: 101 [ Fri Apr 14 17:57:48 2023 ] Mean training loss: 0.0419. Mean training acc: 99.98%. [ Fri Apr 14 17:57:48 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:57:48 2023 ] Eval epoch: 101 [ Fri Apr 14 17:57:56 2023 ] Mean test loss of 8 batches: 0.26161503326147795. [ Fri Apr 14 17:57:56 2023 ] Top1: 92.67% [ Fri Apr 14 17:57:56 2023 ] Top5: 99.35% [ Fri Apr 14 17:57:56 2023 ] Training epoch: 102 [ Fri Apr 14 17:59:52 2023 ] Mean training loss: 0.0407. Mean training acc: 99.98%. [ Fri Apr 14 17:59:52 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 17:59:52 2023 ] Eval epoch: 102 [ Fri Apr 14 18:00:00 2023 ] Mean test loss of 8 batches: 0.2454188708215952. [ Fri Apr 14 18:00:00 2023 ] Top1: 93.10% [ Fri Apr 14 18:00:00 2023 ] Top5: 99.35% [ Fri Apr 14 18:00:00 2023 ] Training epoch: 103 [ Fri Apr 14 18:02:02 2023 ] Mean training loss: 0.0424. Mean training acc: 99.92%. [ Fri Apr 14 18:02:02 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 18:02:02 2023 ] Eval epoch: 103 [ Fri Apr 14 18:02:10 2023 ] Mean test loss of 8 batches: 0.2351202191784978. [ Fri Apr 14 18:02:10 2023 ] Top1: 93.53% [ Fri Apr 14 18:02:10 2023 ] Top5: 99.35% [ Fri Apr 14 18:02:10 2023 ] Training epoch: 104 [ Fri Apr 14 18:04:05 2023 ] Mean training loss: 0.0421. Mean training acc: 99.92%. [ Fri Apr 14 18:04:05 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 18:04:05 2023 ] Eval epoch: 104 [ Fri Apr 14 18:04:13 2023 ] Mean test loss of 8 batches: 0.24127063294872642. [ Fri Apr 14 18:04:13 2023 ] Top1: 93.53% [ Fri Apr 14 18:04:13 2023 ] Top5: 99.35% [ Fri Apr 14 18:04:13 2023 ] Training epoch: 105 [ Fri Apr 14 18:06:12 2023 ] Mean training loss: 0.0390. Mean training acc: 99.96%. [ Fri Apr 14 18:06:12 2023 ] Time consumption: [Data]07%, [Network]92% [ Fri Apr 14 18:06:12 2023 ] Eval epoch: 105 [ Fri Apr 14 18:06:19 2023 ] Mean test loss of 8 batches: 0.23500662064179778. [ Fri Apr 14 18:06:19 2023 ] Top1: 93.97% [ Fri Apr 14 18:06:19 2023 ] Top5: 99.35% [ Fri Apr 14 18:06:19 2023 ] Training epoch: 106 [ Fri Apr 14 18:08:19 2023 ] Mean training loss: 0.0427. Mean training acc: 99.90%. [ Fri Apr 14 18:08:19 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 18:08:19 2023 ] Eval epoch: 106 [ Fri Apr 14 18:08:27 2023 ] Mean test loss of 8 batches: 0.23353653773665428. [ Fri Apr 14 18:08:27 2023 ] Top1: 93.53% [ Fri Apr 14 18:08:27 2023 ] Top5: 99.35% [ Fri Apr 14 18:08:27 2023 ] Training epoch: 107 [ Fri Apr 14 18:10:27 2023 ] Mean training loss: 0.0383. Mean training acc: 99.94%. [ Fri Apr 14 18:10:27 2023 ] Time consumption: [Data]06%, [Network]93% [ Fri Apr 14 18:10:27 2023 ] Eval epoch: 107 [ Fri Apr 14 18:10:34 2023 ] Mean test loss of 8 batches: 0.24807305447757244. [ Fri Apr 14 18:10:34 2023 ] Top1: 93.10% [ Fri Apr 14 18:10:34 2023 ] Top5: 99.35% [ Fri Apr 14 18:10:34 2023 ] Training epoch: 108 [ Fri Apr 14 18:12:32 2023 ] Mean training loss: 0.0404. Mean training acc: 99.92%. [ Fri Apr 14 18:12:32 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 18:12:32 2023 ] Eval epoch: 108 [ Fri Apr 14 18:12:38 2023 ] Mean test loss of 8 batches: 0.24318178417161107. [ Fri Apr 14 18:12:38 2023 ] Top1: 93.10% [ Fri Apr 14 18:12:38 2023 ] Top5: 99.35% [ Fri Apr 14 18:12:38 2023 ] Training epoch: 109 [ Fri Apr 14 18:14:34 2023 ] Mean training loss: 0.0395. Mean training acc: 99.94%. [ Fri Apr 14 18:14:34 2023 ] Time consumption: [Data]07%, [Network]93% [ Fri Apr 14 18:14:34 2023 ] Eval epoch: 109 [ Fri Apr 14 18:14:41 2023 ] Mean test loss of 8 batches: 0.23533811047673225. [ Fri Apr 14 18:14:41 2023 ] Top1: 93.32% [ Fri Apr 14 18:14:41 2023 ] Top5: 99.35% [ Fri Apr 14 18:14:41 2023 ] Training epoch: 110 [ Fri Apr 14 18:16:40 2023 ] Mean training loss: 0.0419. Mean training acc: 99.96%. [ Fri Apr 14 18:16:40 2023 ] Time consumption: [Data]06%, [Network]94% [ Fri Apr 14 18:16:41 2023 ] Eval epoch: 110 [ Fri Apr 14 18:16:46 2023 ] Mean test loss of 8 batches: 0.2387898787856102. [ Fri Apr 14 18:16:46 2023 ] Top1: 93.53% [ Fri Apr 14 18:16:46 2023 ] Top5: 99.35% [ Fri Apr 14 18:16:53 2023 ] Best accuracy: 0.9418103448275862 [ Fri Apr 14 18:16:53 2023 ] Epoch number: 97 [ Fri Apr 14 18:16:53 2023 ] Model name: work_dir/ucla/lst_bone [ Fri Apr 14 18:16:53 2023 ] Model total number of params: 2084817 [ Fri Apr 14 18:16:53 2023 ] Weight decay: 0.0004 [ Fri Apr 14 18:16:53 2023 ] Base LR: 0.2 [ Fri Apr 14 18:16:53 2023 ] Batch Size: 64 [ Fri Apr 14 18:16:53 2023 ] Test Batch Size: 64 [ Fri Apr 14 18:16:53 2023 ] seed: 1 [ Fri Apr 14 14:32:50 2023 ] using warm up, epoch: 5 [ Fri Apr 14 14:32:57 2023 ] Parameters: {'work_dir': 'work_dir/ucla/lst_bone_vel', 'model_saved_name': 'work_dir/ucla/lst_bone_vel/runs', 'config': 'config/ucla/lst_bone_vel.yaml', 'phase': 'train', 'save_score': False, 'seed': 1, 'log_interval': 100, 'save_interval': 1, 'save_epoch': 30, 'eval_interval': 5, 'print_log': True, 'show_topk': [1, 5], 'feeder': 'feeders.feeder_ucla.Feeder', 'num_worker': 32, 'train_feeder_args': {'data_path': 'joint_motion', 'label_path': 'train', 'debug': False, 'random_choose': True, 'random_shift': False, 'random_move': False, 'window_size': 52, 'normalization': False, 'repeat': 5}, 'test_feeder_args': {'data_path': 'joint_motion', 'label_path': 'val', 'debug': False}, 'model': 'model.ctrgcn.Model_lst_4part_ucla', 'model_args': {'num_class': 10, 'num_point': 20, 'num_person': 1, 'graph': 'graph.ucla.Graph', 'graph_args': {'labeling_mode': 'spatial'}, 'head': ['ViT-B/32'], 'k': 1}, 'weights': None, 'ignore_weights': [], 'base_lr': 0.2, 'step': [90, 100], 'device': 0, 'optimizer': 'SGD', 'nesterov': True, 'batch_size': 64, 'test_batch_size': 64, 'start_epoch': 0, 'num_epoch': 110, 'weight_decay': 0.0004, 'lr_decay_rate': 0.1, 'warm_up_epoch': 5, 'loss_alpha': 1.0, 'te_lr_ratio': 1} [ Fri Apr 14 14:32:57 2023 ] # Parameters: 2084817 [ Fri Apr 14 14:32:57 2023 ] Training epoch: 1 [ Fri Apr 14 22:12:15 2023 ] using warm up, epoch: 5 [ Fri Apr 14 22:12:18 2023 ] Parameters: {'work_dir': 'work_dir/ucla/lst_bone_vel', 'model_saved_name': 'work_dir/ucla/lst_bone_vel/runs', 'config': 'config/ucla/lst_bone_vel.yaml', 'phase': 'train', 'save_score': False, 'seed': 1, 'log_interval': 100, 'save_interval': 1, 'save_epoch': 30, 'eval_interval': 5, 'print_log': True, 'show_topk': [1, 5], 'feeder': 'feeders.feeder_ucla.Feeder', 'num_worker': 32, 'train_feeder_args': {'data_path': 'joint_motion', 'label_path': 'train', 'debug': False, 'random_choose': True, 'random_shift': False, 'random_move': False, 'window_size': 52, 'normalization': False, 'repeat': 5}, 'test_feeder_args': {'data_path': 'joint_motion', 'label_path': 'val', 'debug': False}, 'model': 'model.ctrgcn.Model_lst_4part_ucla', 'model_args': {'num_class': 10, 'num_point': 20, 'num_person': 1, 'graph': 'graph.ucla.Graph', 'graph_args': {'labeling_mode': 'spatial'}, 'head': ['ViT-B/32'], 'k': 1}, 'weights': None, 'ignore_weights': [], 'base_lr': 0.2, 'step': [90, 100], 'device': 0, 'optimizer': 'SGD', 'nesterov': True, 'batch_size': 64, 'test_batch_size': 64, 'start_epoch': 0, 'num_epoch': 110, 'weight_decay': 0.0004, 'lr_decay_rate': 0.1, 'warm_up_epoch': 5, 'loss_alpha': 1.0, 'te_lr_ratio': 1} [ Fri Apr 14 22:12:18 2023 ] # Parameters: 2084817 [ Fri Apr 14 22:12:18 2023 ] Training epoch: 1 [ Fri Apr 14 22:13:06 2023 ] Mean training loss: 4.1919. Mean training acc: 40.47%. [ Fri Apr 14 22:13:06 2023 ] Time consumption: [Data]09%, [Network]90% [ Fri Apr 14 22:13:06 2023 ] Eval epoch: 1 [ Fri Apr 14 22:13:10 2023 ] Mean test loss of 8 batches: 1.037980206310749. [ Fri Apr 14 22:13:10 2023 ] Top1: 60.13% [ Fri Apr 14 22:13:10 2023 ] Top5: 96.34% [ Fri Apr 14 22:13:10 2023 ] Training epoch: 2 [ Fri Apr 14 22:13:57 2023 ] Mean training loss: 2.8410. Mean training acc: 62.22%. [ Fri Apr 14 22:13:57 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:13:57 2023 ] Eval epoch: 2 [ Fri Apr 14 22:14:01 2023 ] Mean test loss of 8 batches: 1.819196417927742. [ Fri Apr 14 22:14:01 2023 ] Top1: 53.66% [ Fri Apr 14 22:14:01 2023 ] Top5: 95.91% [ Fri Apr 14 22:14:01 2023 ] Training epoch: 3 [ Fri Apr 14 22:14:47 2023 ] Mean training loss: 2.0374. Mean training acc: 69.36%. [ Fri Apr 14 22:14:47 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:14:47 2023 ] Eval epoch: 3 [ Fri Apr 14 22:14:50 2023 ] Mean test loss of 8 batches: 0.6821741871535778. [ Fri Apr 14 22:14:50 2023 ] Top1: 73.06% [ Fri Apr 14 22:14:50 2023 ] Top5: 98.49% [ Fri Apr 14 22:14:50 2023 ] Training epoch: 4 [ Fri Apr 14 22:15:36 2023 ] Mean training loss: 1.6140. Mean training acc: 74.82%. [ Fri Apr 14 22:15:36 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:15:36 2023 ] Eval epoch: 4 [ Fri Apr 14 22:15:40 2023 ] Mean test loss of 8 batches: 0.8809452578425407. [ Fri Apr 14 22:15:40 2023 ] Top1: 69.18% [ Fri Apr 14 22:15:40 2023 ] Top5: 98.28% [ Fri Apr 14 22:15:40 2023 ] Training epoch: 5 [ Fri Apr 14 22:16:26 2023 ] Mean training loss: 1.3519. Mean training acc: 76.94%. [ Fri Apr 14 22:16:26 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:16:26 2023 ] Eval epoch: 5 [ Fri Apr 14 22:16:29 2023 ] Mean test loss of 8 batches: 0.6300991773605347. [ Fri Apr 14 22:16:29 2023 ] Top1: 77.16% [ Fri Apr 14 22:16:29 2023 ] Top5: 98.28% [ Fri Apr 14 22:16:29 2023 ] Training epoch: 6 [ Fri Apr 14 22:17:15 2023 ] Mean training loss: 1.0652. Mean training acc: 81.53%. [ Fri Apr 14 22:17:15 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:17:15 2023 ] Eval epoch: 6 [ Fri Apr 14 22:17:19 2023 ] Mean test loss of 8 batches: 0.7609496712684631. [ Fri Apr 14 22:17:19 2023 ] Top1: 75.86% [ Fri Apr 14 22:17:19 2023 ] Top5: 98.28% [ Fri Apr 14 22:17:19 2023 ] Training epoch: 7 [ Fri Apr 14 22:18:05 2023 ] Mean training loss: 0.9280. Mean training acc: 84.02%. [ Fri Apr 14 22:18:05 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:18:05 2023 ] Eval epoch: 7 [ Fri Apr 14 22:18:09 2023 ] Mean test loss of 8 batches: 1.264021947979927. [ Fri Apr 14 22:18:09 2023 ] Top1: 61.21% [ Fri Apr 14 22:18:09 2023 ] Top5: 97.84% [ Fri Apr 14 22:18:09 2023 ] Training epoch: 8 [ Fri Apr 14 22:18:55 2023 ] Mean training loss: 0.8655. Mean training acc: 84.93%. [ Fri Apr 14 22:18:55 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 22:18:55 2023 ] Eval epoch: 8 [ Fri Apr 14 22:18:59 2023 ] Mean test loss of 8 batches: 0.597852285951376. [ Fri Apr 14 22:18:59 2023 ] Top1: 76.29% [ Fri Apr 14 22:18:59 2023 ] Top5: 98.49% [ Fri Apr 14 22:18:59 2023 ] Training epoch: 9 [ Fri Apr 14 22:19:45 2023 ] Mean training loss: 0.7377. Mean training acc: 88.01%. [ Fri Apr 14 22:19:45 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:19:45 2023 ] Eval epoch: 9 [ Fri Apr 14 22:19:50 2023 ] Mean test loss of 8 batches: 0.4564897231757641. [ Fri Apr 14 22:19:50 2023 ] Top1: 86.42% [ Fri Apr 14 22:19:50 2023 ] Top5: 99.35% [ Fri Apr 14 22:19:50 2023 ] Training epoch: 10 [ Fri Apr 14 22:20:36 2023 ] Mean training loss: 0.6974. Mean training acc: 87.72%. [ Fri Apr 14 22:20:36 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:20:36 2023 ] Eval epoch: 10 [ Fri Apr 14 22:20:40 2023 ] Mean test loss of 8 batches: 0.48968643695116043. [ Fri Apr 14 22:20:40 2023 ] Top1: 83.62% [ Fri Apr 14 22:20:40 2023 ] Top5: 98.71% [ Fri Apr 14 22:20:40 2023 ] Training epoch: 11 [ Fri Apr 14 22:21:26 2023 ] Mean training loss: 0.5936. Mean training acc: 90.45%. [ Fri Apr 14 22:21:26 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 22:21:26 2023 ] Eval epoch: 11 [ Fri Apr 14 22:21:30 2023 ] Mean test loss of 8 batches: 0.5147222466766834. [ Fri Apr 14 22:21:30 2023 ] Top1: 82.76% [ Fri Apr 14 22:21:30 2023 ] Top5: 98.71% [ Fri Apr 14 22:21:30 2023 ] Training epoch: 12 [ Fri Apr 14 22:22:16 2023 ] Mean training loss: 0.6033. Mean training acc: 89.97%. [ Fri Apr 14 22:22:16 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:22:16 2023 ] Eval epoch: 12 [ Fri Apr 14 22:22:20 2023 ] Mean test loss of 8 batches: 0.8916015550494194. [ Fri Apr 14 22:22:20 2023 ] Top1: 76.51% [ Fri Apr 14 22:22:20 2023 ] Top5: 98.49% [ Fri Apr 14 22:22:20 2023 ] Training epoch: 13 [ Fri Apr 14 22:23:06 2023 ] Mean training loss: 0.5555. Mean training acc: 90.92%. [ Fri Apr 14 22:23:06 2023 ] Time consumption: [Data]09%, [Network]90% [ Fri Apr 14 22:23:06 2023 ] Eval epoch: 13 [ Fri Apr 14 22:23:10 2023 ] Mean test loss of 8 batches: 0.46765569783747196. [ Fri Apr 14 22:23:10 2023 ] Top1: 82.97% [ Fri Apr 14 22:23:10 2023 ] Top5: 98.71% [ Fri Apr 14 22:23:10 2023 ] Training epoch: 14 [ Fri Apr 14 22:23:56 2023 ] Mean training loss: 0.5117. Mean training acc: 91.75%. [ Fri Apr 14 22:23:56 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:23:56 2023 ] Eval epoch: 14 [ Fri Apr 14 22:24:00 2023 ] Mean test loss of 8 batches: 1.0212577693164349. [ Fri Apr 14 22:24:00 2023 ] Top1: 69.40% [ Fri Apr 14 22:24:00 2023 ] Top5: 99.14% [ Fri Apr 14 22:24:00 2023 ] Training epoch: 15 [ Fri Apr 14 22:24:46 2023 ] Mean training loss: 0.4444. Mean training acc: 93.12%. [ Fri Apr 14 22:24:46 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:24:46 2023 ] Eval epoch: 15 [ Fri Apr 14 22:24:49 2023 ] Mean test loss of 8 batches: 0.8558640778064728. [ Fri Apr 14 22:24:49 2023 ] Top1: 76.29% [ Fri Apr 14 22:24:49 2023 ] Top5: 99.14% [ Fri Apr 14 22:24:49 2023 ] Training epoch: 16 [ Fri Apr 14 22:25:35 2023 ] Mean training loss: 0.4600. Mean training acc: 93.08%. [ Fri Apr 14 22:25:35 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:25:35 2023 ] Eval epoch: 16 [ Fri Apr 14 22:25:39 2023 ] Mean test loss of 8 batches: 0.5020748749375343. [ Fri Apr 14 22:25:39 2023 ] Top1: 84.91% [ Fri Apr 14 22:25:39 2023 ] Top5: 98.71% [ Fri Apr 14 22:25:39 2023 ] Training epoch: 17 [ Fri Apr 14 22:26:25 2023 ] Mean training loss: 0.4555. Mean training acc: 92.88%. [ Fri Apr 14 22:26:25 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:26:25 2023 ] Eval epoch: 17 [ Fri Apr 14 22:26:29 2023 ] Mean test loss of 8 batches: 0.8993323221802711. [ Fri Apr 14 22:26:29 2023 ] Top1: 76.51% [ Fri Apr 14 22:26:29 2023 ] Top5: 98.71% [ Fri Apr 14 22:26:29 2023 ] Training epoch: 18 [ Fri Apr 14 22:27:15 2023 ] Mean training loss: 0.4296. Mean training acc: 93.33%. [ Fri Apr 14 22:27:15 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:27:15 2023 ] Eval epoch: 18 [ Fri Apr 14 22:27:19 2023 ] Mean test loss of 8 batches: 0.6480088643729687. [ Fri Apr 14 22:27:19 2023 ] Top1: 77.37% [ Fri Apr 14 22:27:19 2023 ] Top5: 98.92% [ Fri Apr 14 22:27:19 2023 ] Training epoch: 19 [ Fri Apr 14 22:28:05 2023 ] Mean training loss: 0.4061. Mean training acc: 93.95%. [ Fri Apr 14 22:28:05 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:28:05 2023 ] Eval epoch: 19 [ Fri Apr 14 22:28:09 2023 ] Mean test loss of 8 batches: 1.271400772035122. [ Fri Apr 14 22:28:09 2023 ] Top1: 68.97% [ Fri Apr 14 22:28:09 2023 ] Top5: 97.84% [ Fri Apr 14 22:28:09 2023 ] Training epoch: 20 [ Fri Apr 14 22:28:55 2023 ] Mean training loss: 0.4121. Mean training acc: 93.61%. [ Fri Apr 14 22:28:55 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 22:28:55 2023 ] Eval epoch: 20 [ Fri Apr 14 22:28:59 2023 ] Mean test loss of 8 batches: 1.024251714348793. [ Fri Apr 14 22:28:59 2023 ] Top1: 72.20% [ Fri Apr 14 22:28:59 2023 ] Top5: 98.71% [ Fri Apr 14 22:28:59 2023 ] Training epoch: 21 [ Fri Apr 14 22:29:45 2023 ] Mean training loss: 0.3940. Mean training acc: 94.13%. [ Fri Apr 14 22:29:45 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 22:29:45 2023 ] Eval epoch: 21 [ Fri Apr 14 22:29:49 2023 ] Mean test loss of 8 batches: 0.6216142438352108. [ Fri Apr 14 22:29:49 2023 ] Top1: 83.84% [ Fri Apr 14 22:29:49 2023 ] Top5: 99.35% [ Fri Apr 14 22:29:49 2023 ] Training epoch: 22 [ Fri Apr 14 22:30:35 2023 ] Mean training loss: 0.3796. Mean training acc: 94.32%. [ Fri Apr 14 22:30:35 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:30:35 2023 ] Eval epoch: 22 [ Fri Apr 14 22:30:39 2023 ] Mean test loss of 8 batches: 3.570750206708908. [ Fri Apr 14 22:30:39 2023 ] Top1: 63.36% [ Fri Apr 14 22:30:39 2023 ] Top5: 97.63% [ Fri Apr 14 22:30:39 2023 ] Training epoch: 23 [ Fri Apr 14 22:31:25 2023 ] Mean training loss: 0.3169. Mean training acc: 95.33%. [ Fri Apr 14 22:31:25 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:31:25 2023 ] Eval epoch: 23 [ Fri Apr 14 22:31:29 2023 ] Mean test loss of 8 batches: 0.7545811533927917. [ Fri Apr 14 22:31:29 2023 ] Top1: 75.43% [ Fri Apr 14 22:31:29 2023 ] Top5: 99.35% [ Fri Apr 14 22:31:29 2023 ] Training epoch: 24 [ Fri Apr 14 22:32:15 2023 ] Mean training loss: 0.3810. Mean training acc: 94.50%. [ Fri Apr 14 22:32:15 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:32:15 2023 ] Eval epoch: 24 [ Fri Apr 14 22:32:19 2023 ] Mean test loss of 8 batches: 0.6958607286214828. [ Fri Apr 14 22:32:19 2023 ] Top1: 86.42% [ Fri Apr 14 22:32:19 2023 ] Top5: 98.92% [ Fri Apr 14 22:32:19 2023 ] Training epoch: 25 [ Fri Apr 14 22:33:05 2023 ] Mean training loss: 0.3449. Mean training acc: 95.19%. [ Fri Apr 14 22:33:05 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:33:05 2023 ] Eval epoch: 25 [ Fri Apr 14 22:33:09 2023 ] Mean test loss of 8 batches: 0.630146861076355. [ Fri Apr 14 22:33:09 2023 ] Top1: 83.19% [ Fri Apr 14 22:33:09 2023 ] Top5: 99.14% [ Fri Apr 14 22:33:09 2023 ] Training epoch: 26 [ Fri Apr 14 22:33:55 2023 ] Mean training loss: 0.3214. Mean training acc: 95.79%. [ Fri Apr 14 22:33:55 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:33:55 2023 ] Eval epoch: 26 [ Fri Apr 14 22:33:58 2023 ] Mean test loss of 8 batches: 0.618928013369441. [ Fri Apr 14 22:33:58 2023 ] Top1: 80.82% [ Fri Apr 14 22:33:58 2023 ] Top5: 98.49% [ Fri Apr 14 22:33:58 2023 ] Training epoch: 27 [ Fri Apr 14 22:34:44 2023 ] Mean training loss: 0.3391. Mean training acc: 95.11%. [ Fri Apr 14 22:34:44 2023 ] Time consumption: [Data]09%, [Network]91% [ Fri Apr 14 22:34:44 2023 ] Eval epoch: 27 [ Fri Apr 14 22:34:48 2023 ] Mean test loss of 8 batches: 1.5345043316483498. [ Fri Apr 14 22:34:48 2023 ] Top1: 66.16% [ Fri Apr 14 22:34:48 2023 ] Top5: 98.71% [ Fri Apr 14 22:34:48 2023 ] Training epoch: 28 [ Fri Apr 14 22:35:34 2023 ] Mean training loss: 0.3286. Mean training acc: 95.11%. [ Fri Apr 14 22:35:34 2023 ] Time consumption: [Data]08%, [Network]91% [ Fri Apr 14 22:35:34 2023 ] Eval epoch: 28 [ Fri Apr 14 22:35:38 2023 ] Mean test loss of 8 batches: 0.4797500930726528. [ Fri Apr 14 22:35:38 2023 ] Top1: 84.05% [ Fri Apr 14 22:35:38 2023 ] Top5: 98.71% [ Fri Apr 14 22:35:38 2023 ] Training epoch: 29 [ Fri Apr 14 22:36:24 2023 ] Mean training loss: 0.2819. Mean training acc: 95.77%. [ Fri Apr 14 22:36:24 2023 ] Time consumption: [Data]08%, [Network]92% [ Fri Apr 14 22:36:24 2023 ] Eval epoch: 29 [ Fri Apr 14 22:36:28 2023 ] Mean test loss of 8 batches: 0.9707095623016357. [ Fri Apr 14 22:36:28 2023 ] Top1: 76.29% [ Fri Apr 14 22:36:28 2023 ] Top5: 98.71% [ Fri Apr 14 22:36:28 2023 ] Training epoch: 30 [ Fri Apr 14 22:37:14 2023 ] Mean training loss: 0.3038. Mean training

[MartinXM]

The training log seems fine to me. Have you tried CTR-GCN on NW-UCLA? Can you achieve the performance reported in CTR-GCN?

[yanghhx]

I haven't tried it yet, but I personally believe that the method you provided should achieve the similar results in your paper. At 2023-04-15 17:40:43, "Martin" ***@***.***> wrote: The training log seems fine to me. Have you tried CTR-GCN on NW-UCLA? Can you achieve the performance reported in CTR-GCN? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>