Skip to content

Challenges Encountered in the Experimental Results #1

Open
Open
Challenges Encountered in the Experimental Results#1

Description

@KORIYN
KORIYN
opened on Dec 10, 2025

2025-12-02 20:36:07 [trainer.py] Run Name: pina_vit_domainnet_deep16
2025-12-02 20:36:07 [trainer.py] config: configs/domainnet_pina_vit.yaml
2025-12-02 20:36:07 [trainer.py] device: [device(type='cuda', index=0)]
2025-12-02 20:36:07 [trainer.py] dataset: domainnet
2025-12-02 20:36:07 [trainer.py] data_path: /mnt/0e754b13-a8ea-4e74-bfac-793a2a74bccf/wgr/data/DIL/DomainNet
2025-12-02 20:36:07 [trainer.py] log_path: ./_output
2025-12-02 20:36:07 [trainer.py] init_cls: 345
2025-12-02 20:36:07 [trainer.py] increment: 345
2025-12-02 20:36:07 [trainer.py] total_sessions: 6
2025-12-02 20:36:07 [trainer.py] model_name: pina
2025-12-02 20:36:07 [trainer.py] net_type: pina_vit
2025-12-02 20:36:07 [trainer.py] image_dim: 768
2025-12-02 20:36:07 [trainer.py] prompt_length: 10
2025-12-02 20:36:07 [trainer.py] ca_mode: deep
2025-12-02 20:36:07 [trainer.py] hidden_dim: 16
2025-12-02 20:36:07 [trainer.py] init_epoch: 30
2025-12-02 20:36:07 [trainer.py] init_lr: 0.01
2025-12-02 20:36:07 [trainer.py] init_lr_decay: 0.1
2025-12-02 20:36:07 [trainer.py] init_weight_decay: 0.0002
2025-12-02 20:36:07 [trainer.py] epochs: 30
2025-12-02 20:36:07 [trainer.py] lr: 0.01
2025-12-02 20:36:07 [trainer.py] lr_decay: 0.1
2025-12-02 20:36:07 [trainer.py] weight_decay: 0.0002
2025-12-02 20:36:07 [trainer.py] batch_size: 128
2025-12-02 20:36:07 [trainer.py] seed: 0
2025-12-02 20:36:07 [trainer.py] num_workers: 16
2025-12-02 20:36:07 [trainer.py] memory_size: 0
2025-12-02 20:36:07 [trainer.py] memory_per_class: 0
2025-12-02 20:36:07 [trainer.py] fixed_memory: True
2025-12-02 20:36:07 [trainer.py] shuffle: False
2025-12-02 20:36:07 [trainer.py] EPSILON: 1e-08
2025-12-02 20:36:10 [pina.py]
2025-12-02 20:36:10 [pina.py] ==> Training task 0, Learning on 0-345
2025-12-02 20:36:11 [pina.py] len(train_dataset): 33525
2025-12-02 20:36:11 [pina.py] len(test_dataset): 14604
2025-12-02 20:36:12 [pina.py] ==> Checking the parameter
2025-12-02 20:36:12 [pina.py] Total parameters: 93463489
2025-12-02 20:36:12 [pina.py] Trainable parameters: 272985
2025-12-02 20:36:12 [pina.py] Blocks:
2025-12-02 20:36:12 [pina.py] image_encoder: 91326184
2025-12-02 20:36:12 [pina.py] unified_classifier: 265305
2025-12-02 20:36:12 [pina.py] prompt_pool: 46080
2025-12-02 20:36:12 [pina.py] down_pool: 885888
2025-12-02 20:36:12 [pina.py] up_pool: 940032
2025-12-02 20:36:12 [pina.py] Training:
2025-12-02 20:36:12 [pina.py] unified_classifier.weight: torch.Size([345, 768])
2025-12-02 20:36:12 [pina.py] unified_classifier.bias: torch.Size([345])
2025-12-02 20:36:12 [pina.py] prompt_pool.0.weight: torch.Size([10, 768])
2025-12-02 20:38:27 [pina.py] Task 0, Epoch [1/30] lr 0.00997 Loss 3.552, Train_accy 37.00, Test_accy 59.45
2025-12-02 20:40:41 [pina.py] Task 0, Epoch [2/30] lr 0.00989 Loss 2.253, Train_accy 56.22, Test_accy 62.28
2025-12-02 20:42:56 [pina.py] Task 0, Epoch [3/30] lr 0.00976 Loss 1.931, Train_accy 60.64, Test_accy 63.56
2025-12-02 20:45:10 [pina.py] Task 0, Epoch [4/30] lr 0.00957 Loss 1.739, Train_accy 63.80, Test_accy 64.54
2025-12-02 20:47:24 [pina.py] Task 0, Epoch [5/30] lr 0.00933 Loss 1.585, Train_accy 66.48, Test_accy 64.44
2025-12-02 20:49:39 [pina.py] Task 0, Epoch [6/30] lr 0.00905 Loss 1.498, Train_accy 67.98, Test_accy 66.38
2025-12-02 20:51:53 [pina.py] Task 0, Epoch [7/30] lr 0.00872 Loss 1.426, Train_accy 69.01, Test_accy 65.53
2025-12-02 20:54:07 [pina.py] Task 0, Epoch [8/30] lr 0.00835 Loss 1.368, Train_accy 70.08, Test_accy 66.19
2025-12-02 20:56:21 [pina.py] Task 0, Epoch [9/30] lr 0.00794 Loss 1.301, Train_accy 71.16, Test_accy 65.52
2025-12-02 20:58:36 [pina.py] Task 0, Epoch [10/30] lr 0.00750 Loss 1.226, Train_accy 72.57, Test_accy 66.63
2025-12-02 21:00:50 [pina.py] Task 0, Epoch [11/30] lr 0.00703 Loss 1.158, Train_accy 73.64, Test_accy 67.49
2025-12-02 21:03:04 [pina.py] Task 0, Epoch [12/30] lr 0.00655 Loss 1.110, Train_accy 74.69, Test_accy 67.48
2025-12-02 21:05:19 [pina.py] Task 0, Epoch [13/30] lr 0.00604 Loss 1.057, Train_accy 75.57, Test_accy 68.03
2025-12-02 21:07:33 [pina.py] Task 0, Epoch [14/30] lr 0.00552 Loss 1.007, Train_accy 76.66, Test_accy 68.01
2025-12-02 21:09:47 [pina.py] Task 0, Epoch [15/30] lr 0.00500 Loss 0.977, Train_accy 77.15, Test_accy 68.73
2025-12-02 21:12:01 [pina.py] Task 0, Epoch [16/30] lr 0.00448 Loss 0.926, Train_accy 78.39, Test_accy 68.84
2025-12-02 21:14:15 [pina.py] Task 0, Epoch [17/30] lr 0.00396 Loss 0.878, Train_accy 79.33, Test_accy 69.18
2025-12-02 21:16:30 [pina.py] Task 0, Epoch [18/30] lr 0.00345 Loss 0.875, Train_accy 79.58, Test_accy 69.50
2025-12-02 21:18:45 [pina.py] Task 0, Epoch [19/30] lr 0.00297 Loss 0.840, Train_accy 79.96, Test_accy 69.52
2025-12-02 21:20:59 [pina.py] Task 0, Epoch [20/30] lr 0.00250 Loss 0.820, Train_accy 80.62, Test_accy 69.93
2025-12-02 21:23:13 [pina.py] Task 0, Epoch [21/30] lr 0.00206 Loss 0.789, Train_accy 81.30, Test_accy 70.22
2025-12-02 21:25:27 [pina.py] Task 0, Epoch [22/30] lr 0.00165 Loss 0.769, Train_accy 81.70, Test_accy 70.06
2025-12-02 21:27:41 [pina.py] Task 0, Epoch [23/30] lr 0.00128 Loss 0.748, Train_accy 82.46, Test_accy 70.30
2025-12-02 21:29:55 [pina.py] Task 0, Epoch [24/30] lr 0.00095 Loss 0.737, Train_accy 82.89, Test_accy 70.33
2025-12-02 21:32:10 [pina.py] Task 0, Epoch [25/30] lr 0.00067 Loss 0.717, Train_accy 83.10, Test_accy 70.55
2025-12-02 21:34:24 [pina.py] Task 0, Epoch [26/30] lr 0.00043 Loss 0.713, Train_accy 83.25, Test_accy 70.52
2025-12-02 21:36:38 [pina.py] Task 0, Epoch [27/30] lr 0.00024 Loss 0.706, Train_accy 83.53, Test_accy 70.51
2025-12-02 21:38:53 [pina.py] Task 0, Epoch [28/30] lr 0.00011 Loss 0.703, Train_accy 83.51, Test_accy 70.51
2025-12-02 21:41:07 [pina.py] Task 0, Epoch [29/30] lr 0.00003 Loss 0.688, Train_accy 84.00, Test_accy 70.52
2025-12-02 21:43:21 [pina.py] Task 0, Epoch [30/30] lr 0.00000 Loss 0.682, Train_accy 84.24, Test_accy 70.47
2025-12-02 21:43:21 [pina.py]
2025-12-02 21:43:21 [pina.py] ==> Start clustering
2025-12-02 21:44:11 [pina.py] clustering features: (33525, 768)
2025-12-02 21:44:18 [pina.py] clustering centers: (5, 768)
2025-12-02 21:46:43 [trainer.py] CNN: {'total': 70.47, '0-344': 70.47, 'old': 0, 'new': 70.47}
2025-12-02 21:46:43 [trainer.py] CNN top1 curve: [70.47]
2025-12-02 21:46:43 [pina.py] Exemplar size: 0
2025-12-02 21:46:44 [trainer.py] Save the checkpoint task_0.pth
2025-12-02 21:46:44 [pina.py]
2025-12-02 21:46:44 [pina.py] ==> Training task 1, Learning on 345-690
2025-12-02 21:46:45 [pina.py] len(train_dataset): 36023
2025-12-02 21:46:45 [pina.py] len(test_dataset): 30186
2025-12-02 21:46:45 [pina.py] ==> Checking the parameter
2025-12-02 21:46:45 [pina.py] Total parameters: 93463489
2025-12-02 21:46:45 [pina.py] Trainable parameters: 312000
2025-12-02 21:46:45 [pina.py] Blocks:
2025-12-02 21:46:45 [pina.py] image_encoder: 91326184
2025-12-02 21:46:45 [pina.py] unified_classifier: 265305
2025-12-02 21:46:45 [pina.py] prompt_pool: 46080
2025-12-02 21:46:45 [pina.py] down_pool: 885888
2025-12-02 21:46:45 [pina.py] up_pool: 940032
2025-12-02 21:46:45 [pina.py] Training:
2025-12-02 21:46:45 [pina.py] prompt_pool.1.weight: torch.Size([10, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.0.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.0.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.1.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.1.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.2.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.2.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.3.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.3.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.4.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.4.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.5.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.5.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.6.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.6.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.7.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.7.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.8.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.8.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.9.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.9.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.10.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.10.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] down_pool.1.11.weight: torch.Size([16, 768])
2025-12-02 21:46:45 [pina.py] down_pool.1.11.bias: torch.Size([16])
2025-12-02 21:46:45 [pina.py] up_pool.1.0.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.0.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.1.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.1.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.2.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.2.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.3.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.3.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.4.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.4.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.5.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.5.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.6.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.6.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.7.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.7.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.8.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.8.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.9.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.9.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.10.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.10.bias: torch.Size([768])
2025-12-02 21:46:45 [pina.py] up_pool.1.11.weight: torch.Size([768, 16])
2025-12-02 21:46:45 [pina.py] up_pool.1.11.bias: torch.Size([768])
2025-12-02 21:49:34 [pina.py] Task 1, Epoch [1/30] lr 0.00997 Loss 4.379, Train_accy 20.79, Test_accy 38.88
2025-12-02 21:52:23 [pina.py] Task 1, Epoch [2/30] lr 0.00989 Loss 3.537, Train_accy 31.44, Test_accy 43.02
2025-12-02 21:55:12 [pina.py] Task 1, Epoch [3/30] lr 0.00976 Loss 3.296, Train_accy 35.24, Test_accy 43.82
2025-12-02 21:58:00 [pina.py] Task 1, Epoch [4/30] lr 0.00957 Loss 3.166, Train_accy 36.78, Test_accy 44.72
2025-12-02 22:00:49 [pina.py] Task 1, Epoch [5/30] lr 0.00933 Loss 3.066, Train_accy 38.40, Test_accy 46.52
2025-12-02 22:03:37 [pina.py] Task 1, Epoch [6/30] lr 0.00905 Loss 2.976, Train_accy 40.08, Test_accy 47.64
2025-12-02 22:06:25 [pina.py] Task 1, Epoch [7/30] lr 0.00872 Loss 2.916, Train_accy 40.75, Test_accy 46.87
2025-12-02 22:09:14 [pina.py] Task 1, Epoch [8/30] lr 0.00835 Loss 2.840, Train_accy 42.05, Test_accy 46.81
2025-12-02 22:12:02 [pina.py] Task 1, Epoch [9/30] lr 0.00794 Loss 2.804, Train_accy 42.56, Test_accy 47.39
2025-12-02 22:14:49 [pina.py] Task 1, Epoch [10/30] lr 0.00750 Loss 2.745, Train_accy 43.62, Test_accy 48.91
2025-12-02 22:17:36 [pina.py] Task 1, Epoch [11/30] lr 0.00703 Loss 2.690, Train_accy 44.45, Test_accy 48.79
2025-12-02 22:20:23 [pina.py] Task 1, Epoch [12/30] lr 0.00655 Loss 2.669, Train_accy 44.47, Test_accy 49.13
2025-12-02 22:23:35 [pina.py] Task 1, Epoch [13/30] lr 0.00604 Loss 2.628, Train_accy 45.28, Test_accy 49.37
2025-12-02 22:27:00 [pina.py] Task 1, Epoch [14/30] lr 0.00552 Loss 2.598, Train_accy 45.82, Test_accy 49.04
2025-12-02 22:30:25 [pina.py] Task 1, Epoch [15/30] lr 0.00500 Loss 2.558, Train_accy 46.36, Test_accy 49.27
2025-12-02 22:33:50 [pina.py] Task 1, Epoch [16/30] lr 0.00448 Loss 2.518, Train_accy 46.91, Test_accy 49.90
2025-12-02 22:37:15 [pina.py] Task 1, Epoch [17/30] lr 0.00396 Loss 2.498, Train_accy 47.84, Test_accy 49.97
2025-12-02 22:40:39 [pina.py] Task 1, Epoch [18/30] lr 0.00345 Loss 2.453, Train_accy 48.35, Test_accy 50.47
2025-12-02 22:44:05 [pina.py] Task 1, Epoch [19/30] lr 0.00297 Loss 2.432, Train_accy 48.62, Test_accy 50.11
2025-12-02 22:47:31 [pina.py] Task 1, Epoch [20/30] lr 0.00250 Loss 2.405, Train_accy 49.36, Test_accy 50.21
2025-12-02 22:50:56 [pina.py] Task 1, Epoch [21/30] lr 0.00206 Loss 2.383, Train_accy 49.68, Test_accy 50.21
2025-12-02 22:54:22 [pina.py] Task 1, Epoch [22/30] lr 0.00165 Loss 2.354, Train_accy 49.97, Test_accy 49.96
2025-12-02 22:57:48 [pina.py] Task 1, Epoch [23/30] lr 0.00128 Loss 2.327, Train_accy 50.30, Test_accy 49.69
2025-12-02 23:01:14 [pina.py] Task 1, Epoch [24/30] lr 0.00095 Loss 2.304, Train_accy 50.77, Test_accy 49.98
2025-12-02 23:04:39 [pina.py] Task 1, Epoch [25/30] lr 0.00067 Loss 2.298, Train_accy 51.03, Test_accy 50.07
2025-12-02 23:08:05 [pina.py] Task 1, Epoch [26/30] lr 0.00043 Loss 2.284, Train_accy 51.20, Test_accy 50.14
2025-12-02 23:11:31 [pina.py] Task 1, Epoch [27/30] lr 0.00024 Loss 2.280, Train_accy 51.42, Test_accy 50.38
2025-12-02 23:14:56 [pina.py] Task 1, Epoch [28/30] lr 0.00011 Loss 2.245, Train_accy 52.11, Test_accy 50.43
2025-12-02 23:18:22 [pina.py] Task 1, Epoch [29/30] lr 0.00003 Loss 2.247, Train_accy 52.07, Test_accy 50.44
2025-12-02 23:21:48 [pina.py] Task 1, Epoch [30/30] lr 0.00000 Loss 2.245, Train_accy 51.90, Test_accy 50.48
2025-12-02 23:21:48 [pina.py]
2025-12-02 23:21:48 [pina.py] ==> Start clustering
2025-12-02 23:22:53 [pina.py] clustering features: (36023, 768)
2025-12-02 23:23:02 [pina.py] clustering centers: (5, 768)
2025-12-02 23:29:02 [trainer.py] CNN: {'total': 56.03, '0-344': 68.34, '345-689': 44.5, 'old': 68.34, 'new': 44.5}
2025-12-02 23:29:02 [trainer.py] CNN top1 curve: [70.47, 56.03]
2025-12-02 23:29:02 [pina.py] Exemplar size: 0
2025-12-02 23:29:03 [trainer.py] Save the checkpoint task_1.pth
2025-12-02 23:29:03 [pina.py]
2025-12-02 23:29:03 [pina.py] ==> Training task 2, Learning on 690-1035
2025-12-02 23:29:04 [pina.py] len(train_dataset): 50416
2025-12-02 23:29:04 [pina.py] len(test_dataset): 52036
2025-12-02 23:29:04 [pina.py] ==> Checking the parameter
2025-12-02 23:29:04 [pina.py] Total parameters: 93463489
2025-12-02 23:29:04 [pina.py] Trainable parameters: 312000
2025-12-02 23:29:04 [pina.py] Blocks:
2025-12-02 23:29:04 [pina.py] image_encoder: 91326184
2025-12-02 23:29:04 [pina.py] unified_classifier: 265305
2025-12-02 23:29:04 [pina.py] prompt_pool: 46080
2025-12-02 23:29:04 [pina.py] down_pool: 885888
2025-12-02 23:29:04 [pina.py] up_pool: 940032
2025-12-02 23:29:04 [pina.py] Training:
2025-12-02 23:29:04 [pina.py] prompt_pool.2.weight: torch.Size([10, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.0.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.0.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.1.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.1.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.2.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.2.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.3.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.3.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.4.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.4.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.5.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.5.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.6.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.6.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.7.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.7.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.8.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.8.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.9.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.9.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.10.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.10.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] down_pool.2.11.weight: torch.Size([16, 768])
2025-12-02 23:29:04 [pina.py] down_pool.2.11.bias: torch.Size([16])
2025-12-02 23:29:04 [pina.py] up_pool.2.0.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.0.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.1.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.1.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.2.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.2.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.3.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.3.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.4.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.4.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.5.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.5.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.6.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.6.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.7.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.7.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.8.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.8.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.9.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.9.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.10.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.10.bias: torch.Size([768])
2025-12-02 23:29:04 [pina.py] up_pool.2.11.weight: torch.Size([768, 16])
2025-12-02 23:29:04 [pina.py] up_pool.2.11.bias: torch.Size([768])
2025-12-02 23:34:10 [pina.py] Task 2, Epoch [1/30] lr 0.00997 Loss 2.566, Train_accy 49.74, Test_accy 46.24
2025-12-02 23:39:14 [pina.py] Task 2, Epoch [2/30] lr 0.00989 Loss 1.840, Train_accy 61.36, Test_accy 49.20
2025-12-02 23:44:18 [pina.py] Task 2, Epoch [3/30] lr 0.00976 Loss 1.667, Train_accy 64.36, Test_accy 50.21
2025-12-02 23:49:22 [pina.py] Task 2, Epoch [4/30] lr 0.00957 Loss 1.558, Train_accy 66.20, Test_accy 50.22
2025-12-02 23:54:28 [pina.py] Task 2, Epoch [5/30] lr 0.00933 Loss 1.471, Train_accy 67.60, Test_accy 52.16
2025-12-02 23:59:31 [pina.py] Task 2, Epoch [6/30] lr 0.00905 Loss 1.405, Train_accy 68.97, Test_accy 52.54
2025-12-03 00:04:38 [pina.py] Task 2, Epoch [7/30] lr 0.00872 Loss 1.368, Train_accy 69.67, Test_accy 52.36
2025-12-03 00:09:42 [pina.py] Task 2, Epoch [8/30] lr 0.00835 Loss 1.316, Train_accy 70.42, Test_accy 52.71
2025-12-03 00:14:46 [pina.py] Task 2, Epoch [9/30] lr 0.00794 Loss 1.270, Train_accy 71.27, Test_accy 53.92
2025-12-03 00:19:47 [pina.py] Task 2, Epoch [10/30] lr 0.00750 Loss 1.247, Train_accy 71.60, Test_accy 53.53
2025-12-03 00:24:48 [pina.py] Task 2, Epoch [11/30] lr 0.00703 Loss 1.222, Train_accy 72.01, Test_accy 53.47
2025-12-03 00:29:51 [pina.py] Task 2, Epoch [12/30] lr 0.00655 Loss 1.179, Train_accy 73.15, Test_accy 53.61
2025-12-03 00:34:50 [pina.py] Task 2, Epoch [13/30] lr 0.00604 Loss 1.159, Train_accy 73.44, Test_accy 53.90
2025-12-03 00:39:48 [pina.py] Task 2, Epoch [14/30] lr 0.00552 Loss 1.130, Train_accy 74.06, Test_accy 54.26
2025-12-03 00:44:45 [pina.py] Task 2, Epoch [15/30] lr 0.00500 Loss 1.108, Train_accy 74.54, Test_accy 53.73
2025-12-03 00:49:42 [pina.py] Task 2, Epoch [16/30] lr 0.00448 Loss 1.085, Train_accy 74.85, Test_accy 54.37
2025-12-03 00:54:41 [pina.py] Task 2, Epoch [17/30] lr 0.00396 Loss 1.052, Train_accy 75.49, Test_accy 54.31
2025-12-03 00:59:39 [pina.py] Task 2, Epoch [18/30] lr 0.00345 Loss 1.031, Train_accy 76.12, Test_accy 54.26
2025-12-03 01:04:36 [pina.py] Task 2, Epoch [19/30] lr 0.00297 Loss 1.014, Train_accy 76.30, Test_accy 54.10
2025-12-03 01:09:33 [pina.py] Task 2, Epoch [20/30] lr 0.00250 Loss 0.989, Train_accy 76.97, Test_accy 53.99
2025-12-03 01:14:31 [pina.py] Task 2, Epoch [21/30] lr 0.00206 Loss 0.971, Train_accy 77.23, Test_accy 54.56
2025-12-03 01:19:31 [pina.py] Task 2, Epoch [22/30] lr 0.00165 Loss 0.957, Train_accy 77.72, Test_accy 54.36
2025-12-03 01:24:31 [pina.py] Task 2, Epoch [23/30] lr 0.00128 Loss 0.940, Train_accy 77.88, Test_accy 54.37
2025-12-03 01:29:31 [pina.py] Task 2, Epoch [24/30] lr 0.00095 Loss 0.936, Train_accy 77.92, Test_accy 54.36
2025-12-03 01:34:31 [pina.py] Task 2, Epoch [25/30] lr 0.00067 Loss 0.914, Train_accy 78.46, Test_accy 54.34
2025-12-03 01:39:28 [pina.py] Task 2, Epoch [26/30] lr 0.00043 Loss 0.898, Train_accy 78.87, Test_accy 54.20
2025-12-03 01:44:26 [pina.py] Task 2, Epoch [27/30] lr 0.00024 Loss 0.895, Train_accy 79.04, Test_accy 54.46
2025-12-03 01:49:23 [pina.py] Task 2, Epoch [28/30] lr 0.00011 Loss 0.889, Train_accy 79.24, Test_accy 54.51
2025-12-03 01:54:20 [pina.py] Task 2, Epoch [29/30] lr 0.00003 Loss 0.896, Train_accy 78.85, Test_accy 54.61
2025-12-03 01:59:18 [pina.py] Task 2, Epoch [30/30] lr 0.00000 Loss 0.892, Train_accy 79.11, Test_accy 54.59
2025-12-03 01:59:18 [pina.py]
2025-12-03 01:59:18 [pina.py] ==> Start clustering
2025-12-03 02:00:45 [pina.py] clustering features: (50416, 768)
2025-12-03 02:00:57 [pina.py] clustering centers: (5, 768)
2025-12-03 02:11:12 [trainer.py] CNN: {'total': 61.77, '0-344': 68.14, '345-689': 44.01, '690-1034': 70.17, 'old': 55.68, 'new': 70.17}
2025-12-03 02:11:12 [trainer.py] CNN top1 curve: [70.47, 56.03, 61.77]
2025-12-03 02:11:12 [pina.py] Exemplar size: 0
2025-12-03 02:11:13 [trainer.py] Save the checkpoint task_2.pth
2025-12-03 02:11:13 [pina.py]
2025-12-03 02:11:13 [pina.py] ==> Training task 3, Learning on 1035-1380
2025-12-03 02:11:14 [pina.py] len(train_dataset): 120750
2025-12-03 02:11:14 [pina.py] len(test_dataset): 103786
2025-12-03 02:11:14 [pina.py] ==> Checking the parameter
2025-12-03 02:11:14 [pina.py] Total parameters: 93463489
2025-12-03 02:11:14 [pina.py] Trainable parameters: 312000
2025-12-03 02:11:14 [pina.py] Blocks:
2025-12-03 02:11:14 [pina.py] image_encoder: 91326184
2025-12-03 02:11:14 [pina.py] unified_classifier: 265305
2025-12-03 02:11:14 [pina.py] prompt_pool: 46080
2025-12-03 02:11:14 [pina.py] down_pool: 885888
2025-12-03 02:11:14 [pina.py] up_pool: 940032
2025-12-03 02:11:14 [pina.py] Training:
2025-12-03 02:11:14 [pina.py] prompt_pool.3.weight: torch.Size([10, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.0.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.0.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.1.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.1.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.2.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.2.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.3.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.3.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.4.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.4.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.5.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.5.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.6.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.6.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.7.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.7.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.8.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.8.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.9.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.9.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.10.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.10.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] down_pool.3.11.weight: torch.Size([16, 768])
2025-12-03 02:11:14 [pina.py] down_pool.3.11.bias: torch.Size([16])
2025-12-03 02:11:14 [pina.py] up_pool.3.0.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.0.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.1.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.1.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.2.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.2.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.3.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.3.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.4.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.4.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.5.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.5.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.6.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.6.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.7.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.7.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.8.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.8.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.9.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.9.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.10.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.10.bias: torch.Size([768])
2025-12-03 02:11:14 [pina.py] up_pool.3.11.weight: torch.Size([768, 16])
2025-12-03 02:11:14 [pina.py] up_pool.3.11.bias: torch.Size([768])
2025-12-03 02:22:20 [pina.py] Task 3, Epoch [1/30] lr 0.00997 Loss 6.027, Train_accy 0.28, Test_accy 0.30
2025-12-03 02:33:24 [pina.py] Task 3, Epoch [2/30] lr 0.00989 Loss 5.883, Train_accy 0.26, Test_accy 0.24
2025-12-03 02:44:29 [pina.py] Task 3, Epoch [3/30] lr 0.00976 Loss 5.875, Train_accy 0.27, Test_accy 0.36
2025-12-03 02:55:33 [pina.py] Task 3, Epoch [4/30] lr 0.00957 Loss 5.871, Train_accy 0.28, Test_accy 0.32
2025-12-03 03:06:52 [pina.py] Task 3, Epoch [5/30] lr 0.00933 Loss 5.873, Train_accy 0.27, Test_accy 0.44
2025-12-03 03:18:11 [pina.py] Task 3, Epoch [6/30] lr 0.00905 Loss 5.866, Train_accy 0.28, Test_accy 0.36
2025-12-03 03:29:28 [pina.py] Task 3, Epoch [7/30] lr 0.00872 Loss 5.864, Train_accy 0.29, Test_accy 0.26
2025-12-03 03:40:47 [pina.py] Task 3, Epoch [8/30] lr 0.00835 Loss 5.862, Train_accy 0.28, Test_accy 0.30
2025-12-03 03:52:06 [pina.py] Task 3, Epoch [9/30] lr 0.00794 Loss 5.861, Train_accy 0.31, Test_accy 0.27
2025-12-03 04:03:24 [pina.py] Task 3, Epoch [10/30] lr 0.00750 Loss 5.860, Train_accy 0.30, Test_accy 0.25
2025-12-03 04:14:43 [pina.py] Task 3, Epoch [11/30] lr 0.00703 Loss 5.859, Train_accy 0.28, Test_accy 0.39
2025-12-03 04:26:02 [pina.py] Task 3, Epoch [12/30] lr 0.00655 Loss 5.858, Train_accy 0.28, Test_accy 0.29
2025-12-03 04:37:21 [pina.py] Task 3, Epoch [13/30] lr 0.00604 Loss 5.856, Train_accy 0.28, Test_accy 0.22
2025-12-03 04:48:40 [pina.py] Task 3, Epoch [14/30] lr 0.00552 Loss 5.856, Train_accy 0.28, Test_accy 0.36
2025-12-03 04:59:59 [pina.py] Task 3, Epoch [15/30] lr 0.00500 Loss 5.854, Train_accy 0.29, Test_accy 0.28
2025-12-03 05:11:20 [pina.py] Task 3, Epoch [16/30] lr 0.00448 Loss 5.854, Train_accy 0.28, Test_accy 0.34
2025-12-03 05:22:39 [pina.py] Task 3, Epoch [17/30] lr 0.00396 Loss 5.853, Train_accy 0.25, Test_accy 0.23
2025-12-03 05:33:59 [pina.py] Task 3, Epoch [18/30] lr 0.00345 Loss 5.852, Train_accy 0.27, Test_accy 0.25
2025-12-03 05:45:16 [pina.py] Task 3, Epoch [19/30] lr 0.00297 Loss 5.851, Train_accy 0.25, Test_accy 0.33
2025-12-03 05:56:32 [pina.py] Task 3, Epoch [20/30] lr 0.00250 Loss 5.850, Train_accy 0.26, Test_accy 0.27
2025-12-03 06:07:48 [pina.py] Task 3, Epoch [21/30] lr 0.00206 Loss 5.849, Train_accy 0.28, Test_accy 0.23
2025-12-03 06:19:05 [pina.py] Task 3, Epoch [22/30] lr 0.00165 Loss 5.848, Train_accy 0.28, Test_accy 0.23
2025-12-03 06:30:21 [pina.py] Task 3, Epoch [23/30] lr 0.00128 Loss 5.848, Train_accy 0.25, Test_accy 0.26
2025-12-03 06:41:37 [pina.py] Task 3, Epoch [24/30] lr 0.00095 Loss 5.847, Train_accy 0.28, Test_accy 0.44
2025-12-03 06:52:56 [pina.py] Task 3, Epoch [25/30] lr 0.00067 Loss 5.846, Train_accy 0.28, Test_accy 0.28
2025-12-03 07:04:14 [pina.py] Task 3, Epoch [26/30] lr 0.00043 Loss 5.845, Train_accy 0.28, Test_accy 0.33
2025-12-03 07:15:30 [pina.py] Task 3, Epoch [27/30] lr 0.00024 Loss 5.845, Train_accy 0.25, Test_accy 0.28
2025-12-03 07:26:45 [pina.py] Task 3, Epoch [28/30] lr 0.00011 Loss 5.844, Train_accy 0.29, Test_accy 0.30
2025-12-03 07:37:59 [pina.py] Task 3, Epoch [29/30] lr 0.00003 Loss 5.844, Train_accy 0.26, Test_accy 0.24
2025-12-03 07:49:14 [pina.py] Task 3, Epoch [30/30] lr 0.00000 Loss 5.844, Train_accy 0.27, Test_accy 0.24
2025-12-03 07:49:14 [pina.py]
2025-12-03 07:49:14 [pina.py] ==> Start clustering
2025-12-03 07:52:47 [pina.py] clustering features: (120750, 768)
2025-12-03 07:53:10 [pina.py] clustering centers: (5, 768)
2025-12-03 08:13:43 [trainer.py] CNN: {'total': 30.9, '0-344': 66.67, '345-689': 43.99, '690-1034': 70.15, '1035-1379': 0.3, 'old': 61.34, 'new': 0.3}
2025-12-03 08:13:43 [trainer.py] CNN top1 curve: [70.47, 56.03, 61.77, 30.9]
2025-12-03 08:13:43 [pina.py] Exemplar size: 0
2025-12-03 08:13:44 [trainer.py] Save the checkpoint task_3.pth
2025-12-03 08:13:44 [pina.py]
2025-12-03 08:13:44 [pina.py] ==> Training task 4, Learning on 1380-1725
2025-12-03 08:13:45 [pina.py] len(train_dataset): 120906
2025-12-03 08:13:45 [pina.py] len(test_dataset): 155827
2025-12-03 08:13:45 [pina.py] ==> Checking the parameter
2025-12-03 08:13:45 [pina.py] Total parameters: 93463489
2025-12-03 08:13:45 [pina.py] Trainable parameters: 312000
2025-12-03 08:13:45 [pina.py] Blocks:
2025-12-03 08:13:45 [pina.py] image_encoder: 91326184
2025-12-03 08:13:45 [pina.py] unified_classifier: 265305
2025-12-03 08:13:45 [pina.py] prompt_pool: 46080
2025-12-03 08:13:45 [pina.py] down_pool: 885888
2025-12-03 08:13:45 [pina.py] up_pool: 940032
2025-12-03 08:13:45 [pina.py] Training:
2025-12-03 08:13:45 [pina.py] prompt_pool.4.weight: torch.Size([10, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.0.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.0.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.1.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.1.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.2.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.2.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.3.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.3.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.4.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.4.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.5.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.5.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.6.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.6.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.7.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.7.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.8.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.8.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.9.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.9.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.10.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.10.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] down_pool.4.11.weight: torch.Size([16, 768])
2025-12-03 08:13:45 [pina.py] down_pool.4.11.bias: torch.Size([16])
2025-12-03 08:13:45 [pina.py] up_pool.4.0.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.0.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.1.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.1.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.2.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.2.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.3.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.3.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.4.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.4.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.5.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.5.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.6.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.6.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.7.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.7.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.8.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.8.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.9.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.9.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.10.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.10.bias: torch.Size([768])
2025-12-03 08:13:45 [pina.py] up_pool.4.11.weight: torch.Size([768, 16])
2025-12-03 08:13:45 [pina.py] up_pool.4.11.bias: torch.Size([768])
2025-12-03 08:26:50 [pina.py] Task 4, Epoch [1/30] lr 0.00997 Loss 1.357, Train_accy 70.13, Test_accy 43.15
2025-12-03 08:39:59 [pina.py] Task 4, Epoch [2/30] lr 0.00989 Loss 0.979, Train_accy 77.03, Test_accy 44.23
2025-12-03 08:53:07 [pina.py] Task 4, Epoch [3/30] lr 0.00976 Loss 0.900, Train_accy 78.61, Test_accy 44.61
2025-12-03 09:06:15 [pina.py] Task 4, Epoch [4/30] lr 0.00957 Loss 0.838, Train_accy 80.02, Test_accy 45.26
2025-12-03 09:19:20 [pina.py] Task 4, Epoch [5/30] lr 0.00933 Loss 0.806, Train_accy 80.63, Test_accy 45.66
2025-12-03 09:32:24 [pina.py] Task 4, Epoch [6/30] lr 0.00905 Loss 0.779, Train_accy 81.16, Test_accy 45.74
2025-12-03 09:45:32 [pina.py] Task 4, Epoch [7/30] lr 0.00872 Loss 0.751, Train_accy 81.73, Test_accy 46.24
2025-12-03 09:58:37 [pina.py] Task 4, Epoch [8/30] lr 0.00835 Loss 0.731, Train_accy 82.12, Test_accy 46.21
2025-12-03 10:11:41 [pina.py] Task 4, Epoch [9/30] lr 0.00794 Loss 0.707, Train_accy 82.62, Test_accy 46.67
2025-12-03 10:22:59 [pina.py] Task 4, Epoch [10/30] lr 0.00750 Loss 0.692, Train_accy 83.07, Test_accy 45.88
2025-12-03 10:33:40 [pina.py] Task 4, Epoch [11/30] lr 0.00703 Loss 0.679, Train_accy 83.11, Test_accy 46.39
2025-12-03 10:44:28 [pina.py] Task 4, Epoch [12/30] lr 0.00655 Loss 0.664, Train_accy 83.55, Test_accy 46.35
2025-12-03 10:55:16 [pina.py] Task 4, Epoch [13/30] lr 0.00604 Loss 0.650, Train_accy 83.90, Test_accy 46.86
2025-12-03 11:06:03 [pina.py] Task 4, Epoch [14/30] lr 0.00552 Loss 0.637, Train_accy 84.07, Test_accy 46.58
2025-12-03 11:16:48 [pina.py] Task 4, Epoch [15/30] lr 0.00500 Loss 0.618, Train_accy 84.55, Test_accy 46.50
2025-12-03 11:27:38 [pina.py] Task 4, Epoch [16/30] lr 0.00448 Loss 0.608, Train_accy 84.77, Test_accy 46.34
2025-12-03 11:38:28 [pina.py] Task 4, Epoch [17/30] lr 0.00396 Loss 0.601, Train_accy 84.90, Test_accy 47.21
2025-12-03 11:49:07 [pina.py] Task 4, Epoch [18/30] lr 0.00345 Loss 0.592, Train_accy 85.15, Test_accy 46.63
2025-12-03 11:59:50 [pina.py] Task 4, Epoch [19/30] lr 0.00297 Loss 0.572, Train_accy 85.51, Test_accy 46.82
2025-12-03 12:10:30 [pina.py] Task 4, Epoch [20/30] lr 0.00250 Loss 0.571, Train_accy 85.66, Test_accy 47.20
2025-12-03 12:21:08 [pina.py] Task 4, Epoch [21/30] lr 0.00206 Loss 0.559, Train_accy 85.94, Test_accy 46.94
2025-12-03 12:31:51 [pina.py] Task 4, Epoch [22/30] lr 0.00165 Loss 0.545, Train_accy 86.21, Test_accy 46.96
2025-12-03 12:42:28 [pina.py] Task 4, Epoch [23/30] lr 0.00128 Loss 0.537, Train_accy 86.40, Test_accy 47.16
2025-12-03 12:53:07 [pina.py] Task 4, Epoch [24/30] lr 0.00095 Loss 0.533, Train_accy 86.50, Test_accy 46.87
2025-12-03 13:03:46 [pina.py] Task 4, Epoch [25/30] lr 0.00067 Loss 0.524, Train_accy 86.66, Test_accy 46.80
2025-12-03 13:14:23 [pina.py] Task 4, Epoch [26/30] lr 0.00043 Loss 0.526, Train_accy 86.73, Test_accy 47.05
2025-12-03 13:25:05 [pina.py] Task 4, Epoch [27/30] lr 0.00024 Loss 0.516, Train_accy 86.87, Test_accy 46.94
2025-12-03 13:35:48 [pina.py] Task 4, Epoch [28/30] lr 0.00011 Loss 0.512, Train_accy 86.91, Test_accy 47.01
2025-12-03 13:46:27 [pina.py] Task 4, Epoch [29/30] lr 0.00003 Loss 0.514, Train_accy 86.83, Test_accy 47.02
2025-12-03 13:57:19 [pina.py] Task 4, Epoch [30/30] lr 0.00000 Loss 0.511, Train_accy 87.04, Test_accy 47.04
2025-12-03 13:57:19 [pina.py]
2025-12-03 13:57:19 [pina.py] ==> Start clustering
2025-12-03 14:00:11 [pina.py] clustering features: (120906, 768)
2025-12-03 14:00:35 [pina.py] clustering centers: (5, 768)
2025-12-03 14:26:06 [trainer.py] CNN: {'total': 47.44, '0-344': 66.81, '345-689': 44.09, '690-1034': 68.7, '1035-1379': 0.3, '1380-1724': 80.96, 'old': 30.63, 'new': 80.96}
2025-12-03 14:26:06 [trainer.py] CNN top1 curve: [70.47, 56.03, 61.77, 30.9, 47.44]
2025-12-03 14:26:06 [pina.py] Exemplar size: 0
2025-12-03 14:26:07 [trainer.py] Save the checkpoint task_4.pth
2025-12-03 14:26:07 [pina.py]
2025-12-03 14:26:07 [pina.py] ==> Training task 5, Learning on 1725-2070
2025-12-03 14:26:08 [pina.py] len(train_dataset): 48212
2025-12-03 14:26:08 [pina.py] len(test_dataset): 176743
2025-12-03 14:26:08 [pina.py] ==> Checking the parameter
2025-12-03 14:26:08 [pina.py] Total parameters: 93463489
2025-12-03 14:26:08 [pina.py] Trainable parameters: 312000
2025-12-03 14:26:08 [pina.py] Blocks:
2025-12-03 14:26:08 [pina.py] image_encoder: 91326184
2025-12-03 14:26:08 [pina.py] unified_classifier: 265305
2025-12-03 14:26:08 [pina.py] prompt_pool: 46080
2025-12-03 14:26:08 [pina.py] down_pool: 885888
2025-12-03 14:26:08 [pina.py] up_pool: 940032
2025-12-03 14:26:08 [pina.py] Training:
2025-12-03 14:26:08 [pina.py] prompt_pool.5.weight: torch.Size([10, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.0.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.0.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.1.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.1.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.2.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.2.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.3.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.3.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.4.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.4.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.5.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.5.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.6.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.6.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.7.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.7.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.8.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.8.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.9.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.9.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.10.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.10.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] down_pool.5.11.weight: torch.Size([16, 768])
2025-12-03 14:26:08 [pina.py] down_pool.5.11.bias: torch.Size([16])
2025-12-03 14:26:08 [pina.py] up_pool.5.0.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.0.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.1.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.1.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.2.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.2.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.3.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.3.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.4.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.4.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.5.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.5.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.6.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.6.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.7.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.7.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.8.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.8.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.9.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.9.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.10.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.10.bias: torch.Size([768])
2025-12-03 14:26:08 [pina.py] up_pool.5.11.weight: torch.Size([768, 16])
2025-12-03 14:26:08 [pina.py] up_pool.5.11.bias: torch.Size([768])
2025-12-03 14:33:37 [pina.py] Task 5, Epoch [1/30] lr 0.00997 Loss 3.227, Train_accy 38.00, Test_accy 39.11
2025-12-03 14:40:58 [pina.py] Task 5, Epoch [2/30] lr 0.00989 Loss 2.263, Train_accy 52.50, Test_accy 40.65
2025-12-03 14:48:26 [pina.py] Task 5, Epoch [3/30] lr 0.00976 Loss 2.052, Train_accy 55.98, Test_accy 40.72
2025-12-03 14:55:55 [pina.py] Task 5, Epoch [4/30] lr 0.00957 Loss 1.921, Train_accy 58.41, Test_accy 41.25
2025-12-03 15:03:18 [pina.py] Task 5, Epoch [5/30] lr 0.00933 Loss 1.826, Train_accy 60.02, Test_accy 41.98
2025-12-03 15:10:37 [pina.py] Task 5, Epoch [6/30] lr 0.00905 Loss 1.756, Train_accy 61.32, Test_accy 41.19
2025-12-03 15:17:57 [pina.py] Task 5, Epoch [7/30] lr 0.00872 Loss 1.701, Train_accy 62.26, Test_accy 42.10
2025-12-03 15:25:18 [pina.py] Task 5, Epoch [8/30] lr 0.00835 Loss 1.647, Train_accy 63.44, Test_accy 42.27
2025-12-03 15:32:41 [pina.py] Task 5, Epoch [9/30] lr 0.00794 Loss 1.601, Train_accy 64.31, Test_accy 40.87
2025-12-03 15:40:12 [pina.py] Task 5, Epoch [10/30] lr 0.00750 Loss 1.573, Train_accy 64.83, Test_accy 42.23
2025-12-03 15:47:43 [pina.py] Task 5, Epoch [11/30] lr 0.00703 Loss 1.525, Train_accy 65.79, Test_accy 41.25
2025-12-03 15:55:09 [pina.py] Task 5, Epoch [12/30] lr 0.00655 Loss 1.494, Train_accy 66.40, Test_accy 42.74
2025-12-03 16:02:35 [pina.py] Task 5, Epoch [13/30] lr 0.00604 Loss 1.461, Train_accy 66.86, Test_accy 42.38
2025-12-03 16:10:09 [pina.py] Task 5, Epoch [14/30] lr 0.00552 Loss 1.418, Train_accy 67.77, Test_accy 42.83
2025-12-03 16:17:39 [pina.py] Task 5, Epoch [15/30] lr 0.00500 Loss 1.395, Train_accy 68.27, Test_accy 41.99
2025-12-03 16:25:10 [pina.py] Task 5, Epoch [16/30] lr 0.00448 Loss 1.370, Train_accy 68.77, Test_accy 42.69
2025-12-03 16:32:45 [pina.py] Task 5, Epoch [17/30] lr 0.00396 Loss 1.345, Train_accy 69.17, Test_accy 43.00
2025-12-03 16:40:19 [pina.py] Task 5, Epoch [18/30] lr 0.00345 Loss 1.320, Train_accy 69.61, Test_accy 43.25
2025-12-03 16:47:53 [pina.py] Task 5, Epoch [19/30] lr 0.00297 Loss 1.284, Train_accy 70.40, Test_accy 42.33
2025-12-03 16:55:19 [pina.py] Task 5, Epoch [20/30] lr 0.00250 Loss 1.270, Train_accy 70.58, Test_accy 43.15
2025-12-03 17:02:45 [pina.py] Task 5, Epoch [21/30] lr 0.00206 Loss 1.238, Train_accy 71.23, Test_accy 42.51
2025-12-03 17:10:11 [pina.py] Task 5, Epoch [22/30] lr 0.00165 Loss 1.232, Train_accy 71.50, Test_accy 42.97
2025-12-03 17:17:45 [pina.py] Task 5, Epoch [23/30] lr 0.00128 Loss 1.212, Train_accy 71.76, Test_accy 42.06
2025-12-03 17:25:11 [pina.py] Task 5, Epoch [24/30] lr 0.00095 Loss 1.205, Train_accy 72.07, Test_accy 42.33
2025-12-03 17:32:45 [pina.py] Task 5, Epoch [25/30] lr 0.00067 Loss 1.186, Train_accy 72.45, Test_accy 42.47
2025-12-03 17:40:21 [pina.py] Task 5, Epoch [26/30] lr 0.00043 Loss 1.177, Train_accy 72.60, Test_accy 42.66
2025-12-03 17:47:55 [pina.py] Task 5, Epoch [27/30] lr 0.00024 Loss 1.156, Train_accy 73.13, Test_accy 42.53
2025-12-03 17:55:30 [pina.py] Task 5, Epoch [28/30] lr 0.00011 Loss 1.169, Train_accy 72.80, Test_accy 42.59
2025-12-03 18:03:01 [pina.py] Task 5, Epoch [29/30] lr 0.00003 Loss 1.159, Train_accy 72.83, Test_accy 42.43
2025-12-03 18:10:36 [pina.py] Task 5, Epoch [30/30] lr 0.00000 Loss 1.153, Train_accy 73.15, Test_accy 42.41
2025-12-03 18:10:36 [pina.py]
2025-12-03 18:10:36 [pina.py] ==> Start clustering
2025-12-03 18:11:45 [pina.py] clustering features: (48212, 768)
2025-12-03 18:11:56 [pina.py] clustering centers: (5, 768)
2025-12-03 18:40:49 [trainer.py] CNN: {'total': 49.57, '0-344': 68.09, '345-689': 44.17, '690-1034': 68.4, '1035-1379': 0.34, '1380-1724': 80.88, '1725-2069': 64.85, 'old': 47.51, 'new': 64.85}
2025-12-03 18:40:49 [trainer.py] CNN top1 curve: [70.47, 56.03, 61.77, 30.9, 47.44, 49.57]
2025-12-03 18:40:49 [pina.py] Exemplar size: 0
2025-12-03 18:40:50 [trainer.py] Save the checkpoint task_5.pth

您好，这是我按照步骤复现出的DomainNet上的实验结果，与原论文中结果不一致。您可以解答一下吗？

可以联系我吗？wechat：1046960908
谢谢您！

祝好

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions