[2021-03-22 14:25:29] Created experiment 0:
[2021-03-22 14:25:29]  - Model: vgg-pt
[2021-03-22 14:25:29]  - Acquisition function: random
[2021-03-22 14:25:29] Created experiment 1:
[2021-03-22 14:25:29]  - Model: vgg-pt
[2021-03-22 14:25:29]  - Acquisition function: greedy-coreset
[2021-03-22 14:25:29] Created experiment 2:
[2021-03-22 14:25:29]  - Model: vgg-pt
[2021-03-22 14:25:29]  - Acquisition function: lc-beam-pweighted-coreset (beams=10)
[2021-03-22 14:25:29] Loading cifar100 test set...
[2021-03-22 14:25:30] Experiment repeat 1/5
[2021-03-22 14:25:30] Seeded: 5
[2021-03-22 14:25:30] Using 10.00% labels of the dataset (5000/50000)
[2021-03-22 14:25:30] Randomly labelled 5000/50000
[2021-03-22 14:25:30] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 14:25:30] Seeded: 6
[2021-03-22 14:25:30] Running: experiment 0
[2021-03-22 14:25:30] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 14:25:30] Creating pretrained=True VGG16...
[2021-03-22 14:25:32] No parameter reset since we are using a pretrained model.
[2021-03-22 14:25:32] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 14:25:32] Creating trainer with model on device: cuda
[2021-03-22 14:25:37] Training with expected score >= 0.9900
[2021-03-22 14:25:37] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 14:30:35] Training accuracy: 0.9950
[2021-03-22 14:30:35] Testing on 10000 data points...
[2021-03-22 14:30:37] Test score for 5000 training labels: 0.4834
[2021-03-22 14:30:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 14:30:37] Found 50000 unlabelled features.
[2021-03-22 14:30:43] Showing labelled: True (9512/50000 visible, 488 redundant)
[2021-03-22 14:30:43] Creating trainer with model on device: cuda
[2021-03-22 14:30:43] Training with expected score >= 0.9900
[2021-03-22 14:30:43] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 14:33:13] Training accuracy: 0.9922
[2021-03-22 14:33:13] Testing on 10000 data points...
[2021-03-22 14:33:16] Test score for 10000 training labels: 0.5246
[2021-03-22 14:33:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 14:33:16] Found 50000 unlabelled features.
[2021-03-22 14:33:21] Showing labelled: True (13552/50000 visible, 1448 redundant)
[2021-03-22 14:33:21] Creating trainer with model on device: cuda
[2021-03-22 14:33:21] Training with expected score >= 0.9900
[2021-03-22 14:33:21] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 14:37:09] Training accuracy: 0.9955
[2021-03-22 14:37:09] Testing on 10000 data points...
[2021-03-22 14:37:11] Test score for 15000 training labels: 0.5553
[2021-03-22 14:37:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 14:37:11] Found 50000 unlabelled features.
[2021-03-22 14:37:17] Showing labelled: True (17219/50000 visible, 2781 redundant)
[2021-03-22 14:37:17] Creating trainer with model on device: cuda
[2021-03-22 14:37:17] Training with expected score >= 0.9900
[2021-03-22 14:37:17] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 14:42:18] Training accuracy: 0.9987
[2021-03-22 14:42:18] Testing on 10000 data points...
[2021-03-22 14:42:20] Test score for 20000 training labels: 0.5726
[2021-03-22 14:42:20] Seeded: 6
[2021-03-22 14:42:20] Running: experiment 1
[2021-03-22 14:42:20] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 14:42:20] Creating pretrained=True VGG16...
[2021-03-22 14:42:22] No parameter reset since we are using a pretrained model.
[2021-03-22 14:42:22] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 14:42:22] Creating trainer with model on device: cuda
[2021-03-22 14:42:22] Training with expected score >= 0.9900
[2021-03-22 14:42:22] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 14:47:46] Training accuracy: 0.9963
[2021-03-22 14:47:46] Testing on 10000 data points...
[2021-03-22 14:47:48] Test score for 5000 training labels: 0.4902
[2021-03-22 14:47:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 14:47:48] Found 50000 unlabelled features.
[2021-03-22 14:47:59] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 14:48:14] Searching for coresets greedily...
[2021-03-22 14:48:24] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 14:48:24] Creating trainer with model on device: cuda
[2021-03-22 14:48:24] Training with expected score >= 0.9900
[2021-03-22 14:48:24] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 14:50:54] Training accuracy: 0.9919
[2021-03-22 14:50:54] Testing on 10000 data points...
[2021-03-22 14:50:56] Test score for 10000 training labels: 0.5394
[2021-03-22 14:50:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 14:50:56] Found 50000 unlabelled features.
[2021-03-22 14:51:07] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 14:51:37] Searching for coresets greedily...
[2021-03-22 14:51:46] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 14:51:46] Creating trainer with model on device: cuda
[2021-03-22 14:51:46] Training with expected score >= 0.9900
[2021-03-22 14:51:46] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 14:55:32] Training accuracy: 0.9956
[2021-03-22 14:55:32] Testing on 10000 data points...
[2021-03-22 14:55:34] Test score for 15000 training labels: 0.5726
[2021-03-22 14:55:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 14:55:34] Found 50000 unlabelled features.
[2021-03-22 14:55:46] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 14:56:30] Searching for coresets greedily...
[2021-03-22 14:56:39] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 14:56:39] Creating trainer with model on device: cuda
[2021-03-22 14:56:39] Training with expected score >= 0.9900
[2021-03-22 14:56:39] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 15:01:37] Training accuracy: 0.9955
[2021-03-22 15:01:37] Testing on 10000 data points...
[2021-03-22 15:01:39] Test score for 20000 training labels: 0.5947
[2021-03-22 15:01:40] Seeded: 6
[2021-03-22 15:01:40] Running: experiment 2
[2021-03-22 15:01:40] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 15:01:40] Creating pretrained=True VGG16...
[2021-03-22 15:01:41] No parameter reset since we are using a pretrained model.
[2021-03-22 15:01:41] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 15:01:41] Creating trainer with model on device: cuda
[2021-03-22 15:01:41] Training with expected score >= 0.9900
[2021-03-22 15:01:41] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 15:06:49] Training accuracy: 0.9969
[2021-03-22 15:06:49] Testing on 10000 data points...
[2021-03-22 15:06:50] Test score for 5000 training labels: 0.4887
[2021-03-22 15:06:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:06:50] Found 50000 unlabelled features.
[2021-03-22 15:07:01] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 15:07:16] Searching for coresets with 10 beams...
[2021-03-22 15:09:55] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 15:09:55] Creating trainer with model on device: cuda
[2021-03-22 15:09:55] Training with expected score >= 0.9900
[2021-03-22 15:09:55] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 15:12:25] Training accuracy: 0.9889
[2021-03-22 15:12:25] Training score (0.9889) was below expectations. Retraining...
[2021-03-22 15:12:25] Creating trainer with model on device: cuda
[2021-03-22 15:12:25] Training with expected score >= 0.9900
[2021-03-22 15:12:25] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 15:14:55] Training accuracy: 0.9980
[2021-03-22 15:14:55] Testing on 10000 data points...
[2021-03-22 15:14:57] Test score for 10000 training labels: 0.5477
[2021-03-22 15:14:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:14:57] Found 50000 unlabelled features.
[2021-03-22 15:15:11] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 15:15:41] Searching for coresets with 10 beams...
[2021-03-22 15:18:19] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 15:18:19] Creating trainer with model on device: cuda
[2021-03-22 15:18:19] Training with expected score >= 0.9900
[2021-03-22 15:18:19] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 15:22:05] Training accuracy: 0.9935
[2021-03-22 15:22:05] Testing on 10000 data points...
[2021-03-22 15:22:06] Test score for 15000 training labels: 0.5798
[2021-03-22 15:22:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:22:06] Found 50000 unlabelled features.
[2021-03-22 15:22:20] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 15:23:05] Searching for coresets with 10 beams...
[2021-03-22 15:25:44] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 15:25:44] Creating trainer with model on device: cuda
[2021-03-22 15:25:44] Training with expected score >= 0.9900
[2021-03-22 15:25:44] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 15:30:41] Training accuracy: 0.9979
[2021-03-22 15:30:41] Testing on 10000 data points...
[2021-03-22 15:30:43] Test score for 20000 training labels: 0.6041
[2021-03-22 15:30:43] Experiment repeat 2/5
[2021-03-22 15:30:43] Seeded: 6
[2021-03-22 15:30:43] Using 10.00% labels of the dataset (5000/50000)
[2021-03-22 15:30:43] Randomly labelled 5000/50000
[2021-03-22 15:30:43] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 15:30:43] Seeded: 7
[2021-03-22 15:30:43] Running: experiment 0
[2021-03-22 15:30:43] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 15:30:43] Creating pretrained=True VGG16...
[2021-03-22 15:30:45] No parameter reset since we are using a pretrained model.
[2021-03-22 15:30:45] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 15:30:45] Creating trainer with model on device: cuda
[2021-03-22 15:30:45] Training with expected score >= 0.9900
[2021-03-22 15:30:45] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 15:35:43] Training accuracy: 0.9964
[2021-03-22 15:35:43] Testing on 10000 data points...
[2021-03-22 15:35:45] Test score for 5000 training labels: 0.4877
[2021-03-22 15:35:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:35:45] Found 50000 unlabelled features.
[2021-03-22 15:35:52] Showing labelled: True (9529/50000 visible, 471 redundant)
[2021-03-22 15:35:52] Creating trainer with model on device: cuda
[2021-03-22 15:35:52] Training with expected score >= 0.9900
[2021-03-22 15:35:52] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 15:38:23] Training accuracy: 0.9896
[2021-03-22 15:38:23] Training score (0.9896) was below expectations. Retraining...
[2021-03-22 15:38:23] Creating trainer with model on device: cuda
[2021-03-22 15:38:23] Training with expected score >= 0.9900
[2021-03-22 15:38:23] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 15:40:53] Training accuracy: 0.9975
[2021-03-22 15:40:53] Testing on 10000 data points...
[2021-03-22 15:40:55] Test score for 10000 training labels: 0.5354
[2021-03-22 15:40:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:40:55] Found 50000 unlabelled features.
[2021-03-22 15:41:01] Showing labelled: True (13568/50000 visible, 1432 redundant)
[2021-03-22 15:41:01] Creating trainer with model on device: cuda
[2021-03-22 15:41:01] Training with expected score >= 0.9900
[2021-03-22 15:41:01] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 15:44:46] Training accuracy: 0.9985
[2021-03-22 15:44:46] Testing on 10000 data points...
[2021-03-22 15:44:48] Test score for 15000 training labels: 0.5567
[2021-03-22 15:44:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:44:48] Found 50000 unlabelled features.
[2021-03-22 15:44:55] Showing labelled: True (17156/50000 visible, 2844 redundant)
[2021-03-22 15:44:55] Creating trainer with model on device: cuda
[2021-03-22 15:44:55] Training with expected score >= 0.9900
[2021-03-22 15:44:55] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 15:49:54] Training accuracy: 0.9991
[2021-03-22 15:49:54] Testing on 10000 data points...
[2021-03-22 15:49:56] Test score for 20000 training labels: 0.5725
[2021-03-22 15:49:56] Seeded: 7
[2021-03-22 15:49:56] Running: experiment 1
[2021-03-22 15:49:56] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 15:49:56] Creating pretrained=True VGG16...
[2021-03-22 15:49:57] No parameter reset since we are using a pretrained model.
[2021-03-22 15:49:57] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 15:49:57] Creating trainer with model on device: cuda
[2021-03-22 15:49:57] Training with expected score >= 0.9900
[2021-03-22 15:49:57] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 15:54:57] Training accuracy: 0.9934
[2021-03-22 15:54:57] Testing on 10000 data points...
[2021-03-22 15:54:59] Test score for 5000 training labels: 0.4871
[2021-03-22 15:54:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:54:59] Found 50000 unlabelled features.
[2021-03-22 15:55:11] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 15:55:25] Searching for coresets greedily...
[2021-03-22 15:55:35] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 15:55:35] Creating trainer with model on device: cuda
[2021-03-22 15:55:35] Training with expected score >= 0.9900
[2021-03-22 15:55:35] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 15:58:06] Training accuracy: 0.9914
[2021-03-22 15:58:06] Testing on 10000 data points...
[2021-03-22 15:58:08] Test score for 10000 training labels: 0.5354
[2021-03-22 15:58:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 15:58:08] Found 50000 unlabelled features.
[2021-03-22 15:58:19] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 15:58:49] Searching for coresets greedily...
[2021-03-22 15:58:59] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 15:58:59] Creating trainer with model on device: cuda
[2021-03-22 15:58:59] Training with expected score >= 0.9900
[2021-03-22 15:58:59] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 16:02:42] Training accuracy: 0.9935
[2021-03-22 16:02:42] Testing on 10000 data points...
[2021-03-22 16:02:44] Test score for 15000 training labels: 0.5683
[2021-03-22 16:02:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:02:44] Found 50000 unlabelled features.
[2021-03-22 16:02:58] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 16:03:43] Searching for coresets greedily...
[2021-03-22 16:03:53] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 16:03:53] Creating trainer with model on device: cuda
[2021-03-22 16:03:53] Training with expected score >= 0.9900
[2021-03-22 16:03:53] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 16:08:51] Training accuracy: 0.9983
[2021-03-22 16:08:51] Testing on 10000 data points...
[2021-03-22 16:08:53] Test score for 20000 training labels: 0.5965
[2021-03-22 16:08:53] Seeded: 7
[2021-03-22 16:08:53] Running: experiment 2
[2021-03-22 16:08:53] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 16:08:53] Creating pretrained=True VGG16...
[2021-03-22 16:08:55] No parameter reset since we are using a pretrained model.
[2021-03-22 16:08:55] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 16:08:55] Creating trainer with model on device: cuda
[2021-03-22 16:08:55] Training with expected score >= 0.9900
[2021-03-22 16:08:55] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 16:13:55] Training accuracy: 0.9932
[2021-03-22 16:13:55] Testing on 10000 data points...
[2021-03-22 16:13:57] Test score for 5000 training labels: 0.4822
[2021-03-22 16:13:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:13:57] Found 50000 unlabelled features.
[2021-03-22 16:14:09] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 16:14:24] Searching for coresets with 10 beams...
[2021-03-22 16:17:03] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 16:17:03] Creating trainer with model on device: cuda
[2021-03-22 16:17:03] Training with expected score >= 0.9900
[2021-03-22 16:17:03] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 16:19:33] Training accuracy: 0.9851
[2021-03-22 16:19:33] Training score (0.9851) was below expectations. Retraining...
[2021-03-22 16:19:33] Creating trainer with model on device: cuda
[2021-03-22 16:19:33] Training with expected score >= 0.9900
[2021-03-22 16:19:33] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 16:22:02] Training accuracy: 0.9978
[2021-03-22 16:22:02] Testing on 10000 data points...
[2021-03-22 16:22:04] Test score for 10000 training labels: 0.5414
[2021-03-22 16:22:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:22:04] Found 50000 unlabelled features.
[2021-03-22 16:22:17] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 16:22:47] Searching for coresets with 10 beams...
[2021-03-22 16:25:25] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 16:25:25] Creating trainer with model on device: cuda
[2021-03-22 16:25:25] Training with expected score >= 0.9900
[2021-03-22 16:25:25] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 16:29:11] Training accuracy: 0.9937
[2021-03-22 16:29:11] Testing on 10000 data points...
[2021-03-22 16:29:13] Test score for 15000 training labels: 0.5702
[2021-03-22 16:29:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:29:13] Found 50000 unlabelled features.
[2021-03-22 16:29:27] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 16:30:08] Searching for coresets with 10 beams...
[2021-03-22 16:32:47] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 16:32:47] Creating trainer with model on device: cuda
[2021-03-22 16:32:47] Training with expected score >= 0.9900
[2021-03-22 16:32:47] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 16:37:45] Training accuracy: 0.9969
[2021-03-22 16:37:45] Testing on 10000 data points...
[2021-03-22 16:37:47] Test score for 20000 training labels: 0.5956
[2021-03-22 16:37:47] Experiment repeat 3/5
[2021-03-22 16:37:47] Seeded: 7
[2021-03-22 16:37:47] Using 10.00% labels of the dataset (5000/50000)
[2021-03-22 16:37:47] Randomly labelled 5000/50000
[2021-03-22 16:37:47] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 16:37:47] Seeded: 8
[2021-03-22 16:37:47] Running: experiment 0
[2021-03-22 16:37:47] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 16:37:47] Creating pretrained=True VGG16...
[2021-03-22 16:37:48] No parameter reset since we are using a pretrained model.
[2021-03-22 16:37:48] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 16:37:48] Creating trainer with model on device: cuda
[2021-03-22 16:37:48] Training with expected score >= 0.9900
[2021-03-22 16:37:48] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 16:42:48] Training accuracy: 0.9974
[2021-03-22 16:42:48] Testing on 10000 data points...
[2021-03-22 16:42:49] Test score for 5000 training labels: 0.4905
[2021-03-22 16:42:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:42:49] Found 50000 unlabelled features.
[2021-03-22 16:42:56] Showing labelled: True (9506/50000 visible, 494 redundant)
[2021-03-22 16:42:56] Creating trainer with model on device: cuda
[2021-03-22 16:42:56] Training with expected score >= 0.9900
[2021-03-22 16:42:56] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 16:45:26] Training accuracy: 0.9923
[2021-03-22 16:45:26] Testing on 10000 data points...
[2021-03-22 16:45:28] Test score for 10000 training labels: 0.5359
[2021-03-22 16:45:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:45:28] Found 50000 unlabelled features.
[2021-03-22 16:45:34] Showing labelled: True (13561/50000 visible, 1439 redundant)
[2021-03-22 16:45:34] Creating trainer with model on device: cuda
[2021-03-22 16:45:34] Training with expected score >= 0.9900
[2021-03-22 16:45:34] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 16:49:19] Training accuracy: 0.9964
[2021-03-22 16:49:19] Testing on 10000 data points...
[2021-03-22 16:49:21] Test score for 15000 training labels: 0.5597
[2021-03-22 16:49:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:49:21] Found 50000 unlabelled features.
[2021-03-22 16:49:28] Showing labelled: True (17247/50000 visible, 2753 redundant)
[2021-03-22 16:49:28] Creating trainer with model on device: cuda
[2021-03-22 16:49:28] Training with expected score >= 0.9900
[2021-03-22 16:49:28] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 16:54:28] Training accuracy: 0.9989
[2021-03-22 16:54:28] Testing on 10000 data points...
[2021-03-22 16:54:30] Test score for 20000 training labels: 0.5737
[2021-03-22 16:54:30] Seeded: 8
[2021-03-22 16:54:30] Running: experiment 1
[2021-03-22 16:54:30] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 16:54:30] Creating pretrained=True VGG16...
[2021-03-22 16:54:31] No parameter reset since we are using a pretrained model.
[2021-03-22 16:54:31] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 16:54:31] Creating trainer with model on device: cuda
[2021-03-22 16:54:31] Training with expected score >= 0.9900
[2021-03-22 16:54:31] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 16:59:30] Training accuracy: 0.9965
[2021-03-22 16:59:30] Testing on 10000 data points...
[2021-03-22 16:59:32] Test score for 5000 training labels: 0.4925
[2021-03-22 16:59:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 16:59:32] Found 50000 unlabelled features.
[2021-03-22 16:59:43] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 16:59:58] Searching for coresets greedily...
[2021-03-22 17:00:08] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 17:00:08] Creating trainer with model on device: cuda
[2021-03-22 17:00:08] Training with expected score >= 0.9900
[2021-03-22 17:00:08] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 17:02:38] Training accuracy: 0.9905
[2021-03-22 17:02:38] Testing on 10000 data points...
[2021-03-22 17:02:40] Test score for 10000 training labels: 0.5381
[2021-03-22 17:02:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:02:40] Found 50000 unlabelled features.
[2021-03-22 17:02:53] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 17:03:20] Searching for coresets greedily...
[2021-03-22 17:03:30] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 17:03:30] Creating trainer with model on device: cuda
[2021-03-22 17:03:30] Training with expected score >= 0.9900
[2021-03-22 17:03:30] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 17:07:17] Training accuracy: 0.9939
[2021-03-22 17:07:17] Testing on 10000 data points...
[2021-03-22 17:07:19] Test score for 15000 training labels: 0.5706
[2021-03-22 17:07:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:07:19] Found 50000 unlabelled features.
[2021-03-22 17:07:32] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 17:08:17] Searching for coresets greedily...
[2021-03-22 17:08:26] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 17:08:26] Creating trainer with model on device: cuda
[2021-03-22 17:08:26] Training with expected score >= 0.9900
[2021-03-22 17:08:26] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 17:13:25] Training accuracy: 0.9969
[2021-03-22 17:13:25] Testing on 10000 data points...
[2021-03-22 17:13:27] Test score for 20000 training labels: 0.5906
[2021-03-22 17:13:27] Seeded: 8
[2021-03-22 17:13:27] Running: experiment 2
[2021-03-22 17:13:27] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 17:13:27] Creating pretrained=True VGG16...
[2021-03-22 17:13:28] No parameter reset since we are using a pretrained model.
[2021-03-22 17:13:28] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 17:13:28] Creating trainer with model on device: cuda
[2021-03-22 17:13:28] Training with expected score >= 0.9900
[2021-03-22 17:13:28] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 17:18:28] Training accuracy: 0.9932
[2021-03-22 17:18:28] Testing on 10000 data points...
[2021-03-22 17:18:30] Test score for 5000 training labels: 0.4905
[2021-03-22 17:18:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:18:30] Found 50000 unlabelled features.
[2021-03-22 17:18:41] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 17:18:55] Searching for coresets with 10 beams...
[2021-03-22 17:21:33] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 17:21:33] Creating trainer with model on device: cuda
[2021-03-22 17:21:33] Training with expected score >= 0.9900
[2021-03-22 17:21:33] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 17:24:03] Training accuracy: 0.9877
[2021-03-22 17:24:03] Training score (0.9877) was below expectations. Retraining...
[2021-03-22 17:24:03] Creating trainer with model on device: cuda
[2021-03-22 17:24:03] Training with expected score >= 0.9900
[2021-03-22 17:24:03] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 17:26:33] Training accuracy: 0.9976
[2021-03-22 17:26:33] Testing on 10000 data points...
[2021-03-22 17:26:35] Test score for 10000 training labels: 0.5491
[2021-03-22 17:26:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:26:35] Found 50000 unlabelled features.
[2021-03-22 17:26:48] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 17:27:18] Searching for coresets with 10 beams...
[2021-03-22 17:29:56] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 17:29:56] Creating trainer with model on device: cuda
[2021-03-22 17:29:56] Training with expected score >= 0.9900
[2021-03-22 17:29:56] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 17:33:43] Training accuracy: 0.9952
[2021-03-22 17:33:43] Testing on 10000 data points...
[2021-03-22 17:33:45] Test score for 15000 training labels: 0.5712
[2021-03-22 17:33:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:33:45] Found 50000 unlabelled features.
[2021-03-22 17:33:58] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 17:34:40] Searching for coresets with 10 beams...
[2021-03-22 17:37:19] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 17:37:19] Creating trainer with model on device: cuda
[2021-03-22 17:37:19] Training with expected score >= 0.9900
[2021-03-22 17:37:19] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 17:42:19] Training accuracy: 0.9977
[2021-03-22 17:42:19] Testing on 10000 data points...
[2021-03-22 17:42:21] Test score for 20000 training labels: 0.5968
[2021-03-22 17:42:21] Experiment repeat 4/5
[2021-03-22 17:42:21] Seeded: 8
[2021-03-22 17:42:21] Using 10.00% labels of the dataset (5000/50000)
[2021-03-22 17:42:21] Randomly labelled 5000/50000
[2021-03-22 17:42:21] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 17:42:21] Seeded: 9
[2021-03-22 17:42:21] Running: experiment 0
[2021-03-22 17:42:21] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 17:42:21] Creating pretrained=True VGG16...
[2021-03-22 17:42:23] No parameter reset since we are using a pretrained model.
[2021-03-22 17:42:23] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 17:42:23] Creating trainer with model on device: cuda
[2021-03-22 17:42:23] Training with expected score >= 0.9900
[2021-03-22 17:42:23] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 17:47:21] Training accuracy: 0.9965
[2021-03-22 17:47:21] Testing on 10000 data points...
[2021-03-22 17:47:23] Test score for 5000 training labels: 0.4974
[2021-03-22 17:47:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:47:24] Found 50000 unlabelled features.
[2021-03-22 17:47:30] Showing labelled: True (9469/50000 visible, 531 redundant)
[2021-03-22 17:47:30] Creating trainer with model on device: cuda
[2021-03-22 17:47:30] Training with expected score >= 0.9900
[2021-03-22 17:47:30] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 17:50:00] Training accuracy: 0.9923
[2021-03-22 17:50:00] Testing on 10000 data points...
[2021-03-22 17:50:02] Test score for 10000 training labels: 0.5366
[2021-03-22 17:50:02] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:50:02] Found 50000 unlabelled features.
[2021-03-22 17:50:08] Showing labelled: True (13522/50000 visible, 1478 redundant)
[2021-03-22 17:50:08] Creating trainer with model on device: cuda
[2021-03-22 17:50:08] Training with expected score >= 0.9900
[2021-03-22 17:50:08] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 17:53:56] Training accuracy: 0.9961
[2021-03-22 17:53:56] Testing on 10000 data points...
[2021-03-22 17:53:59] Test score for 15000 training labels: 0.5557
[2021-03-22 17:53:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 17:53:59] Found 50000 unlabelled features.
[2021-03-22 17:54:05] Showing labelled: True (17159/50000 visible, 2841 redundant)
[2021-03-22 17:54:05] Creating trainer with model on device: cuda
[2021-03-22 17:54:05] Training with expected score >= 0.9900
[2021-03-22 17:54:05] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 17:59:05] Training accuracy: 0.9988
[2021-03-22 17:59:05] Testing on 10000 data points...
[2021-03-22 17:59:07] Test score for 20000 training labels: 0.5798
[2021-03-22 17:59:07] Seeded: 9
[2021-03-22 17:59:07] Running: experiment 1
[2021-03-22 17:59:07] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 17:59:07] Creating pretrained=True VGG16...
[2021-03-22 17:59:09] No parameter reset since we are using a pretrained model.
[2021-03-22 17:59:09] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 17:59:09] Creating trainer with model on device: cuda
[2021-03-22 17:59:09] Training with expected score >= 0.9900
[2021-03-22 17:59:09] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 18:04:08] Training accuracy: 0.9968
[2021-03-22 18:04:08] Testing on 10000 data points...
[2021-03-22 18:04:10] Test score for 5000 training labels: 0.4940
[2021-03-22 18:04:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:04:11] Found 50000 unlabelled features.
[2021-03-22 18:04:22] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 18:04:37] Searching for coresets greedily...
[2021-03-22 18:04:47] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 18:04:47] Creating trainer with model on device: cuda
[2021-03-22 18:04:47] Training with expected score >= 0.9900
[2021-03-22 18:04:47] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 18:07:18] Training accuracy: 0.9904
[2021-03-22 18:07:18] Testing on 10000 data points...
[2021-03-22 18:07:20] Test score for 10000 training labels: 0.5455
[2021-03-22 18:07:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:07:21] Found 50000 unlabelled features.
[2021-03-22 18:07:33] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 18:08:03] Searching for coresets greedily...
[2021-03-22 18:08:13] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 18:08:13] Creating trainer with model on device: cuda
[2021-03-22 18:08:13] Training with expected score >= 0.9900
[2021-03-22 18:08:13] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 18:12:03] Training accuracy: 0.9936
[2021-03-22 18:12:03] Testing on 10000 data points...
[2021-03-22 18:12:06] Test score for 15000 training labels: 0.5745
[2021-03-22 18:12:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:12:06] Found 50000 unlabelled features.
[2021-03-22 18:12:18] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 18:13:03] Searching for coresets greedily...
[2021-03-22 18:13:12] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 18:13:12] Creating trainer with model on device: cuda
[2021-03-22 18:13:12] Training with expected score >= 0.9900
[2021-03-22 18:13:12] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 18:18:20] Training accuracy: 0.9950
[2021-03-22 18:18:20] Testing on 10000 data points...
[2021-03-22 18:18:22] Test score for 20000 training labels: 0.6000
[2021-03-22 18:18:22] Seeded: 9
[2021-03-22 18:18:22] Running: experiment 2
[2021-03-22 18:18:22] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 18:18:22] Creating pretrained=True VGG16...
[2021-03-22 18:18:23] No parameter reset since we are using a pretrained model.
[2021-03-22 18:18:23] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 18:18:23] Creating trainer with model on device: cuda
[2021-03-22 18:18:23] Training with expected score >= 0.9900
[2021-03-22 18:18:23] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 18:23:25] Training accuracy: 0.9990
[2021-03-22 18:23:25] Testing on 10000 data points...
[2021-03-22 18:23:27] Test score for 5000 training labels: 0.4894
[2021-03-22 18:23:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:23:27] Found 50000 unlabelled features.
[2021-03-22 18:23:39] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 18:23:53] Searching for coresets with 10 beams...
[2021-03-22 18:26:31] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 18:26:31] Creating trainer with model on device: cuda
[2021-03-22 18:26:31] Training with expected score >= 0.9900
[2021-03-22 18:26:31] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 18:29:01] Training accuracy: 0.9866
[2021-03-22 18:29:01] Training score (0.9866) was below expectations. Retraining...
[2021-03-22 18:29:01] Creating trainer with model on device: cuda
[2021-03-22 18:29:01] Training with expected score >= 0.9900
[2021-03-22 18:29:01] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 18:31:30] Training accuracy: 0.9969
[2021-03-22 18:31:30] Testing on 10000 data points...
[2021-03-22 18:31:33] Test score for 10000 training labels: 0.5445
[2021-03-22 18:31:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:31:33] Found 50000 unlabelled features.
[2021-03-22 18:31:45] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 18:32:13] Searching for coresets with 10 beams...
[2021-03-22 18:34:51] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 18:34:51] Creating trainer with model on device: cuda
[2021-03-22 18:34:51] Training with expected score >= 0.9900
[2021-03-22 18:34:51] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 18:38:36] Training accuracy: 0.9952
[2021-03-22 18:38:36] Testing on 10000 data points...
[2021-03-22 18:38:38] Test score for 15000 training labels: 0.5764
[2021-03-22 18:38:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:38:38] Found 50000 unlabelled features.
[2021-03-22 18:38:51] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 18:39:33] Searching for coresets with 10 beams...
[2021-03-22 18:42:12] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 18:42:12] Creating trainer with model on device: cuda
[2021-03-22 18:42:12] Training with expected score >= 0.9900
[2021-03-22 18:42:12] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 18:47:10] Training accuracy: 0.9965
[2021-03-22 18:47:10] Testing on 10000 data points...
[2021-03-22 18:47:12] Test score for 20000 training labels: 0.5980
[2021-03-22 18:47:12] Experiment repeat 5/5
[2021-03-22 18:47:12] Seeded: 9
[2021-03-22 18:47:12] Using 10.00% labels of the dataset (5000/50000)
[2021-03-22 18:47:12] Randomly labelled 5000/50000
[2021-03-22 18:47:12] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 18:47:12] Seeded: 10
[2021-03-22 18:47:12] Running: experiment 0
[2021-03-22 18:47:12] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 18:47:12] Creating pretrained=True VGG16...
[2021-03-22 18:47:14] No parameter reset since we are using a pretrained model.
[2021-03-22 18:47:14] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 18:47:14] Creating trainer with model on device: cuda
[2021-03-22 18:47:14] Training with expected score >= 0.9900
[2021-03-22 18:47:14] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 18:52:18] Training accuracy: 0.9976
[2021-03-22 18:52:18] Testing on 10000 data points...
[2021-03-22 18:52:20] Test score for 5000 training labels: 0.4847
[2021-03-22 18:52:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:52:20] Found 50000 unlabelled features.
[2021-03-22 18:52:26] Showing labelled: True (9518/50000 visible, 482 redundant)
[2021-03-22 18:52:26] Creating trainer with model on device: cuda
[2021-03-22 18:52:26] Training with expected score >= 0.9900
[2021-03-22 18:52:26] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 18:54:59] Training accuracy: 0.9900
[2021-03-22 18:54:59] Testing on 10000 data points...
[2021-03-22 18:55:01] Test score for 10000 training labels: 0.5254
[2021-03-22 18:55:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:55:01] Found 50000 unlabelled features.
[2021-03-22 18:55:06] Showing labelled: True (13557/50000 visible, 1443 redundant)
[2021-03-22 18:55:06] Creating trainer with model on device: cuda
[2021-03-22 18:55:06] Training with expected score >= 0.9900
[2021-03-22 18:55:06] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 18:58:53] Training accuracy: 0.9974
[2021-03-22 18:58:53] Testing on 10000 data points...
[2021-03-22 18:58:55] Test score for 15000 training labels: 0.5562
[2021-03-22 18:58:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 18:58:55] Found 50000 unlabelled features.
[2021-03-22 18:59:01] Showing labelled: True (17167/50000 visible, 2833 redundant)
[2021-03-22 18:59:01] Creating trainer with model on device: cuda
[2021-03-22 18:59:01] Training with expected score >= 0.9900
[2021-03-22 18:59:01] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 19:04:01] Training accuracy: 0.9983
[2021-03-22 19:04:01] Testing on 10000 data points...
[2021-03-22 19:04:03] Test score for 20000 training labels: 0.5740
[2021-03-22 19:04:03] Seeded: 10
[2021-03-22 19:04:03] Running: experiment 1
[2021-03-22 19:04:03] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 19:04:03] Creating pretrained=True VGG16...
[2021-03-22 19:04:05] No parameter reset since we are using a pretrained model.
[2021-03-22 19:04:05] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 19:04:05] Creating trainer with model on device: cuda
[2021-03-22 19:04:05] Training with expected score >= 0.9900
[2021-03-22 19:04:05] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 19:09:04] Training accuracy: 0.9983
[2021-03-22 19:09:04] Testing on 10000 data points...
[2021-03-22 19:09:06] Test score for 5000 training labels: 0.4817
[2021-03-22 19:09:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 19:09:06] Found 50000 unlabelled features.
[2021-03-22 19:09:18] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 19:09:33] Searching for coresets greedily...
[2021-03-22 19:09:42] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 19:09:42] Creating trainer with model on device: cuda
[2021-03-22 19:09:42] Training with expected score >= 0.9900
[2021-03-22 19:09:42] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 19:12:13] Training accuracy: 0.9883
[2021-03-22 19:12:13] Training score (0.9883) was below expectations. Retraining...
[2021-03-22 19:12:13] Creating trainer with model on device: cuda
[2021-03-22 19:12:13] Training with expected score >= 0.9900
[2021-03-22 19:12:13] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 19:15:07] Training accuracy: 0.9986
[2021-03-22 19:15:07] Testing on 10000 data points...
[2021-03-22 19:15:09] Test score for 10000 training labels: 0.5381
[2021-03-22 19:15:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 19:15:09] Found 50000 unlabelled features.
[2021-03-22 19:15:21] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 19:15:51] Searching for coresets greedily...
[2021-03-22 19:16:01] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 19:16:01] Creating trainer with model on device: cuda
[2021-03-22 19:16:01] Training with expected score >= 0.9900
[2021-03-22 19:16:01] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 19:19:46] Training accuracy: 0.9972
[2021-03-22 19:19:46] Testing on 10000 data points...
[2021-03-22 19:19:48] Test score for 15000 training labels: 0.5665
[2021-03-22 19:19:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 19:19:48] Found 50000 unlabelled features.
[2021-03-22 19:20:02] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 19:20:47] Searching for coresets greedily...
[2021-03-22 19:20:56] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 19:20:56] Creating trainer with model on device: cuda
[2021-03-22 19:20:56] Training with expected score >= 0.9900
[2021-03-22 19:20:56] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 19:25:56] Training accuracy: 0.9980
[2021-03-22 19:25:56] Testing on 10000 data points...
[2021-03-22 19:25:58] Test score for 20000 training labels: 0.5965
[2021-03-22 19:25:58] Seeded: 10
[2021-03-22 19:25:58] Running: experiment 2
[2021-03-22 19:25:58] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-22 19:25:58] Creating pretrained=True VGG16...
[2021-03-22 19:26:00] No parameter reset since we are using a pretrained model.
[2021-03-22 19:26:00] vgg-pretrained: initialized 15291300 parameters.
[2021-03-22 19:26:00] Creating trainer with model on device: cuda
[2021-03-22 19:26:00] Training with expected score >= 0.9900
[2021-03-22 19:26:00] Training vgg-pretrained across 5000 data points in cifar100...
[2021-03-22 19:30:59] Training accuracy: 0.9973
[2021-03-22 19:30:59] Testing on 10000 data points...
[2021-03-22 19:31:01] Test score for 5000 training labels: 0.4832
[2021-03-22 19:31:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 19:31:01] Found 50000 unlabelled features.
[2021-03-22 19:31:11] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 19:31:26] Searching for coresets with 10 beams...
[2021-03-22 19:34:04] Showing labelled: True (10000/50000 visible, 0 redundant)
[2021-03-22 19:34:04] Creating trainer with model on device: cuda
[2021-03-22 19:34:04] Training with expected score >= 0.9900
[2021-03-22 19:34:04] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 19:36:34] Training accuracy: 0.9875
[2021-03-22 19:36:34] Training score (0.9875) was below expectations. Retraining...
[2021-03-22 19:36:34] Creating trainer with model on device: cuda
[2021-03-22 19:36:34] Training with expected score >= 0.9900
[2021-03-22 19:36:34] Training vgg-pretrained across 10000 data points in cifar100...
[2021-03-22 19:39:04] Training accuracy: 0.9980
[2021-03-22 19:39:04] Testing on 10000 data points...
[2021-03-22 19:39:06] Test score for 10000 training labels: 0.5426
[2021-03-22 19:39:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 19:39:06] Found 50000 unlabelled features.
[2021-03-22 19:39:19] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 19:39:49] Searching for coresets with 10 beams...
[2021-03-22 19:42:27] Showing labelled: True (15000/50000 visible, 0 redundant)
[2021-03-22 19:42:27] Creating trainer with model on device: cuda
[2021-03-22 19:42:27] Training with expected score >= 0.9900
[2021-03-22 19:42:27] Training vgg-pretrained across 15000 data points in cifar100...
[2021-03-22 19:46:12] Training accuracy: 0.9946
[2021-03-22 19:46:12] Testing on 10000 data points...
[2021-03-22 19:46:14] Test score for 15000 training labels: 0.5710
[2021-03-22 19:46:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-22 19:46:14] Found 50000 unlabelled features.
[2021-03-22 19:46:26] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-22 19:47:12] Searching for coresets with 10 beams...
[2021-03-22 19:49:50] Showing labelled: True (20000/50000 visible, 0 redundant)
[2021-03-22 19:49:50] Creating trainer with model on device: cuda
[2021-03-22 19:49:50] Training with expected score >= 0.9900
[2021-03-22 19:49:50] Training vgg-pretrained across 20000 data points in cifar100...
[2021-03-22 19:54:50] Training accuracy: 0.9976
[2021-03-22 19:54:50] Testing on 10000 data points...
[2021-03-22 19:54:52] Test score for 20000 training labels: 0.5883
[2021-03-22 19:54:52] Updated results: ../results/cifar100/main/results.json
