[2021-03-21 00:52:02] Created experiment 0:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: random
[2021-03-21 00:52:02] Created experiment 1:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: greedy-coreset
[2021-03-21 00:52:02] Created experiment 2:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: least-confidence
[2021-03-21 00:52:02] Created experiment 3:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: lc-beam-coreset
[2021-03-21 00:52:02] Created experiment 4:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: lc-beam-coreset
[2021-03-21 00:52:02] Created experiment 5:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: lc-beam-pweighted-coreset
[2021-03-21 00:52:02] Created experiment 6:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: lc-beam-pweighted-coreset
[2021-03-21 00:52:02] Created experiment 7:
[2021-03-21 00:52:02]  - Model: vgg-pretrained
[2021-03-21 00:52:02]  - Acquisition function: lc-beam-pweighted-relconf-optimcore
[2021-03-21 00:52:02] Loading cifar10 test set...
[2021-03-21 00:52:03] Experiment repeat 1/5
[2021-03-21 00:52:03] Seeded: 5
[2021-03-21 00:52:03] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 00:52:03] Randomly labelled 1000/50000
[2021-03-21 00:52:03] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 00:52:03] Seeded: 6
[2021-03-21 00:52:03] Running: experiment 0
[2021-03-21 00:52:03] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 00:52:03] Creating pretrained=True VGG16...
[2021-03-21 00:52:05] No parameter reset since we are using a pretrained model.
[2021-03-21 00:52:05] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 00:52:05] Creating trainer with model on device: cuda
[2021-03-21 00:52:10] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 00:53:09] Training accuracy: 0.9997
[2021-03-21 00:53:09] Testing on 10000 data points...
[2021-03-21 00:53:11] Test score for 1000 training labels: 0.7154
[2021-03-21 00:53:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:53:11] Found 50000 unlabelled features.
[2021-03-21 00:53:18] Showing labelled: True (1382/50000 visible, 18 redundant)
[2021-03-21 00:53:18] Creating trainer with model on device: cuda
[2021-03-21 00:53:18] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 00:53:38] Training accuracy: 0.9987
[2021-03-21 00:53:38] Testing on 10000 data points...
[2021-03-21 00:53:40] Test score for 1400 training labels: 0.7376
[2021-03-21 00:53:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:53:40] Found 50000 unlabelled features.
[2021-03-21 00:53:46] Showing labelled: True (1769/50000 visible, 31 redundant)
[2021-03-21 00:53:46] Creating trainer with model on device: cuda
[2021-03-21 00:53:46] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 00:54:13] Training accuracy: 0.9973
[2021-03-21 00:54:13] Testing on 10000 data points...
[2021-03-21 00:54:15] Test score for 1800 training labels: 0.7532
[2021-03-21 00:54:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:54:15] Found 50000 unlabelled features.
[2021-03-21 00:54:20] Showing labelled: True (2150/50000 visible, 50 redundant)
[2021-03-21 00:54:20] Creating trainer with model on device: cuda
[2021-03-21 00:54:20] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 00:54:54] Training accuracy: 1.0000
[2021-03-21 00:54:54] Testing on 10000 data points...
[2021-03-21 00:54:56] Test score for 2200 training labels: 0.7614
[2021-03-21 00:54:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:54:56] Found 50000 unlabelled features.
[2021-03-21 00:55:01] Showing labelled: True (2531/50000 visible, 69 redundant)
[2021-03-21 00:55:01] Creating trainer with model on device: cuda
[2021-03-21 00:55:01] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 00:55:40] Training accuracy: 0.9989
[2021-03-21 00:55:40] Testing on 10000 data points...
[2021-03-21 00:55:42] Test score for 2600 training labels: 0.7697
[2021-03-21 00:55:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:55:42] Found 50000 unlabelled features.
[2021-03-21 00:55:47] Showing labelled: True (2910/50000 visible, 90 redundant)
[2021-03-21 00:55:47] Creating trainer with model on device: cuda
[2021-03-21 00:55:47] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 00:56:31] Training accuracy: 0.9999
[2021-03-21 00:56:31] Testing on 10000 data points...
[2021-03-21 00:56:33] Test score for 3000 training labels: 0.7827
[2021-03-21 00:56:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:56:33] Found 50000 unlabelled features.
[2021-03-21 00:56:40] Showing labelled: True (3286/50000 visible, 114 redundant)
[2021-03-21 00:56:40] Creating trainer with model on device: cuda
[2021-03-21 00:56:40] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 00:57:32] Training accuracy: 1.0000
[2021-03-21 00:57:32] Testing on 10000 data points...
[2021-03-21 00:57:33] Test score for 3400 training labels: 0.7924
[2021-03-21 00:57:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:57:33] Found 50000 unlabelled features.
[2021-03-21 00:57:39] Showing labelled: True (3662/50000 visible, 138 redundant)
[2021-03-21 00:57:39] Creating trainer with model on device: cuda
[2021-03-21 00:57:39] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 00:58:36] Training accuracy: 0.9995
[2021-03-21 00:58:36] Testing on 10000 data points...
[2021-03-21 00:58:38] Test score for 3800 training labels: 0.7921
[2021-03-21 00:58:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:58:38] Found 50000 unlabelled features.
[2021-03-21 00:58:43] Showing labelled: True (4033/50000 visible, 167 redundant)
[2021-03-21 00:58:43] Creating trainer with model on device: cuda
[2021-03-21 00:58:43] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 00:59:46] Training accuracy: 1.0000
[2021-03-21 00:59:46] Testing on 10000 data points...
[2021-03-21 00:59:48] Test score for 4200 training labels: 0.7981
[2021-03-21 00:59:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 00:59:48] Found 50000 unlabelled features.
[2021-03-21 00:59:53] Showing labelled: True (4398/50000 visible, 202 redundant)
[2021-03-21 00:59:53] Creating trainer with model on device: cuda
[2021-03-21 00:59:53] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 01:01:02] Training accuracy: 1.0000
[2021-03-21 01:01:02] Testing on 10000 data points...
[2021-03-21 01:01:03] Test score for 4600 training labels: 0.7977
[2021-03-21 01:01:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:01:04] Found 50000 unlabelled features.
[2021-03-21 01:01:10] Showing labelled: True (4760/50000 visible, 240 redundant)
[2021-03-21 01:01:10] Creating trainer with model on device: cuda
[2021-03-21 01:01:10] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 01:02:24] Training accuracy: 1.0000
[2021-03-21 01:02:24] Testing on 10000 data points...
[2021-03-21 01:02:26] Test score for 5000 training labels: 0.8053
[2021-03-21 01:02:26] Seeded: 6
[2021-03-21 01:02:26] Running: experiment 1
[2021-03-21 01:02:27] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 01:02:27] Creating pretrained=True VGG16...
[2021-03-21 01:02:28] No parameter reset since we are using a pretrained model.
[2021-03-21 01:02:28] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 01:02:28] Creating trainer with model on device: cuda
[2021-03-21 01:02:28] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 01:03:29] Training accuracy: 0.9996
[2021-03-21 01:03:29] Testing on 10000 data points...
[2021-03-21 01:03:31] Test score for 1000 training labels: 0.7233
[2021-03-21 01:03:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:03:31] Found 50000 unlabelled features.
[2021-03-21 01:03:42] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:03:45] Searching for coresets greedily...
[2021-03-21 01:03:46] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 01:03:46] Creating trainer with model on device: cuda
[2021-03-21 01:03:46] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 01:04:07] Training accuracy: 0.9974
[2021-03-21 01:04:07] Testing on 10000 data points...
[2021-03-21 01:04:08] Test score for 1400 training labels: 0.7404
[2021-03-21 01:04:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:04:08] Found 50000 unlabelled features.
[2021-03-21 01:04:18] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:04:22] Searching for coresets greedily...
[2021-03-21 01:04:23] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 01:04:23] Creating trainer with model on device: cuda
[2021-03-21 01:04:23] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 01:04:51] Training accuracy: 0.9993
[2021-03-21 01:04:51] Testing on 10000 data points...
[2021-03-21 01:04:53] Test score for 1800 training labels: 0.7633
[2021-03-21 01:04:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:04:53] Found 50000 unlabelled features.
[2021-03-21 01:05:04] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:05:09] Searching for coresets greedily...
[2021-03-21 01:05:10] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 01:05:10] Creating trainer with model on device: cuda
[2021-03-21 01:05:10] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 01:05:44] Training accuracy: 0.9992
[2021-03-21 01:05:44] Testing on 10000 data points...
[2021-03-21 01:05:46] Test score for 2200 training labels: 0.7743
[2021-03-21 01:05:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:05:46] Found 50000 unlabelled features.
[2021-03-21 01:05:58] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:06:04] Searching for coresets greedily...
[2021-03-21 01:06:05] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 01:06:05] Creating trainer with model on device: cuda
[2021-03-21 01:06:05] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 01:06:44] Training accuracy: 0.9996
[2021-03-21 01:06:44] Testing on 10000 data points...
[2021-03-21 01:06:45] Test score for 2600 training labels: 0.7849
[2021-03-21 01:06:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:06:45] Found 50000 unlabelled features.
[2021-03-21 01:06:56] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:07:04] Searching for coresets greedily...
[2021-03-21 01:07:05] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 01:07:05] Creating trainer with model on device: cuda
[2021-03-21 01:07:05] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 01:07:50] Training accuracy: 0.9999
[2021-03-21 01:07:50] Testing on 10000 data points...
[2021-03-21 01:07:52] Test score for 3000 training labels: 0.7954
[2021-03-21 01:07:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:07:52] Found 50000 unlabelled features.
[2021-03-21 01:08:03] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:08:12] Searching for coresets greedily...
[2021-03-21 01:08:12] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 01:08:12] Creating trainer with model on device: cuda
[2021-03-21 01:08:12] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 01:09:04] Training accuracy: 0.9997
[2021-03-21 01:09:04] Testing on 10000 data points...
[2021-03-21 01:09:06] Test score for 3400 training labels: 0.8023
[2021-03-21 01:09:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:09:06] Found 50000 unlabelled features.
[2021-03-21 01:09:17] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:09:27] Searching for coresets greedily...
[2021-03-21 01:09:27] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 01:09:27] Creating trainer with model on device: cuda
[2021-03-21 01:09:27] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 01:10:26] Training accuracy: 1.0000
[2021-03-21 01:10:26] Testing on 10000 data points...
[2021-03-21 01:10:28] Test score for 3800 training labels: 0.8144
[2021-03-21 01:10:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:10:28] Found 50000 unlabelled features.
[2021-03-21 01:10:40] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:10:51] Searching for coresets greedily...
[2021-03-21 01:10:51] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 01:10:51] Creating trainer with model on device: cuda
[2021-03-21 01:10:51] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 01:11:54] Training accuracy: 1.0000
[2021-03-21 01:11:54] Testing on 10000 data points...
[2021-03-21 01:11:56] Test score for 4200 training labels: 0.8201
[2021-03-21 01:11:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:11:56] Found 50000 unlabelled features.
[2021-03-21 01:12:07] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:12:20] Searching for coresets greedily...
[2021-03-21 01:12:21] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 01:12:21] Creating trainer with model on device: cuda
[2021-03-21 01:12:21] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 01:13:30] Training accuracy: 1.0000
[2021-03-21 01:13:30] Testing on 10000 data points...
[2021-03-21 01:13:32] Test score for 4600 training labels: 0.8280
[2021-03-21 01:13:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:13:32] Found 50000 unlabelled features.
[2021-03-21 01:13:44] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:13:58] Searching for coresets greedily...
[2021-03-21 01:13:59] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 01:13:59] Creating trainer with model on device: cuda
[2021-03-21 01:13:59] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 01:15:15] Training accuracy: 1.0000
[2021-03-21 01:15:15] Testing on 10000 data points...
[2021-03-21 01:15:17] Test score for 5000 training labels: 0.8310
[2021-03-21 01:15:17] Seeded: 6
[2021-03-21 01:15:17] Running: experiment 2
[2021-03-21 01:15:17] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 01:15:17] Creating pretrained=True VGG16...
[2021-03-21 01:15:19] No parameter reset since we are using a pretrained model.
[2021-03-21 01:15:19] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 01:15:19] Creating trainer with model on device: cuda
[2021-03-21 01:15:19] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 01:16:20] Training accuracy: 0.9301
[2021-03-21 01:16:20] Testing on 10000 data points...
[2021-03-21 01:16:22] Test score for 1000 training labels: 0.6751
[2021-03-21 01:16:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:16:22] Found 50000 unlabelled features.
[2021-03-21 01:16:32] Showing labelled: True (1396/50000 visible, 4 redundant)
[2021-03-21 01:16:32] Creating trainer with model on device: cuda
[2021-03-21 01:16:32] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 01:16:53] Training accuracy: 0.9967
[2021-03-21 01:16:53] Testing on 10000 data points...
[2021-03-21 01:16:55] Test score for 1400 training labels: 0.7328
[2021-03-21 01:16:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:16:55] Found 50000 unlabelled features.
[2021-03-21 01:17:05] Showing labelled: True (1795/50000 visible, 5 redundant)
[2021-03-21 01:17:05] Creating trainer with model on device: cuda
[2021-03-21 01:17:05] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 01:17:32] Training accuracy: 0.9944
[2021-03-21 01:17:32] Testing on 10000 data points...
[2021-03-21 01:17:34] Test score for 1800 training labels: 0.7490
[2021-03-21 01:17:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:17:34] Found 50000 unlabelled features.
[2021-03-21 01:17:45] Showing labelled: True (2191/50000 visible, 9 redundant)
[2021-03-21 01:17:45] Creating trainer with model on device: cuda
[2021-03-21 01:17:45] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 01:18:18] Training accuracy: 0.9992
[2021-03-21 01:18:18] Testing on 10000 data points...
[2021-03-21 01:18:21] Test score for 2200 training labels: 0.7765
[2021-03-21 01:18:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:18:21] Found 50000 unlabelled features.
[2021-03-21 01:18:30] Showing labelled: True (2590/50000 visible, 10 redundant)
[2021-03-21 01:18:30] Creating trainer with model on device: cuda
[2021-03-21 01:18:30] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 01:19:09] Training accuracy: 0.9978
[2021-03-21 01:19:09] Testing on 10000 data points...
[2021-03-21 01:19:11] Test score for 2600 training labels: 0.7885
[2021-03-21 01:19:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:19:11] Found 50000 unlabelled features.
[2021-03-21 01:19:21] Showing labelled: True (2989/50000 visible, 11 redundant)
[2021-03-21 01:19:21] Creating trainer with model on device: cuda
[2021-03-21 01:19:21] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 01:20:07] Training accuracy: 1.0000
[2021-03-21 01:20:07] Testing on 10000 data points...
[2021-03-21 01:20:09] Test score for 3000 training labels: 0.7996
[2021-03-21 01:20:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:20:09] Found 50000 unlabelled features.
[2021-03-21 01:20:19] Showing labelled: True (3389/50000 visible, 11 redundant)
[2021-03-21 01:20:19] Creating trainer with model on device: cuda
[2021-03-21 01:20:19] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 01:21:10] Training accuracy: 0.9993
[2021-03-21 01:21:10] Testing on 10000 data points...
[2021-03-21 01:21:12] Test score for 3400 training labels: 0.8095
[2021-03-21 01:21:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:21:12] Found 50000 unlabelled features.
[2021-03-21 01:21:22] Showing labelled: True (3789/50000 visible, 11 redundant)
[2021-03-21 01:21:22] Creating trainer with model on device: cuda
[2021-03-21 01:21:22] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 01:22:20] Training accuracy: 1.0000
[2021-03-21 01:22:20] Testing on 10000 data points...
[2021-03-21 01:22:22] Test score for 3800 training labels: 0.8148
[2021-03-21 01:22:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:22:22] Found 50000 unlabelled features.
[2021-03-21 01:22:32] Showing labelled: True (4189/50000 visible, 11 redundant)
[2021-03-21 01:22:32] Creating trainer with model on device: cuda
[2021-03-21 01:22:32] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 01:23:35] Training accuracy: 1.0000
[2021-03-21 01:23:35] Testing on 10000 data points...
[2021-03-21 01:23:37] Test score for 4200 training labels: 0.8188
[2021-03-21 01:23:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:23:37] Found 50000 unlabelled features.
[2021-03-21 01:23:48] Showing labelled: True (4589/50000 visible, 11 redundant)
[2021-03-21 01:23:48] Creating trainer with model on device: cuda
[2021-03-21 01:23:48] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 01:24:58] Training accuracy: 0.9999
[2021-03-21 01:24:58] Testing on 10000 data points...
[2021-03-21 01:25:00] Test score for 4600 training labels: 0.8274
[2021-03-21 01:25:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:25:00] Found 50000 unlabelled features.
[2021-03-21 01:25:10] Showing labelled: True (4987/50000 visible, 13 redundant)
[2021-03-21 01:25:10] Creating trainer with model on device: cuda
[2021-03-21 01:25:10] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 01:26:26] Training accuracy: 0.9994
[2021-03-21 01:26:26] Testing on 10000 data points...
[2021-03-21 01:26:28] Test score for 5000 training labels: 0.8323
[2021-03-21 01:26:28] Seeded: 6
[2021-03-21 01:26:28] Running: experiment 3
[2021-03-21 01:26:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 01:26:28] Creating pretrained=True VGG16...
[2021-03-21 01:26:29] No parameter reset since we are using a pretrained model.
[2021-03-21 01:26:29] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 01:26:29] Creating trainer with model on device: cuda
[2021-03-21 01:26:30] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 01:27:29] Training accuracy: 0.9996
[2021-03-21 01:27:29] Testing on 10000 data points...
[2021-03-21 01:27:31] Test score for 1000 training labels: 0.7220
[2021-03-21 01:27:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:27:31] Found 50000 unlabelled features.
[2021-03-21 01:27:42] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:27:45] Searching for coresets with 1 beams...
[2021-03-21 01:27:46] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 01:27:46] Creating trainer with model on device: cuda
[2021-03-21 01:27:46] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 01:28:07] Training accuracy: 0.9992
[2021-03-21 01:28:07] Testing on 10000 data points...
[2021-03-21 01:28:09] Test score for 1400 training labels: 0.7402
[2021-03-21 01:28:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:28:09] Found 50000 unlabelled features.
[2021-03-21 01:28:20] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:28:24] Searching for coresets with 1 beams...
[2021-03-21 01:28:25] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 01:28:25] Creating trainer with model on device: cuda
[2021-03-21 01:28:25] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 01:28:53] Training accuracy: 0.9991
[2021-03-21 01:28:53] Testing on 10000 data points...
[2021-03-21 01:28:55] Test score for 1800 training labels: 0.7635
[2021-03-21 01:28:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:28:55] Found 50000 unlabelled features.
[2021-03-21 01:29:05] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:29:10] Searching for coresets with 1 beams...
[2021-03-21 01:29:11] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 01:29:11] Creating trainer with model on device: cuda
[2021-03-21 01:29:11] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 01:29:46] Training accuracy: 0.9994
[2021-03-21 01:29:46] Testing on 10000 data points...
[2021-03-21 01:29:48] Test score for 2200 training labels: 0.7768
[2021-03-21 01:29:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:29:48] Found 50000 unlabelled features.
[2021-03-21 01:29:57] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:30:04] Searching for coresets with 1 beams...
[2021-03-21 01:30:05] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 01:30:05] Creating trainer with model on device: cuda
[2021-03-21 01:30:05] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 01:30:44] Training accuracy: 1.0000
[2021-03-21 01:30:44] Testing on 10000 data points...
[2021-03-21 01:30:46] Test score for 2600 training labels: 0.7903
[2021-03-21 01:30:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:30:46] Found 50000 unlabelled features.
[2021-03-21 01:30:58] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:31:05] Searching for coresets with 1 beams...
[2021-03-21 01:31:06] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 01:31:06] Creating trainer with model on device: cuda
[2021-03-21 01:31:06] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 01:31:52] Training accuracy: 0.9992
[2021-03-21 01:31:52] Testing on 10000 data points...
[2021-03-21 01:31:54] Test score for 3000 training labels: 0.8000
[2021-03-21 01:31:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:31:54] Found 50000 unlabelled features.
[2021-03-21 01:32:04] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:32:13] Searching for coresets with 1 beams...
[2021-03-21 01:32:14] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 01:32:14] Creating trainer with model on device: cuda
[2021-03-21 01:32:14] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 01:33:05] Training accuracy: 0.9997
[2021-03-21 01:33:05] Testing on 10000 data points...
[2021-03-21 01:33:07] Test score for 3400 training labels: 0.8070
[2021-03-21 01:33:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:33:07] Found 50000 unlabelled features.
[2021-03-21 01:33:18] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:33:27] Searching for coresets with 1 beams...
[2021-03-21 01:33:29] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 01:33:29] Creating trainer with model on device: cuda
[2021-03-21 01:33:29] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 01:34:27] Training accuracy: 1.0000
[2021-03-21 01:34:27] Testing on 10000 data points...
[2021-03-21 01:34:28] Test score for 3800 training labels: 0.8124
[2021-03-21 01:34:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:34:29] Found 50000 unlabelled features.
[2021-03-21 01:34:39] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:34:50] Searching for coresets with 1 beams...
[2021-03-21 01:34:52] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 01:34:52] Creating trainer with model on device: cuda
[2021-03-21 01:34:52] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 01:35:55] Training accuracy: 1.0000
[2021-03-21 01:35:55] Testing on 10000 data points...
[2021-03-21 01:35:57] Test score for 4200 training labels: 0.8199
[2021-03-21 01:35:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:35:57] Found 50000 unlabelled features.
[2021-03-21 01:36:08] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:36:20] Searching for coresets with 1 beams...
[2021-03-21 01:36:22] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 01:36:22] Creating trainer with model on device: cuda
[2021-03-21 01:36:22] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 01:37:31] Training accuracy: 0.9998
[2021-03-21 01:37:31] Testing on 10000 data points...
[2021-03-21 01:37:33] Test score for 4600 training labels: 0.8247
[2021-03-21 01:37:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:37:33] Found 50000 unlabelled features.
[2021-03-21 01:37:44] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:37:57] Searching for coresets with 1 beams...
[2021-03-21 01:37:59] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 01:37:59] Creating trainer with model on device: cuda
[2021-03-21 01:37:59] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 01:39:15] Training accuracy: 1.0000
[2021-03-21 01:39:15] Testing on 10000 data points...
[2021-03-21 01:39:17] Test score for 5000 training labels: 0.8304
[2021-03-21 01:39:17] Seeded: 6
[2021-03-21 01:39:17] Running: experiment 4
[2021-03-21 01:39:17] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 01:39:17] Creating pretrained=True VGG16...
[2021-03-21 01:39:18] No parameter reset since we are using a pretrained model.
[2021-03-21 01:39:18] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 01:39:18] Creating trainer with model on device: cuda
[2021-03-21 01:39:18] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 01:40:18] Training accuracy: 0.9979
[2021-03-21 01:40:18] Testing on 10000 data points...
[2021-03-21 01:40:20] Test score for 1000 training labels: 0.7101
[2021-03-21 01:40:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:40:20] Found 50000 unlabelled features.
[2021-03-21 01:40:30] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:40:33] Searching for coresets with 20 beams...
[2021-03-21 01:41:02] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 01:41:02] Creating trainer with model on device: cuda
[2021-03-21 01:41:02] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 01:41:23] Training accuracy: 0.9986
[2021-03-21 01:41:23] Testing on 10000 data points...
[2021-03-21 01:41:25] Test score for 1400 training labels: 0.7404
[2021-03-21 01:41:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:41:25] Found 50000 unlabelled features.
[2021-03-21 01:41:35] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:41:39] Searching for coresets with 20 beams...
[2021-03-21 01:42:08] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 01:42:08] Creating trainer with model on device: cuda
[2021-03-21 01:42:08] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 01:42:35] Training accuracy: 0.9987
[2021-03-21 01:42:35] Testing on 10000 data points...
[2021-03-21 01:42:37] Test score for 1800 training labels: 0.7662
[2021-03-21 01:42:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:42:37] Found 50000 unlabelled features.
[2021-03-21 01:42:48] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:42:53] Searching for coresets with 20 beams...
[2021-03-21 01:43:22] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 01:43:22] Creating trainer with model on device: cuda
[2021-03-21 01:43:22] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 01:43:55] Training accuracy: 1.0000
[2021-03-21 01:43:55] Testing on 10000 data points...
[2021-03-21 01:43:57] Test score for 2200 training labels: 0.7752
[2021-03-21 01:43:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:43:57] Found 50000 unlabelled features.
[2021-03-21 01:44:08] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:44:15] Searching for coresets with 20 beams...
[2021-03-21 01:44:43] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 01:44:43] Creating trainer with model on device: cuda
[2021-03-21 01:44:43] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 01:45:23] Training accuracy: 0.9994
[2021-03-21 01:45:23] Testing on 10000 data points...
[2021-03-21 01:45:25] Test score for 2600 training labels: 0.7870
[2021-03-21 01:45:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:45:25] Found 50000 unlabelled features.
[2021-03-21 01:45:36] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:45:44] Searching for coresets with 20 beams...
[2021-03-21 01:46:13] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 01:46:13] Creating trainer with model on device: cuda
[2021-03-21 01:46:13] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 01:46:59] Training accuracy: 0.9999
[2021-03-21 01:46:59] Testing on 10000 data points...
[2021-03-21 01:47:01] Test score for 3000 training labels: 0.7993
[2021-03-21 01:47:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:47:01] Found 50000 unlabelled features.
[2021-03-21 01:47:12] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:47:21] Searching for coresets with 20 beams...
[2021-03-21 01:47:50] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 01:47:50] Creating trainer with model on device: cuda
[2021-03-21 01:47:50] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 01:48:42] Training accuracy: 0.9996
[2021-03-21 01:48:42] Testing on 10000 data points...
[2021-03-21 01:48:44] Test score for 3400 training labels: 0.8032
[2021-03-21 01:48:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:48:44] Found 50000 unlabelled features.
[2021-03-21 01:48:55] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:49:06] Searching for coresets with 20 beams...
[2021-03-21 01:49:35] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 01:49:35] Creating trainer with model on device: cuda
[2021-03-21 01:49:35] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 01:50:32] Training accuracy: 0.9995
[2021-03-21 01:50:32] Testing on 10000 data points...
[2021-03-21 01:50:34] Test score for 3800 training labels: 0.8117
[2021-03-21 01:50:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:50:34] Found 50000 unlabelled features.
[2021-03-21 01:50:46] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:50:57] Searching for coresets with 20 beams...
[2021-03-21 01:51:26] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 01:51:26] Creating trainer with model on device: cuda
[2021-03-21 01:51:26] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 01:52:30] Training accuracy: 1.0000
[2021-03-21 01:52:30] Testing on 10000 data points...
[2021-03-21 01:52:32] Test score for 4200 training labels: 0.8156
[2021-03-21 01:52:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:52:32] Found 50000 unlabelled features.
[2021-03-21 01:52:44] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:52:56] Searching for coresets with 20 beams...
[2021-03-21 01:53:24] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 01:53:24] Creating trainer with model on device: cuda
[2021-03-21 01:53:24] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 01:54:34] Training accuracy: 0.9998
[2021-03-21 01:54:34] Testing on 10000 data points...
[2021-03-21 01:54:36] Test score for 4600 training labels: 0.8235
[2021-03-21 01:54:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:54:36] Found 50000 unlabelled features.
[2021-03-21 01:54:47] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:55:02] Searching for coresets with 20 beams...
[2021-03-21 01:55:31] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 01:55:31] Creating trainer with model on device: cuda
[2021-03-21 01:55:31] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 01:56:48] Training accuracy: 1.0000
[2021-03-21 01:56:48] Testing on 10000 data points...
[2021-03-21 01:56:50] Test score for 5000 training labels: 0.8303
[2021-03-21 01:56:50] Seeded: 6
[2021-03-21 01:56:50] Running: experiment 5
[2021-03-21 01:56:50] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 01:56:50] Creating pretrained=True VGG16...
[2021-03-21 01:56:51] No parameter reset since we are using a pretrained model.
[2021-03-21 01:56:51] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 01:56:51] Creating trainer with model on device: cuda
[2021-03-21 01:56:51] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 01:57:52] Training accuracy: 0.9958
[2021-03-21 01:57:52] Testing on 10000 data points...
[2021-03-21 01:57:54] Test score for 1000 training labels: 0.7070
[2021-03-21 01:57:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:57:54] Found 50000 unlabelled features.
[2021-03-21 01:58:04] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:58:07] Searching for coresets with 1 beams...
[2021-03-21 01:58:09] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 01:58:09] Creating trainer with model on device: cuda
[2021-03-21 01:58:09] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 01:58:30] Training accuracy: 0.9987
[2021-03-21 01:58:30] Testing on 10000 data points...
[2021-03-21 01:58:32] Test score for 1400 training labels: 0.7494
[2021-03-21 01:58:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:58:32] Found 50000 unlabelled features.
[2021-03-21 01:58:43] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:58:47] Searching for coresets with 1 beams...
[2021-03-21 01:58:48] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 01:58:48] Creating trainer with model on device: cuda
[2021-03-21 01:58:48] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 01:59:15] Training accuracy: 0.9967
[2021-03-21 01:59:15] Testing on 10000 data points...
[2021-03-21 01:59:17] Test score for 1800 training labels: 0.7628
[2021-03-21 01:59:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 01:59:17] Found 50000 unlabelled features.
[2021-03-21 01:59:28] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 01:59:34] Searching for coresets with 1 beams...
[2021-03-21 01:59:35] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 01:59:35] Creating trainer with model on device: cuda
[2021-03-21 01:59:35] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 02:00:09] Training accuracy: 0.9996
[2021-03-21 02:00:09] Testing on 10000 data points...
[2021-03-21 02:00:11] Test score for 2200 training labels: 0.7816
[2021-03-21 02:00:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:00:11] Found 50000 unlabelled features.
[2021-03-21 02:00:21] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:00:28] Searching for coresets with 1 beams...
[2021-03-21 02:00:29] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 02:00:29] Creating trainer with model on device: cuda
[2021-03-21 02:00:29] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 02:01:08] Training accuracy: 0.9999
[2021-03-21 02:01:08] Testing on 10000 data points...
[2021-03-21 02:01:10] Test score for 2600 training labels: 0.7929
[2021-03-21 02:01:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:01:10] Found 50000 unlabelled features.
[2021-03-21 02:01:21] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:01:29] Searching for coresets with 1 beams...
[2021-03-21 02:01:31] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 02:01:31] Creating trainer with model on device: cuda
[2021-03-21 02:01:31] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 02:02:15] Training accuracy: 0.9997
[2021-03-21 02:02:15] Testing on 10000 data points...
[2021-03-21 02:02:17] Test score for 3000 training labels: 0.8031
[2021-03-21 02:02:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:02:17] Found 50000 unlabelled features.
[2021-03-21 02:02:29] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:02:37] Searching for coresets with 1 beams...
[2021-03-21 02:02:39] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 02:02:39] Creating trainer with model on device: cuda
[2021-03-21 02:02:39] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 02:03:29] Training accuracy: 0.9999
[2021-03-21 02:03:29] Testing on 10000 data points...
[2021-03-21 02:03:31] Test score for 3400 training labels: 0.8113
[2021-03-21 02:03:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:03:32] Found 50000 unlabelled features.
[2021-03-21 02:03:43] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:03:53] Searching for coresets with 1 beams...
[2021-03-21 02:03:55] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 02:03:55] Creating trainer with model on device: cuda
[2021-03-21 02:03:55] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 02:04:52] Training accuracy: 0.9998
[2021-03-21 02:04:52] Testing on 10000 data points...
[2021-03-21 02:04:54] Test score for 3800 training labels: 0.8115
[2021-03-21 02:04:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:04:54] Found 50000 unlabelled features.
[2021-03-21 02:05:05] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:05:16] Searching for coresets with 1 beams...
[2021-03-21 02:05:18] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 02:05:18] Creating trainer with model on device: cuda
[2021-03-21 02:05:18] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 02:06:21] Training accuracy: 0.9999
[2021-03-21 02:06:21] Testing on 10000 data points...
[2021-03-21 02:06:23] Test score for 4200 training labels: 0.8231
[2021-03-21 02:06:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:06:23] Found 50000 unlabelled features.
[2021-03-21 02:06:35] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:06:47] Searching for coresets with 1 beams...
[2021-03-21 02:06:49] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 02:06:49] Creating trainer with model on device: cuda
[2021-03-21 02:06:49] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 02:07:57] Training accuracy: 1.0000
[2021-03-21 02:07:57] Testing on 10000 data points...
[2021-03-21 02:07:59] Test score for 4600 training labels: 0.8259
[2021-03-21 02:07:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:07:59] Found 50000 unlabelled features.
[2021-03-21 02:08:11] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:08:25] Searching for coresets with 1 beams...
[2021-03-21 02:08:26] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 02:08:26] Creating trainer with model on device: cuda
[2021-03-21 02:08:26] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 02:09:42] Training accuracy: 1.0000
[2021-03-21 02:09:42] Testing on 10000 data points...
[2021-03-21 02:09:44] Test score for 5000 training labels: 0.8361
[2021-03-21 02:09:44] Seeded: 6
[2021-03-21 02:09:44] Running: experiment 6
[2021-03-21 02:09:44] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 02:09:44] Creating pretrained=True VGG16...
[2021-03-21 02:09:46] No parameter reset since we are using a pretrained model.
[2021-03-21 02:09:46] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 02:09:46] Creating trainer with model on device: cuda
[2021-03-21 02:09:46] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 02:10:46] Training accuracy: 0.9991
[2021-03-21 02:10:46] Testing on 10000 data points...
[2021-03-21 02:10:47] Test score for 1000 training labels: 0.7115
[2021-03-21 02:10:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:10:48] Found 50000 unlabelled features.
[2021-03-21 02:10:59] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:11:02] Searching for coresets with 20 beams...
[2021-03-21 02:11:30] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 02:11:30] Creating trainer with model on device: cuda
[2021-03-21 02:11:30] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 02:11:51] Training accuracy: 0.9979
[2021-03-21 02:11:51] Testing on 10000 data points...
[2021-03-21 02:11:54] Test score for 1400 training labels: 0.7523
[2021-03-21 02:11:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:11:54] Found 50000 unlabelled features.
[2021-03-21 02:12:04] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:12:08] Searching for coresets with 20 beams...
[2021-03-21 02:12:37] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 02:12:37] Creating trainer with model on device: cuda
[2021-03-21 02:12:37] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 02:13:05] Training accuracy: 0.9968
[2021-03-21 02:13:05] Testing on 10000 data points...
[2021-03-21 02:13:06] Test score for 1800 training labels: 0.7698
[2021-03-21 02:13:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:13:06] Found 50000 unlabelled features.
[2021-03-21 02:13:17] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:13:23] Searching for coresets with 20 beams...
[2021-03-21 02:13:52] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 02:13:52] Creating trainer with model on device: cuda
[2021-03-21 02:13:52] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 02:14:26] Training accuracy: 0.9998
[2021-03-21 02:14:26] Testing on 10000 data points...
[2021-03-21 02:14:28] Test score for 2200 training labels: 0.7884
[2021-03-21 02:14:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:14:28] Found 50000 unlabelled features.
[2021-03-21 02:14:38] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:14:45] Searching for coresets with 20 beams...
[2021-03-21 02:15:14] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 02:15:14] Creating trainer with model on device: cuda
[2021-03-21 02:15:14] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 02:15:54] Training accuracy: 0.9995
[2021-03-21 02:15:54] Testing on 10000 data points...
[2021-03-21 02:15:56] Test score for 2600 training labels: 0.7965
[2021-03-21 02:15:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:15:56] Found 50000 unlabelled features.
[2021-03-21 02:16:07] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:16:15] Searching for coresets with 20 beams...
[2021-03-21 02:16:44] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 02:16:44] Creating trainer with model on device: cuda
[2021-03-21 02:16:44] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 02:17:29] Training accuracy: 0.9996
[2021-03-21 02:17:29] Testing on 10000 data points...
[2021-03-21 02:17:31] Test score for 3000 training labels: 0.8044
[2021-03-21 02:17:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:17:31] Found 50000 unlabelled features.
[2021-03-21 02:17:42] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:17:52] Searching for coresets with 20 beams...
[2021-03-21 02:18:20] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 02:18:20] Creating trainer with model on device: cuda
[2021-03-21 02:18:20] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 02:19:12] Training accuracy: 1.0000
[2021-03-21 02:19:12] Testing on 10000 data points...
[2021-03-21 02:19:14] Test score for 3400 training labels: 0.8127
[2021-03-21 02:19:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:19:14] Found 50000 unlabelled features.
[2021-03-21 02:19:26] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:19:36] Searching for coresets with 20 beams...
[2021-03-21 02:20:05] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 02:20:05] Creating trainer with model on device: cuda
[2021-03-21 02:20:05] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 02:21:04] Training accuracy: 0.9998
[2021-03-21 02:21:04] Testing on 10000 data points...
[2021-03-21 02:21:06] Test score for 3800 training labels: 0.8211
[2021-03-21 02:21:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:21:06] Found 50000 unlabelled features.
[2021-03-21 02:21:17] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:21:28] Searching for coresets with 20 beams...
[2021-03-21 02:21:57] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 02:21:57] Creating trainer with model on device: cuda
[2021-03-21 02:21:57] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 02:23:01] Training accuracy: 0.9997
[2021-03-21 02:23:01] Testing on 10000 data points...
[2021-03-21 02:23:03] Test score for 4200 training labels: 0.8203
[2021-03-21 02:23:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:23:03] Found 50000 unlabelled features.
[2021-03-21 02:23:14] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:23:26] Searching for coresets with 20 beams...
[2021-03-21 02:23:55] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 02:23:55] Creating trainer with model on device: cuda
[2021-03-21 02:23:55] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 02:25:06] Training accuracy: 1.0000
[2021-03-21 02:25:06] Testing on 10000 data points...
[2021-03-21 02:25:08] Test score for 4600 training labels: 0.8277
[2021-03-21 02:25:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:25:08] Found 50000 unlabelled features.
[2021-03-21 02:25:20] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:25:34] Searching for coresets with 20 beams...
[2021-03-21 02:26:03] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 02:26:03] Creating trainer with model on device: cuda
[2021-03-21 02:26:03] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 02:27:19] Training accuracy: 0.9999
[2021-03-21 02:27:19] Testing on 10000 data points...
[2021-03-21 02:27:21] Test score for 5000 training labels: 0.8325
[2021-03-21 02:27:21] Seeded: 6
[2021-03-21 02:27:21] Running: experiment 7
[2021-03-21 02:27:21] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 02:27:21] Creating pretrained=True VGG16...
[2021-03-21 02:27:23] No parameter reset since we are using a pretrained model.
[2021-03-21 02:27:23] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 02:27:23] Creating trainer with model on device: cuda
[2021-03-21 02:27:23] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 02:28:24] Training accuracy: 0.9993
[2021-03-21 02:28:24] Testing on 10000 data points...
[2021-03-21 02:28:25] Test score for 1000 training labels: 0.7239
[2021-03-21 02:28:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:28:26] Found 50000 unlabelled features.
[2021-03-21 02:28:36] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:28:39] Searching for coresets with 20 beams...
[2021-03-21 02:29:07] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 02:29:07] Creating trainer with model on device: cuda
[2021-03-21 02:29:07] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 02:29:29] Training accuracy: 0.9949
[2021-03-21 02:29:29] Testing on 10000 data points...
[2021-03-21 02:29:31] Test score for 1400 training labels: 0.7440
[2021-03-21 02:29:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:29:31] Found 50000 unlabelled features.
[2021-03-21 02:29:42] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:29:46] Searching for coresets with 20 beams...
[2021-03-21 02:30:14] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 02:30:14] Creating trainer with model on device: cuda
[2021-03-21 02:30:14] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 02:30:42] Training accuracy: 0.9970
[2021-03-21 02:30:42] Testing on 10000 data points...
[2021-03-21 02:30:44] Test score for 1800 training labels: 0.7696
[2021-03-21 02:30:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:30:44] Found 50000 unlabelled features.
[2021-03-21 02:30:54] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:30:59] Searching for coresets with 20 beams...
[2021-03-21 02:31:27] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 02:31:27] Creating trainer with model on device: cuda
[2021-03-21 02:31:27] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 02:32:01] Training accuracy: 0.9997
[2021-03-21 02:32:01] Testing on 10000 data points...
[2021-03-21 02:32:03] Test score for 2200 training labels: 0.7840
[2021-03-21 02:32:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:32:03] Found 50000 unlabelled features.
[2021-03-21 02:32:14] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:32:20] Searching for coresets with 20 beams...
[2021-03-21 02:32:48] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 02:32:48] Creating trainer with model on device: cuda
[2021-03-21 02:32:48] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 02:33:28] Training accuracy: 0.9994
[2021-03-21 02:33:28] Testing on 10000 data points...
[2021-03-21 02:33:30] Test score for 2600 training labels: 0.7944
[2021-03-21 02:33:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:33:30] Found 50000 unlabelled features.
[2021-03-21 02:33:40] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:33:49] Searching for coresets with 20 beams...
[2021-03-21 02:34:17] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 02:34:17] Creating trainer with model on device: cuda
[2021-03-21 02:34:17] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 02:35:02] Training accuracy: 0.9999
[2021-03-21 02:35:02] Testing on 10000 data points...
[2021-03-21 02:35:04] Test score for 3000 training labels: 0.8053
[2021-03-21 02:35:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:35:04] Found 50000 unlabelled features.
[2021-03-21 02:35:15] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:35:24] Searching for coresets with 20 beams...
[2021-03-21 02:35:52] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 02:35:52] Creating trainer with model on device: cuda
[2021-03-21 02:35:52] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 02:36:43] Training accuracy: 1.0000
[2021-03-21 02:36:43] Testing on 10000 data points...
[2021-03-21 02:36:45] Test score for 3400 training labels: 0.8073
[2021-03-21 02:36:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:36:45] Found 50000 unlabelled features.
[2021-03-21 02:36:56] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:37:06] Searching for coresets with 20 beams...
[2021-03-21 02:37:34] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 02:37:34] Creating trainer with model on device: cuda
[2021-03-21 02:37:34] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 02:38:32] Training accuracy: 1.0000
[2021-03-21 02:38:32] Testing on 10000 data points...
[2021-03-21 02:38:34] Test score for 3800 training labels: 0.8141
[2021-03-21 02:38:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:38:34] Found 50000 unlabelled features.
[2021-03-21 02:38:45] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:38:56] Searching for coresets with 20 beams...
[2021-03-21 02:39:24] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 02:39:24] Creating trainer with model on device: cuda
[2021-03-21 02:39:24] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 02:40:27] Training accuracy: 1.0000
[2021-03-21 02:40:27] Testing on 10000 data points...
[2021-03-21 02:40:29] Test score for 4200 training labels: 0.8209
[2021-03-21 02:40:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:40:30] Found 50000 unlabelled features.
[2021-03-21 02:40:41] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:40:53] Searching for coresets with 20 beams...
[2021-03-21 02:41:21] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 02:41:21] Creating trainer with model on device: cuda
[2021-03-21 02:41:21] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 02:42:30] Training accuracy: 0.9995
[2021-03-21 02:42:30] Testing on 10000 data points...
[2021-03-21 02:42:32] Test score for 4600 training labels: 0.8254
[2021-03-21 02:42:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:42:32] Found 50000 unlabelled features.
[2021-03-21 02:42:43] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:42:56] Searching for coresets with 20 beams...
[2021-03-21 02:43:25] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 02:43:25] Creating trainer with model on device: cuda
[2021-03-21 02:43:25] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 02:44:40] Training accuracy: 0.9998
[2021-03-21 02:44:40] Testing on 10000 data points...
[2021-03-21 02:44:43] Test score for 5000 training labels: 0.8262
[2021-03-21 02:44:43] Experiment repeat 2/5
[2021-03-21 02:44:43] Seeded: 6
[2021-03-21 02:44:43] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 02:44:43] Randomly labelled 1000/50000
[2021-03-21 02:44:43] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 02:44:43] Seeded: 7
[2021-03-21 02:44:43] Running: experiment 0
[2021-03-21 02:44:43] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 02:44:43] Creating pretrained=True VGG16...
[2021-03-21 02:44:44] No parameter reset since we are using a pretrained model.
[2021-03-21 02:44:44] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 02:44:44] Creating trainer with model on device: cuda
[2021-03-21 02:44:44] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 02:45:45] Training accuracy: 0.9997
[2021-03-21 02:45:45] Testing on 10000 data points...
[2021-03-21 02:45:47] Test score for 1000 training labels: 0.7439
[2021-03-21 02:45:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:45:47] Found 50000 unlabelled features.
[2021-03-21 02:45:53] Showing labelled: True (1389/50000 visible, 11 redundant)
[2021-03-21 02:45:53] Creating trainer with model on device: cuda
[2021-03-21 02:45:53] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 02:46:13] Training accuracy: 0.9996
[2021-03-21 02:46:13] Testing on 10000 data points...
[2021-03-21 02:46:15] Test score for 1400 training labels: 0.7517
[2021-03-21 02:46:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:46:15] Found 50000 unlabelled features.
[2021-03-21 02:46:22] Showing labelled: True (1779/50000 visible, 21 redundant)
[2021-03-21 02:46:22] Creating trainer with model on device: cuda
[2021-03-21 02:46:22] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 02:46:49] Training accuracy: 0.9998
[2021-03-21 02:46:49] Testing on 10000 data points...
[2021-03-21 02:46:51] Test score for 1800 training labels: 0.7637
[2021-03-21 02:46:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:46:51] Found 50000 unlabelled features.
[2021-03-21 02:46:57] Showing labelled: True (2163/50000 visible, 37 redundant)
[2021-03-21 02:46:57] Creating trainer with model on device: cuda
[2021-03-21 02:46:57] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 02:47:30] Training accuracy: 0.9999
[2021-03-21 02:47:30] Testing on 10000 data points...
[2021-03-21 02:47:32] Test score for 2200 training labels: 0.7719
[2021-03-21 02:47:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:47:32] Found 50000 unlabelled features.
[2021-03-21 02:47:38] Showing labelled: True (2547/50000 visible, 53 redundant)
[2021-03-21 02:47:38] Creating trainer with model on device: cuda
[2021-03-21 02:47:38] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 02:48:16] Training accuracy: 0.9994
[2021-03-21 02:48:17] Testing on 10000 data points...
[2021-03-21 02:48:19] Test score for 2600 training labels: 0.7809
[2021-03-21 02:48:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:48:19] Found 50000 unlabelled features.
[2021-03-21 02:48:24] Showing labelled: True (2917/50000 visible, 83 redundant)
[2021-03-21 02:48:24] Creating trainer with model on device: cuda
[2021-03-21 02:48:24] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 02:49:09] Training accuracy: 0.9998
[2021-03-21 02:49:09] Testing on 10000 data points...
[2021-03-21 02:49:11] Test score for 3000 training labels: 0.7897
[2021-03-21 02:49:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:49:11] Found 50000 unlabelled features.
[2021-03-21 02:49:16] Showing labelled: True (3299/50000 visible, 101 redundant)
[2021-03-21 02:49:16] Creating trainer with model on device: cuda
[2021-03-21 02:49:16] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 02:50:08] Training accuracy: 1.0000
[2021-03-21 02:50:08] Testing on 10000 data points...
[2021-03-21 02:50:10] Test score for 3400 training labels: 0.7986
[2021-03-21 02:50:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:50:10] Found 50000 unlabelled features.
[2021-03-21 02:50:16] Showing labelled: True (3672/50000 visible, 128 redundant)
[2021-03-21 02:50:16] Creating trainer with model on device: cuda
[2021-03-21 02:50:16] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 02:51:13] Training accuracy: 1.0000
[2021-03-21 02:51:13] Testing on 10000 data points...
[2021-03-21 02:51:16] Test score for 3800 training labels: 0.8033
[2021-03-21 02:51:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:51:16] Found 50000 unlabelled features.
[2021-03-21 02:51:22] Showing labelled: True (4044/50000 visible, 156 redundant)
[2021-03-21 02:51:22] Creating trainer with model on device: cuda
[2021-03-21 02:51:22] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 02:52:25] Training accuracy: 1.0000
[2021-03-21 02:52:25] Testing on 10000 data points...
[2021-03-21 02:52:27] Test score for 4200 training labels: 0.8070
[2021-03-21 02:52:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:52:27] Found 50000 unlabelled features.
[2021-03-21 02:52:33] Showing labelled: True (4415/50000 visible, 185 redundant)
[2021-03-21 02:52:33] Creating trainer with model on device: cuda
[2021-03-21 02:52:33] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 02:53:43] Training accuracy: 1.0000
[2021-03-21 02:53:43] Testing on 10000 data points...
[2021-03-21 02:53:45] Test score for 4600 training labels: 0.8109
[2021-03-21 02:53:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:53:45] Found 50000 unlabelled features.
[2021-03-21 02:53:50] Showing labelled: True (4769/50000 visible, 231 redundant)
[2021-03-21 02:53:50] Creating trainer with model on device: cuda
[2021-03-21 02:53:50] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 02:55:05] Training accuracy: 0.9997
[2021-03-21 02:55:05] Testing on 10000 data points...
[2021-03-21 02:55:07] Test score for 5000 training labels: 0.8116
[2021-03-21 02:55:07] Seeded: 7
[2021-03-21 02:55:07] Running: experiment 1
[2021-03-21 02:55:07] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 02:55:07] Creating pretrained=True VGG16...
[2021-03-21 02:55:09] No parameter reset since we are using a pretrained model.
[2021-03-21 02:55:09] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 02:55:09] Creating trainer with model on device: cuda
[2021-03-21 02:55:09] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 02:56:10] Training accuracy: 0.9998
[2021-03-21 02:56:10] Testing on 10000 data points...
[2021-03-21 02:56:12] Test score for 1000 training labels: 0.7411
[2021-03-21 02:56:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:56:12] Found 50000 unlabelled features.
[2021-03-21 02:56:22] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:56:25] Searching for coresets greedily...
[2021-03-21 02:56:26] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 02:56:26] Creating trainer with model on device: cuda
[2021-03-21 02:56:26] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 02:56:47] Training accuracy: 0.9999
[2021-03-21 02:56:47] Testing on 10000 data points...
[2021-03-21 02:56:50] Test score for 1400 training labels: 0.7589
[2021-03-21 02:56:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:56:50] Found 50000 unlabelled features.
[2021-03-21 02:57:00] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:57:05] Searching for coresets greedily...
[2021-03-21 02:57:06] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 02:57:06] Creating trainer with model on device: cuda
[2021-03-21 02:57:06] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 02:57:32] Training accuracy: 0.9991
[2021-03-21 02:57:32] Testing on 10000 data points...
[2021-03-21 02:57:34] Test score for 1800 training labels: 0.7747
[2021-03-21 02:57:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:57:34] Found 50000 unlabelled features.
[2021-03-21 02:57:45] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:57:51] Searching for coresets greedily...
[2021-03-21 02:57:52] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 02:57:52] Creating trainer with model on device: cuda
[2021-03-21 02:57:52] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 02:58:26] Training accuracy: 0.9979
[2021-03-21 02:58:26] Testing on 10000 data points...
[2021-03-21 02:58:28] Test score for 2200 training labels: 0.7871
[2021-03-21 02:58:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:58:28] Found 50000 unlabelled features.
[2021-03-21 02:58:38] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:58:45] Searching for coresets greedily...
[2021-03-21 02:58:46] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 02:58:46] Creating trainer with model on device: cuda
[2021-03-21 02:58:46] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 02:59:24] Training accuracy: 0.9996
[2021-03-21 02:59:24] Testing on 10000 data points...
[2021-03-21 02:59:26] Test score for 2600 training labels: 0.7945
[2021-03-21 02:59:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 02:59:26] Found 50000 unlabelled features.
[2021-03-21 02:59:37] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 02:59:45] Searching for coresets greedily...
[2021-03-21 02:59:46] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 02:59:46] Creating trainer with model on device: cuda
[2021-03-21 02:59:46] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 03:00:31] Training accuracy: 0.9992
[2021-03-21 03:00:31] Testing on 10000 data points...
[2021-03-21 03:00:33] Test score for 3000 training labels: 0.8056
[2021-03-21 03:00:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:00:33] Found 50000 unlabelled features.
[2021-03-21 03:00:44] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:00:52] Searching for coresets greedily...
[2021-03-21 03:00:53] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 03:00:53] Creating trainer with model on device: cuda
[2021-03-21 03:00:53] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 03:01:44] Training accuracy: 0.9999
[2021-03-21 03:01:44] Testing on 10000 data points...
[2021-03-21 03:01:46] Test score for 3400 training labels: 0.8105
[2021-03-21 03:01:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:01:46] Found 50000 unlabelled features.
[2021-03-21 03:01:58] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:02:08] Searching for coresets greedily...
[2021-03-21 03:02:09] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 03:02:09] Creating trainer with model on device: cuda
[2021-03-21 03:02:09] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 03:03:06] Training accuracy: 1.0000
[2021-03-21 03:03:06] Testing on 10000 data points...
[2021-03-21 03:03:08] Test score for 3800 training labels: 0.8199
[2021-03-21 03:03:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:03:08] Found 50000 unlabelled features.
[2021-03-21 03:03:19] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:03:31] Searching for coresets greedily...
[2021-03-21 03:03:32] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 03:03:32] Creating trainer with model on device: cuda
[2021-03-21 03:03:32] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 03:04:36] Training accuracy: 1.0000
[2021-03-21 03:04:36] Testing on 10000 data points...
[2021-03-21 03:04:38] Test score for 4200 training labels: 0.8267
[2021-03-21 03:04:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:04:38] Found 50000 unlabelled features.
[2021-03-21 03:04:49] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:05:03] Searching for coresets greedily...
[2021-03-21 03:05:04] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 03:05:04] Creating trainer with model on device: cuda
[2021-03-21 03:05:04] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 03:06:14] Training accuracy: 1.0000
[2021-03-21 03:06:14] Testing on 10000 data points...
[2021-03-21 03:06:16] Test score for 4600 training labels: 0.8313
[2021-03-21 03:06:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:06:16] Found 50000 unlabelled features.
[2021-03-21 03:06:28] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:06:41] Searching for coresets greedily...
[2021-03-21 03:06:42] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 03:06:42] Creating trainer with model on device: cuda
[2021-03-21 03:06:42] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 03:07:58] Training accuracy: 1.0000
[2021-03-21 03:07:58] Testing on 10000 data points...
[2021-03-21 03:08:00] Test score for 5000 training labels: 0.8327
[2021-03-21 03:08:00] Seeded: 7
[2021-03-21 03:08:00] Running: experiment 2
[2021-03-21 03:08:00] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 03:08:00] Creating pretrained=True VGG16...
[2021-03-21 03:08:02] No parameter reset since we are using a pretrained model.
[2021-03-21 03:08:02] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 03:08:02] Creating trainer with model on device: cuda
[2021-03-21 03:08:02] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 03:09:03] Training accuracy: 0.9989
[2021-03-21 03:09:03] Testing on 10000 data points...
[2021-03-21 03:09:05] Test score for 1000 training labels: 0.7407
[2021-03-21 03:09:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:09:05] Found 50000 unlabelled features.
[2021-03-21 03:09:14] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 03:09:14] Creating trainer with model on device: cuda
[2021-03-21 03:09:14] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 03:09:35] Training accuracy: 0.9975
[2021-03-21 03:09:35] Testing on 10000 data points...
[2021-03-21 03:09:37] Test score for 1400 training labels: 0.7619
[2021-03-21 03:09:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:09:37] Found 50000 unlabelled features.
[2021-03-21 03:09:47] Showing labelled: True (1797/50000 visible, 3 redundant)
[2021-03-21 03:09:47] Creating trainer with model on device: cuda
[2021-03-21 03:09:47] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 03:10:14] Training accuracy: 0.9506
[2021-03-21 03:10:14] Testing on 10000 data points...
[2021-03-21 03:10:16] Test score for 1800 training labels: 0.7537
[2021-03-21 03:10:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:10:16] Found 50000 unlabelled features.
[2021-03-21 03:10:25] Showing labelled: True (2187/50000 visible, 13 redundant)
[2021-03-21 03:10:25] Creating trainer with model on device: cuda
[2021-03-21 03:10:25] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 03:10:59] Training accuracy: 0.9988
[2021-03-21 03:10:59] Testing on 10000 data points...
[2021-03-21 03:11:01] Test score for 2200 training labels: 0.7896
[2021-03-21 03:11:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:11:01] Found 50000 unlabelled features.
[2021-03-21 03:11:12] Showing labelled: True (2586/50000 visible, 14 redundant)
[2021-03-21 03:11:12] Creating trainer with model on device: cuda
[2021-03-21 03:11:12] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 03:11:50] Training accuracy: 0.9995
[2021-03-21 03:11:50] Testing on 10000 data points...
[2021-03-21 03:11:52] Test score for 2600 training labels: 0.8022
[2021-03-21 03:11:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:11:52] Found 50000 unlabelled features.
[2021-03-21 03:12:03] Showing labelled: True (2985/50000 visible, 15 redundant)
[2021-03-21 03:12:03] Creating trainer with model on device: cuda
[2021-03-21 03:12:03] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 03:12:48] Training accuracy: 1.0000
[2021-03-21 03:12:48] Testing on 10000 data points...
[2021-03-21 03:12:50] Test score for 3000 training labels: 0.8122
[2021-03-21 03:12:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:12:50] Found 50000 unlabelled features.
[2021-03-21 03:13:01] Showing labelled: True (3384/50000 visible, 16 redundant)
[2021-03-21 03:13:01] Creating trainer with model on device: cuda
[2021-03-21 03:13:01] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 03:13:52] Training accuracy: 0.9994
[2021-03-21 03:13:52] Testing on 10000 data points...
[2021-03-21 03:13:54] Test score for 3400 training labels: 0.8195
[2021-03-21 03:13:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:13:54] Found 50000 unlabelled features.
[2021-03-21 03:14:04] Showing labelled: True (3784/50000 visible, 16 redundant)
[2021-03-21 03:14:04] Creating trainer with model on device: cuda
[2021-03-21 03:14:04] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 03:15:02] Training accuracy: 0.9994
[2021-03-21 03:15:02] Testing on 10000 data points...
[2021-03-21 03:15:04] Test score for 3800 training labels: 0.8253
[2021-03-21 03:15:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:15:04] Found 50000 unlabelled features.
[2021-03-21 03:15:14] Showing labelled: True (4184/50000 visible, 16 redundant)
[2021-03-21 03:15:14] Creating trainer with model on device: cuda
[2021-03-21 03:15:14] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 03:16:18] Training accuracy: 0.9999
[2021-03-21 03:16:18] Testing on 10000 data points...
[2021-03-21 03:16:20] Test score for 4200 training labels: 0.8310
[2021-03-21 03:16:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:16:20] Found 50000 unlabelled features.
[2021-03-21 03:16:30] Showing labelled: True (4582/50000 visible, 18 redundant)
[2021-03-21 03:16:30] Creating trainer with model on device: cuda
[2021-03-21 03:16:30] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 03:17:40] Training accuracy: 1.0000
[2021-03-21 03:17:40] Testing on 10000 data points...
[2021-03-21 03:17:42] Test score for 4600 training labels: 0.8336
[2021-03-21 03:17:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:17:42] Found 50000 unlabelled features.
[2021-03-21 03:17:53] Showing labelled: True (4982/50000 visible, 18 redundant)
[2021-03-21 03:17:53] Creating trainer with model on device: cuda
[2021-03-21 03:17:53] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 03:19:08] Training accuracy: 1.0000
[2021-03-21 03:19:08] Testing on 10000 data points...
[2021-03-21 03:19:10] Test score for 5000 training labels: 0.8374
[2021-03-21 03:19:10] Seeded: 7
[2021-03-21 03:19:10] Running: experiment 3
[2021-03-21 03:19:10] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 03:19:10] Creating pretrained=True VGG16...
[2021-03-21 03:19:11] No parameter reset since we are using a pretrained model.
[2021-03-21 03:19:11] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 03:19:11] Creating trainer with model on device: cuda
[2021-03-21 03:19:11] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 03:20:12] Training accuracy: 0.9996
[2021-03-21 03:20:12] Testing on 10000 data points...
[2021-03-21 03:20:14] Test score for 1000 training labels: 0.7396
[2021-03-21 03:20:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:20:14] Found 50000 unlabelled features.
[2021-03-21 03:20:24] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:20:27] Searching for coresets with 1 beams...
[2021-03-21 03:20:29] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 03:20:29] Creating trainer with model on device: cuda
[2021-03-21 03:20:29] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 03:20:50] Training accuracy: 0.9971
[2021-03-21 03:20:50] Testing on 10000 data points...
[2021-03-21 03:20:52] Test score for 1400 training labels: 0.7555
[2021-03-21 03:20:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:20:52] Found 50000 unlabelled features.
[2021-03-21 03:21:02] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:21:07] Searching for coresets with 1 beams...
[2021-03-21 03:21:08] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 03:21:08] Creating trainer with model on device: cuda
[2021-03-21 03:21:08] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 03:21:35] Training accuracy: 0.9993
[2021-03-21 03:21:35] Testing on 10000 data points...
[2021-03-21 03:21:37] Test score for 1800 training labels: 0.7706
[2021-03-21 03:21:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:21:37] Found 50000 unlabelled features.
[2021-03-21 03:21:48] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:21:53] Searching for coresets with 1 beams...
[2021-03-21 03:21:55] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 03:21:55] Creating trainer with model on device: cuda
[2021-03-21 03:21:55] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 03:22:29] Training accuracy: 0.9993
[2021-03-21 03:22:29] Testing on 10000 data points...
[2021-03-21 03:22:31] Test score for 2200 training labels: 0.7860
[2021-03-21 03:22:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:22:31] Found 50000 unlabelled features.
[2021-03-21 03:22:42] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:22:49] Searching for coresets with 1 beams...
[2021-03-21 03:22:50] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 03:22:50] Creating trainer with model on device: cuda
[2021-03-21 03:22:50] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 03:23:30] Training accuracy: 0.9994
[2021-03-21 03:23:30] Testing on 10000 data points...
[2021-03-21 03:23:32] Test score for 2600 training labels: 0.7984
[2021-03-21 03:23:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:23:32] Found 50000 unlabelled features.
[2021-03-21 03:23:42] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:23:50] Searching for coresets with 1 beams...
[2021-03-21 03:23:51] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 03:23:51] Creating trainer with model on device: cuda
[2021-03-21 03:23:51] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 03:24:36] Training accuracy: 0.9988
[2021-03-21 03:24:36] Testing on 10000 data points...
[2021-03-21 03:24:38] Test score for 3000 training labels: 0.8070
[2021-03-21 03:24:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:24:38] Found 50000 unlabelled features.
[2021-03-21 03:24:49] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:24:58] Searching for coresets with 1 beams...
[2021-03-21 03:24:59] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 03:24:59] Creating trainer with model on device: cuda
[2021-03-21 03:24:59] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 03:25:51] Training accuracy: 0.9992
[2021-03-21 03:25:51] Testing on 10000 data points...
[2021-03-21 03:25:53] Test score for 3400 training labels: 0.8140
[2021-03-21 03:25:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:25:53] Found 50000 unlabelled features.
[2021-03-21 03:26:05] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:26:15] Searching for coresets with 1 beams...
[2021-03-21 03:26:17] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 03:26:17] Creating trainer with model on device: cuda
[2021-03-21 03:26:17] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 03:27:15] Training accuracy: 0.9994
[2021-03-21 03:27:15] Testing on 10000 data points...
[2021-03-21 03:27:17] Test score for 3800 training labels: 0.8157
[2021-03-21 03:27:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:27:17] Found 50000 unlabelled features.
[2021-03-21 03:27:28] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:27:39] Searching for coresets with 1 beams...
[2021-03-21 03:27:40] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 03:27:40] Creating trainer with model on device: cuda
[2021-03-21 03:27:40] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 03:28:44] Training accuracy: 0.9999
[2021-03-21 03:28:44] Testing on 10000 data points...
[2021-03-21 03:28:46] Test score for 4200 training labels: 0.8258
[2021-03-21 03:28:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:28:46] Found 50000 unlabelled features.
[2021-03-21 03:28:57] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:29:10] Searching for coresets with 1 beams...
[2021-03-21 03:29:12] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 03:29:12] Creating trainer with model on device: cuda
[2021-03-21 03:29:12] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 03:30:21] Training accuracy: 1.0000
[2021-03-21 03:30:21] Testing on 10000 data points...
[2021-03-21 03:30:23] Test score for 4600 training labels: 0.8279
[2021-03-21 03:30:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:30:23] Found 50000 unlabelled features.
[2021-03-21 03:30:33] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:30:46] Searching for coresets with 1 beams...
[2021-03-21 03:30:48] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 03:30:48] Creating trainer with model on device: cuda
[2021-03-21 03:30:48] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 03:32:03] Training accuracy: 1.0000
[2021-03-21 03:32:03] Testing on 10000 data points...
[2021-03-21 03:32:06] Test score for 5000 training labels: 0.8377
[2021-03-21 03:32:06] Seeded: 7
[2021-03-21 03:32:06] Running: experiment 4
[2021-03-21 03:32:06] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 03:32:06] Creating pretrained=True VGG16...
[2021-03-21 03:32:07] No parameter reset since we are using a pretrained model.
[2021-03-21 03:32:07] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 03:32:07] Creating trainer with model on device: cuda
[2021-03-21 03:32:07] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 03:33:08] Training accuracy: 0.9989
[2021-03-21 03:33:08] Testing on 10000 data points...
[2021-03-21 03:33:10] Test score for 1000 training labels: 0.7308
[2021-03-21 03:33:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:33:10] Found 50000 unlabelled features.
[2021-03-21 03:33:21] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:33:24] Searching for coresets with 20 beams...
[2021-03-21 03:33:53] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 03:33:53] Creating trainer with model on device: cuda
[2021-03-21 03:33:53] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 03:34:14] Training accuracy: 0.9977
[2021-03-21 03:34:14] Testing on 10000 data points...
[2021-03-21 03:34:16] Test score for 1400 training labels: 0.7490
[2021-03-21 03:34:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:34:16] Found 50000 unlabelled features.
[2021-03-21 03:34:27] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:34:31] Searching for coresets with 20 beams...
[2021-03-21 03:35:00] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 03:35:00] Creating trainer with model on device: cuda
[2021-03-21 03:35:00] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 03:35:27] Training accuracy: 0.9975
[2021-03-21 03:35:27] Testing on 10000 data points...
[2021-03-21 03:35:29] Test score for 1800 training labels: 0.7703
[2021-03-21 03:35:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:35:29] Found 50000 unlabelled features.
[2021-03-21 03:35:40] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:35:45] Searching for coresets with 20 beams...
[2021-03-21 03:36:14] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 03:36:14] Creating trainer with model on device: cuda
[2021-03-21 03:36:14] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 03:36:48] Training accuracy: 0.9985
[2021-03-21 03:36:48] Testing on 10000 data points...
[2021-03-21 03:36:50] Test score for 2200 training labels: 0.7816
[2021-03-21 03:36:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:36:50] Found 50000 unlabelled features.
[2021-03-21 03:37:01] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:37:08] Searching for coresets with 20 beams...
[2021-03-21 03:37:36] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 03:37:36] Creating trainer with model on device: cuda
[2021-03-21 03:37:36] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 03:38:16] Training accuracy: 0.9991
[2021-03-21 03:38:16] Testing on 10000 data points...
[2021-03-21 03:38:17] Test score for 2600 training labels: 0.7928
[2021-03-21 03:38:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:38:18] Found 50000 unlabelled features.
[2021-03-21 03:38:28] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:38:36] Searching for coresets with 20 beams...
[2021-03-21 03:39:05] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 03:39:05] Creating trainer with model on device: cuda
[2021-03-21 03:39:05] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 03:39:50] Training accuracy: 0.9989
[2021-03-21 03:39:50] Testing on 10000 data points...
[2021-03-21 03:39:52] Test score for 3000 training labels: 0.7988
[2021-03-21 03:39:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:39:52] Found 50000 unlabelled features.
[2021-03-21 03:40:04] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:40:12] Searching for coresets with 20 beams...
[2021-03-21 03:40:41] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 03:40:41] Creating trainer with model on device: cuda
[2021-03-21 03:40:41] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 03:41:32] Training accuracy: 0.9998
[2021-03-21 03:41:32] Testing on 10000 data points...
[2021-03-21 03:41:34] Test score for 3400 training labels: 0.8142
[2021-03-21 03:41:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:41:34] Found 50000 unlabelled features.
[2021-03-21 03:41:45] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:41:55] Searching for coresets with 20 beams...
[2021-03-21 03:42:24] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 03:42:24] Creating trainer with model on device: cuda
[2021-03-21 03:42:24] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 03:43:22] Training accuracy: 0.9995
[2021-03-21 03:43:22] Testing on 10000 data points...
[2021-03-21 03:43:24] Test score for 3800 training labels: 0.8162
[2021-03-21 03:43:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:43:24] Found 50000 unlabelled features.
[2021-03-21 03:43:35] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:43:46] Searching for coresets with 20 beams...
[2021-03-21 03:44:15] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 03:44:15] Creating trainer with model on device: cuda
[2021-03-21 03:44:15] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 03:45:18] Training accuracy: 1.0000
[2021-03-21 03:45:18] Testing on 10000 data points...
[2021-03-21 03:45:20] Test score for 4200 training labels: 0.8241
[2021-03-21 03:45:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:45:20] Found 50000 unlabelled features.
[2021-03-21 03:45:32] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:45:44] Searching for coresets with 20 beams...
[2021-03-21 03:46:13] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 03:46:13] Creating trainer with model on device: cuda
[2021-03-21 03:46:13] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 03:47:23] Training accuracy: 1.0000
[2021-03-21 03:47:23] Testing on 10000 data points...
[2021-03-21 03:47:25] Test score for 4600 training labels: 0.8314
[2021-03-21 03:47:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:47:25] Found 50000 unlabelled features.
[2021-03-21 03:47:37] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:47:51] Searching for coresets with 20 beams...
[2021-03-21 03:48:20] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 03:48:20] Creating trainer with model on device: cuda
[2021-03-21 03:48:20] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 03:49:36] Training accuracy: 0.9997
[2021-03-21 03:49:36] Testing on 10000 data points...
[2021-03-21 03:49:38] Test score for 5000 training labels: 0.8341
[2021-03-21 03:49:38] Seeded: 7
[2021-03-21 03:49:38] Running: experiment 5
[2021-03-21 03:49:38] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 03:49:38] Creating pretrained=True VGG16...
[2021-03-21 03:49:40] No parameter reset since we are using a pretrained model.
[2021-03-21 03:49:40] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 03:49:40] Creating trainer with model on device: cuda
[2021-03-21 03:49:40] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 03:50:40] Training accuracy: 0.9984
[2021-03-21 03:50:40] Testing on 10000 data points...
[2021-03-21 03:50:42] Test score for 1000 training labels: 0.7124
[2021-03-21 03:50:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:50:42] Found 50000 unlabelled features.
[2021-03-21 03:50:53] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:50:56] Searching for coresets with 1 beams...
[2021-03-21 03:50:58] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 03:50:58] Creating trainer with model on device: cuda
[2021-03-21 03:50:58] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 03:51:19] Training accuracy: 0.9967
[2021-03-21 03:51:19] Testing on 10000 data points...
[2021-03-21 03:51:21] Test score for 1400 training labels: 0.7484
[2021-03-21 03:51:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:51:21] Found 50000 unlabelled features.
[2021-03-21 03:51:31] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:51:36] Searching for coresets with 1 beams...
[2021-03-21 03:51:37] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 03:51:37] Creating trainer with model on device: cuda
[2021-03-21 03:51:37] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 03:52:04] Training accuracy: 0.9977
[2021-03-21 03:52:04] Testing on 10000 data points...
[2021-03-21 03:52:06] Test score for 1800 training labels: 0.7700
[2021-03-21 03:52:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:52:06] Found 50000 unlabelled features.
[2021-03-21 03:52:17] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:52:22] Searching for coresets with 1 beams...
[2021-03-21 03:52:24] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 03:52:24] Creating trainer with model on device: cuda
[2021-03-21 03:52:24] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 03:52:57] Training accuracy: 1.0000
[2021-03-21 03:52:57] Testing on 10000 data points...
[2021-03-21 03:52:59] Test score for 2200 training labels: 0.7836
[2021-03-21 03:52:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:52:59] Found 50000 unlabelled features.
[2021-03-21 03:53:10] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:53:17] Searching for coresets with 1 beams...
[2021-03-21 03:53:18] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 03:53:18] Creating trainer with model on device: cuda
[2021-03-21 03:53:18] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 03:53:58] Training accuracy: 0.9971
[2021-03-21 03:53:58] Testing on 10000 data points...
[2021-03-21 03:54:00] Test score for 2600 training labels: 0.7932
[2021-03-21 03:54:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:54:00] Found 50000 unlabelled features.
[2021-03-21 03:54:11] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:54:19] Searching for coresets with 1 beams...
[2021-03-21 03:54:20] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 03:54:20] Creating trainer with model on device: cuda
[2021-03-21 03:54:20] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 03:55:05] Training accuracy: 1.0000
[2021-03-21 03:55:05] Testing on 10000 data points...
[2021-03-21 03:55:07] Test score for 3000 training labels: 0.8052
[2021-03-21 03:55:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:55:07] Found 50000 unlabelled features.
[2021-03-21 03:55:19] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:55:27] Searching for coresets with 1 beams...
[2021-03-21 03:55:28] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 03:55:28] Creating trainer with model on device: cuda
[2021-03-21 03:55:28] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 03:56:20] Training accuracy: 1.0000
[2021-03-21 03:56:20] Testing on 10000 data points...
[2021-03-21 03:56:22] Test score for 3400 training labels: 0.8093
[2021-03-21 03:56:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:56:22] Found 50000 unlabelled features.
[2021-03-21 03:56:33] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:56:44] Searching for coresets with 1 beams...
[2021-03-21 03:56:45] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 03:56:45] Creating trainer with model on device: cuda
[2021-03-21 03:56:45] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 03:57:43] Training accuracy: 1.0000
[2021-03-21 03:57:43] Testing on 10000 data points...
[2021-03-21 03:57:45] Test score for 3800 training labels: 0.8133
[2021-03-21 03:57:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:57:45] Found 50000 unlabelled features.
[2021-03-21 03:57:56] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:58:08] Searching for coresets with 1 beams...
[2021-03-21 03:58:09] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 03:58:09] Creating trainer with model on device: cuda
[2021-03-21 03:58:09] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 03:59:13] Training accuracy: 0.9997
[2021-03-21 03:59:13] Testing on 10000 data points...
[2021-03-21 03:59:15] Test score for 4200 training labels: 0.8190
[2021-03-21 03:59:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 03:59:15] Found 50000 unlabelled features.
[2021-03-21 03:59:26] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 03:59:40] Searching for coresets with 1 beams...
[2021-03-21 03:59:41] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 03:59:41] Creating trainer with model on device: cuda
[2021-03-21 03:59:41] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 04:00:51] Training accuracy: 1.0000
[2021-03-21 04:00:51] Testing on 10000 data points...
[2021-03-21 04:00:53] Test score for 4600 training labels: 0.8258
[2021-03-21 04:00:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:00:53] Found 50000 unlabelled features.
[2021-03-21 04:01:05] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:01:18] Searching for coresets with 1 beams...
[2021-03-21 04:01:20] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 04:01:20] Creating trainer with model on device: cuda
[2021-03-21 04:01:20] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 04:02:36] Training accuracy: 1.0000
[2021-03-21 04:02:36] Testing on 10000 data points...
[2021-03-21 04:02:38] Test score for 5000 training labels: 0.8318
[2021-03-21 04:02:38] Seeded: 7
[2021-03-21 04:02:38] Running: experiment 6
[2021-03-21 04:02:38] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 04:02:38] Creating pretrained=True VGG16...
[2021-03-21 04:02:39] No parameter reset since we are using a pretrained model.
[2021-03-21 04:02:39] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 04:02:39] Creating trainer with model on device: cuda
[2021-03-21 04:02:39] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 04:03:39] Training accuracy: 0.9996
[2021-03-21 04:03:39] Testing on 10000 data points...
[2021-03-21 04:03:41] Test score for 1000 training labels: 0.7391
[2021-03-21 04:03:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:03:41] Found 50000 unlabelled features.
[2021-03-21 04:03:52] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:03:55] Searching for coresets with 20 beams...
[2021-03-21 04:04:24] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 04:04:24] Creating trainer with model on device: cuda
[2021-03-21 04:04:24] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 04:04:45] Training accuracy: 0.9980
[2021-03-21 04:04:45] Testing on 10000 data points...
[2021-03-21 04:04:47] Test score for 1400 training labels: 0.7668
[2021-03-21 04:04:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:04:47] Found 50000 unlabelled features.
[2021-03-21 04:04:58] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:05:03] Searching for coresets with 20 beams...
[2021-03-21 04:05:32] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 04:05:32] Creating trainer with model on device: cuda
[2021-03-21 04:05:32] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 04:05:59] Training accuracy: 0.9993
[2021-03-21 04:05:59] Testing on 10000 data points...
[2021-03-21 04:06:01] Test score for 1800 training labels: 0.7772
[2021-03-21 04:06:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:06:01] Found 50000 unlabelled features.
[2021-03-21 04:06:12] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:06:17] Searching for coresets with 20 beams...
[2021-03-21 04:06:46] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 04:06:46] Creating trainer with model on device: cuda
[2021-03-21 04:06:46] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 04:07:20] Training accuracy: 0.9983
[2021-03-21 04:07:20] Testing on 10000 data points...
[2021-03-21 04:07:22] Test score for 2200 training labels: 0.7880
[2021-03-21 04:07:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:07:22] Found 50000 unlabelled features.
[2021-03-21 04:07:33] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:07:39] Searching for coresets with 20 beams...
[2021-03-21 04:08:08] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 04:08:08] Creating trainer with model on device: cuda
[2021-03-21 04:08:08] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 04:08:48] Training accuracy: 0.9999
[2021-03-21 04:08:48] Testing on 10000 data points...
[2021-03-21 04:08:50] Test score for 2600 training labels: 0.8025
[2021-03-21 04:08:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:08:50] Found 50000 unlabelled features.
[2021-03-21 04:09:01] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:09:09] Searching for coresets with 20 beams...
[2021-03-21 04:09:38] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 04:09:38] Creating trainer with model on device: cuda
[2021-03-21 04:09:38] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 04:10:23] Training accuracy: 0.9993
[2021-03-21 04:10:23] Testing on 10000 data points...
[2021-03-21 04:10:25] Test score for 3000 training labels: 0.8102
[2021-03-21 04:10:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:10:25] Found 50000 unlabelled features.
[2021-03-21 04:10:36] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:10:45] Searching for coresets with 20 beams...
[2021-03-21 04:11:14] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 04:11:14] Creating trainer with model on device: cuda
[2021-03-21 04:11:14] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 04:12:05] Training accuracy: 1.0000
[2021-03-21 04:12:05] Testing on 10000 data points...
[2021-03-21 04:12:07] Test score for 3400 training labels: 0.8249
[2021-03-21 04:12:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:12:07] Found 50000 unlabelled features.
[2021-03-21 04:12:19] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:12:29] Searching for coresets with 20 beams...
[2021-03-21 04:12:58] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 04:12:58] Creating trainer with model on device: cuda
[2021-03-21 04:12:58] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 04:13:56] Training accuracy: 0.9998
[2021-03-21 04:13:56] Testing on 10000 data points...
[2021-03-21 04:13:58] Test score for 3800 training labels: 0.8254
[2021-03-21 04:13:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:13:58] Found 50000 unlabelled features.
[2021-03-21 04:14:09] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:14:20] Searching for coresets with 20 beams...
[2021-03-21 04:14:49] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 04:14:49] Creating trainer with model on device: cuda
[2021-03-21 04:14:49] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 04:15:53] Training accuracy: 1.0000
[2021-03-21 04:15:53] Testing on 10000 data points...
[2021-03-21 04:15:55] Test score for 4200 training labels: 0.8330
[2021-03-21 04:15:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:15:55] Found 50000 unlabelled features.
[2021-03-21 04:16:07] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:16:19] Searching for coresets with 20 beams...
[2021-03-21 04:16:47] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 04:16:47] Creating trainer with model on device: cuda
[2021-03-21 04:16:47] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 04:17:58] Training accuracy: 1.0000
[2021-03-21 04:17:58] Testing on 10000 data points...
[2021-03-21 04:18:00] Test score for 4600 training labels: 0.8389
[2021-03-21 04:18:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:18:00] Found 50000 unlabelled features.
[2021-03-21 04:18:11] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:18:24] Searching for coresets with 20 beams...
[2021-03-21 04:18:53] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 04:18:53] Creating trainer with model on device: cuda
[2021-03-21 04:18:53] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 04:20:08] Training accuracy: 1.0000
[2021-03-21 04:20:08] Testing on 10000 data points...
[2021-03-21 04:20:10] Test score for 5000 training labels: 0.8452
[2021-03-21 04:20:10] Seeded: 7
[2021-03-21 04:20:10] Running: experiment 7
[2021-03-21 04:20:10] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 04:20:10] Creating pretrained=True VGG16...
[2021-03-21 04:20:12] No parameter reset since we are using a pretrained model.
[2021-03-21 04:20:12] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 04:20:12] Creating trainer with model on device: cuda
[2021-03-21 04:20:12] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 04:21:11] Training accuracy: 0.9947
[2021-03-21 04:21:11] Testing on 10000 data points...
[2021-03-21 04:21:13] Test score for 1000 training labels: 0.7272
[2021-03-21 04:21:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:21:13] Found 50000 unlabelled features.
[2021-03-21 04:21:24] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:21:27] Searching for coresets with 20 beams...
[2021-03-21 04:21:55] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 04:21:55] Creating trainer with model on device: cuda
[2021-03-21 04:21:55] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 04:22:17] Training accuracy: 0.9975
[2021-03-21 04:22:17] Testing on 10000 data points...
[2021-03-21 04:22:19] Test score for 1400 training labels: 0.7638
[2021-03-21 04:22:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:22:19] Found 50000 unlabelled features.
[2021-03-21 04:22:30] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:22:34] Searching for coresets with 20 beams...
[2021-03-21 04:23:02] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 04:23:02] Creating trainer with model on device: cuda
[2021-03-21 04:23:02] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 04:23:29] Training accuracy: 0.9996
[2021-03-21 04:23:29] Testing on 10000 data points...
[2021-03-21 04:23:31] Test score for 1800 training labels: 0.7757
[2021-03-21 04:23:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:23:31] Found 50000 unlabelled features.
[2021-03-21 04:23:43] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:23:48] Searching for coresets with 20 beams...
[2021-03-21 04:24:16] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 04:24:16] Creating trainer with model on device: cuda
[2021-03-21 04:24:16] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 04:24:50] Training accuracy: 0.9487
[2021-03-21 04:24:50] Testing on 10000 data points...
[2021-03-21 04:24:52] Test score for 2200 training labels: 0.7700
[2021-03-21 04:24:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:24:52] Found 50000 unlabelled features.
[2021-03-21 04:25:04] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:25:11] Searching for coresets with 20 beams...
[2021-03-21 04:25:39] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 04:25:39] Creating trainer with model on device: cuda
[2021-03-21 04:25:39] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 04:26:19] Training accuracy: 0.9991
[2021-03-21 04:26:19] Testing on 10000 data points...
[2021-03-21 04:26:21] Test score for 2600 training labels: 0.8086
[2021-03-21 04:26:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:26:21] Found 50000 unlabelled features.
[2021-03-21 04:26:32] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:26:40] Searching for coresets with 20 beams...
[2021-03-21 04:27:08] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 04:27:08] Creating trainer with model on device: cuda
[2021-03-21 04:27:08] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 04:27:54] Training accuracy: 0.9995
[2021-03-21 04:27:54] Testing on 10000 data points...
[2021-03-21 04:27:55] Test score for 3000 training labels: 0.8138
[2021-03-21 04:27:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:27:55] Found 50000 unlabelled features.
[2021-03-21 04:28:06] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:28:15] Searching for coresets with 20 beams...
[2021-03-21 04:28:43] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 04:28:43] Creating trainer with model on device: cuda
[2021-03-21 04:28:43] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 04:29:34] Training accuracy: 0.9992
[2021-03-21 04:29:34] Testing on 10000 data points...
[2021-03-21 04:29:37] Test score for 3400 training labels: 0.8188
[2021-03-21 04:29:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:29:37] Found 50000 unlabelled features.
[2021-03-21 04:29:48] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:29:58] Searching for coresets with 20 beams...
[2021-03-21 04:30:27] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 04:30:27] Creating trainer with model on device: cuda
[2021-03-21 04:30:27] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 04:31:25] Training accuracy: 0.9994
[2021-03-21 04:31:25] Testing on 10000 data points...
[2021-03-21 04:31:27] Test score for 3800 training labels: 0.8257
[2021-03-21 04:31:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:31:27] Found 50000 unlabelled features.
[2021-03-21 04:31:38] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:31:49] Searching for coresets with 20 beams...
[2021-03-21 04:32:17] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 04:32:17] Creating trainer with model on device: cuda
[2021-03-21 04:32:17] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 04:33:21] Training accuracy: 1.0000
[2021-03-21 04:33:21] Testing on 10000 data points...
[2021-03-21 04:33:23] Test score for 4200 training labels: 0.8324
[2021-03-21 04:33:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:33:23] Found 50000 unlabelled features.
[2021-03-21 04:33:35] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:33:48] Searching for coresets with 20 beams...
[2021-03-21 04:34:16] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 04:34:16] Creating trainer with model on device: cuda
[2021-03-21 04:34:16] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 04:35:26] Training accuracy: 1.0000
[2021-03-21 04:35:26] Testing on 10000 data points...
[2021-03-21 04:35:28] Test score for 4600 training labels: 0.8392
[2021-03-21 04:35:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:35:28] Found 50000 unlabelled features.
[2021-03-21 04:35:39] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:35:53] Searching for coresets with 20 beams...
[2021-03-21 04:36:22] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 04:36:22] Creating trainer with model on device: cuda
[2021-03-21 04:36:22] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 04:37:38] Training accuracy: 1.0000
[2021-03-21 04:37:38] Testing on 10000 data points...
[2021-03-21 04:37:40] Test score for 5000 training labels: 0.8401
[2021-03-21 04:37:40] Experiment repeat 3/5
[2021-03-21 04:37:40] Seeded: 7
[2021-03-21 04:37:40] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 04:37:40] Randomly labelled 1000/50000
[2021-03-21 04:37:40] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 04:37:40] Seeded: 8
[2021-03-21 04:37:40] Running: experiment 0
[2021-03-21 04:37:40] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 04:37:40] Creating pretrained=True VGG16...
[2021-03-21 04:37:42] No parameter reset since we are using a pretrained model.
[2021-03-21 04:37:42] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 04:37:42] Creating trainer with model on device: cuda
[2021-03-21 04:37:42] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 04:38:42] Training accuracy: 0.9993
[2021-03-21 04:38:42] Testing on 10000 data points...
[2021-03-21 04:38:44] Test score for 1000 training labels: 0.7346
[2021-03-21 04:38:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:38:44] Found 50000 unlabelled features.
[2021-03-21 04:38:50] Showing labelled: True (1392/50000 visible, 8 redundant)
[2021-03-21 04:38:50] Creating trainer with model on device: cuda
[2021-03-21 04:38:50] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 04:39:11] Training accuracy: 0.9992
[2021-03-21 04:39:11] Testing on 10000 data points...
[2021-03-21 04:39:13] Test score for 1400 training labels: 0.7513
[2021-03-21 04:39:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:39:13] Found 50000 unlabelled features.
[2021-03-21 04:39:18] Showing labelled: True (1783/50000 visible, 17 redundant)
[2021-03-21 04:39:18] Creating trainer with model on device: cuda
[2021-03-21 04:39:18] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 04:39:45] Training accuracy: 0.9990
[2021-03-21 04:39:45] Testing on 10000 data points...
[2021-03-21 04:39:47] Test score for 1800 training labels: 0.7573
[2021-03-21 04:39:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:39:47] Found 50000 unlabelled features.
[2021-03-21 04:39:52] Showing labelled: True (2170/50000 visible, 30 redundant)
[2021-03-21 04:39:52] Creating trainer with model on device: cuda
[2021-03-21 04:39:52] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 04:40:26] Training accuracy: 0.9994
[2021-03-21 04:40:26] Testing on 10000 data points...
[2021-03-21 04:40:28] Test score for 2200 training labels: 0.7700
[2021-03-21 04:40:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:40:28] Found 50000 unlabelled features.
[2021-03-21 04:40:33] Showing labelled: True (2549/50000 visible, 51 redundant)
[2021-03-21 04:40:33] Creating trainer with model on device: cuda
[2021-03-21 04:40:33] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 04:41:12] Training accuracy: 1.0000
[2021-03-21 04:41:12] Testing on 10000 data points...
[2021-03-21 04:41:14] Test score for 2600 training labels: 0.7721
[2021-03-21 04:41:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:41:14] Found 50000 unlabelled features.
[2021-03-21 04:41:19] Showing labelled: True (2927/50000 visible, 73 redundant)
[2021-03-21 04:41:19] Creating trainer with model on device: cuda
[2021-03-21 04:41:19] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 04:42:04] Training accuracy: 0.9999
[2021-03-21 04:42:04] Testing on 10000 data points...
[2021-03-21 04:42:06] Test score for 3000 training labels: 0.7775
[2021-03-21 04:42:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:42:06] Found 50000 unlabelled features.
[2021-03-21 04:42:11] Showing labelled: True (3306/50000 visible, 94 redundant)
[2021-03-21 04:42:11] Creating trainer with model on device: cuda
[2021-03-21 04:42:11] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 04:43:02] Training accuracy: 1.0000
[2021-03-21 04:43:02] Testing on 10000 data points...
[2021-03-21 04:43:04] Test score for 3400 training labels: 0.7853
[2021-03-21 04:43:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:43:04] Found 50000 unlabelled features.
[2021-03-21 04:43:11] Showing labelled: True (3681/50000 visible, 119 redundant)
[2021-03-21 04:43:11] Creating trainer with model on device: cuda
[2021-03-21 04:43:11] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 04:44:09] Training accuracy: 1.0000
[2021-03-21 04:44:09] Testing on 10000 data points...
[2021-03-21 04:44:11] Test score for 3800 training labels: 0.7870
[2021-03-21 04:44:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:44:11] Found 50000 unlabelled features.
[2021-03-21 04:44:17] Showing labelled: True (4049/50000 visible, 151 redundant)
[2021-03-21 04:44:17] Creating trainer with model on device: cuda
[2021-03-21 04:44:17] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 04:45:21] Training accuracy: 1.0000
[2021-03-21 04:45:21] Testing on 10000 data points...
[2021-03-21 04:45:22] Test score for 4200 training labels: 0.7878
[2021-03-21 04:45:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:45:22] Found 50000 unlabelled features.
[2021-03-21 04:45:28] Showing labelled: True (4412/50000 visible, 188 redundant)
[2021-03-21 04:45:28] Creating trainer with model on device: cuda
[2021-03-21 04:45:28] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 04:46:38] Training accuracy: 0.9995
[2021-03-21 04:46:38] Testing on 10000 data points...
[2021-03-21 04:46:40] Test score for 4600 training labels: 0.7965
[2021-03-21 04:46:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:46:40] Found 50000 unlabelled features.
[2021-03-21 04:46:46] Showing labelled: True (4781/50000 visible, 219 redundant)
[2021-03-21 04:46:46] Creating trainer with model on device: cuda
[2021-03-21 04:46:46] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 04:48:01] Training accuracy: 0.9998
[2021-03-21 04:48:01] Testing on 10000 data points...
[2021-03-21 04:48:03] Test score for 5000 training labels: 0.7977
[2021-03-21 04:48:03] Seeded: 8
[2021-03-21 04:48:03] Running: experiment 1
[2021-03-21 04:48:03] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 04:48:03] Creating pretrained=True VGG16...
[2021-03-21 04:48:05] No parameter reset since we are using a pretrained model.
[2021-03-21 04:48:05] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 04:48:05] Creating trainer with model on device: cuda
[2021-03-21 04:48:05] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 04:49:05] Training accuracy: 0.9985
[2021-03-21 04:49:05] Testing on 10000 data points...
[2021-03-21 04:49:07] Test score for 1000 training labels: 0.7265
[2021-03-21 04:49:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:49:07] Found 50000 unlabelled features.
[2021-03-21 04:49:18] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:49:21] Searching for coresets greedily...
[2021-03-21 04:49:22] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 04:49:22] Creating trainer with model on device: cuda
[2021-03-21 04:49:22] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 04:49:43] Training accuracy: 0.9988
[2021-03-21 04:49:43] Testing on 10000 data points...
[2021-03-21 04:49:45] Test score for 1400 training labels: 0.7474
[2021-03-21 04:49:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:49:45] Found 50000 unlabelled features.
[2021-03-21 04:49:55] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:49:59] Searching for coresets greedily...
[2021-03-21 04:50:00] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 04:50:00] Creating trainer with model on device: cuda
[2021-03-21 04:50:00] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 04:50:27] Training accuracy: 0.9965
[2021-03-21 04:50:27] Testing on 10000 data points...
[2021-03-21 04:50:30] Test score for 1800 training labels: 0.7525
[2021-03-21 04:50:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:50:30] Found 50000 unlabelled features.
[2021-03-21 04:50:40] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:50:46] Searching for coresets greedily...
[2021-03-21 04:50:46] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 04:50:46] Creating trainer with model on device: cuda
[2021-03-21 04:50:46] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 04:51:20] Training accuracy: 0.9995
[2021-03-21 04:51:20] Testing on 10000 data points...
[2021-03-21 04:51:22] Test score for 2200 training labels: 0.7644
[2021-03-21 04:51:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:51:22] Found 50000 unlabelled features.
[2021-03-21 04:51:32] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:51:38] Searching for coresets greedily...
[2021-03-21 04:51:39] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 04:51:39] Creating trainer with model on device: cuda
[2021-03-21 04:51:39] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 04:52:18] Training accuracy: 0.9996
[2021-03-21 04:52:18] Testing on 10000 data points...
[2021-03-21 04:52:20] Test score for 2600 training labels: 0.7827
[2021-03-21 04:52:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:52:20] Found 50000 unlabelled features.
[2021-03-21 04:52:31] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:52:39] Searching for coresets greedily...
[2021-03-21 04:52:39] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 04:52:39] Creating trainer with model on device: cuda
[2021-03-21 04:52:39] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 04:53:25] Training accuracy: 0.9995
[2021-03-21 04:53:25] Testing on 10000 data points...
[2021-03-21 04:53:27] Test score for 3000 training labels: 0.7974
[2021-03-21 04:53:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:53:27] Found 50000 unlabelled features.
[2021-03-21 04:53:38] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:53:47] Searching for coresets greedily...
[2021-03-21 04:53:48] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 04:53:48] Creating trainer with model on device: cuda
[2021-03-21 04:53:48] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 04:54:39] Training accuracy: 0.9998
[2021-03-21 04:54:39] Testing on 10000 data points...
[2021-03-21 04:54:41] Test score for 3400 training labels: 0.8014
[2021-03-21 04:54:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:54:41] Found 50000 unlabelled features.
[2021-03-21 04:54:52] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:55:04] Searching for coresets greedily...
[2021-03-21 04:55:04] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 04:55:04] Creating trainer with model on device: cuda
[2021-03-21 04:55:04] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 04:56:02] Training accuracy: 0.9993
[2021-03-21 04:56:02] Testing on 10000 data points...
[2021-03-21 04:56:04] Test score for 3800 training labels: 0.8106
[2021-03-21 04:56:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:56:04] Found 50000 unlabelled features.
[2021-03-21 04:56:15] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:56:27] Searching for coresets greedily...
[2021-03-21 04:56:28] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 04:56:28] Creating trainer with model on device: cuda
[2021-03-21 04:56:28] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 04:57:32] Training accuracy: 1.0000
[2021-03-21 04:57:32] Testing on 10000 data points...
[2021-03-21 04:57:34] Test score for 4200 training labels: 0.8147
[2021-03-21 04:57:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:57:34] Found 50000 unlabelled features.
[2021-03-21 04:57:44] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:57:56] Searching for coresets greedily...
[2021-03-21 04:57:57] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 04:57:57] Creating trainer with model on device: cuda
[2021-03-21 04:57:57] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 04:59:06] Training accuracy: 0.9998
[2021-03-21 04:59:06] Testing on 10000 data points...
[2021-03-21 04:59:08] Test score for 4600 training labels: 0.8236
[2021-03-21 04:59:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 04:59:08] Found 50000 unlabelled features.
[2021-03-21 04:59:19] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 04:59:33] Searching for coresets greedily...
[2021-03-21 04:59:33] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 04:59:33] Creating trainer with model on device: cuda
[2021-03-21 04:59:33] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 05:00:50] Training accuracy: 1.0000
[2021-03-21 05:00:50] Testing on 10000 data points...
[2021-03-21 05:00:52] Test score for 5000 training labels: 0.8283
[2021-03-21 05:00:52] Seeded: 8
[2021-03-21 05:00:52] Running: experiment 2
[2021-03-21 05:00:52] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 05:00:52] Creating pretrained=True VGG16...
[2021-03-21 05:00:54] No parameter reset since we are using a pretrained model.
[2021-03-21 05:00:54] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 05:00:54] Creating trainer with model on device: cuda
[2021-03-21 05:00:54] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 05:01:55] Training accuracy: 0.9986
[2021-03-21 05:01:55] Testing on 10000 data points...
[2021-03-21 05:01:57] Test score for 1000 training labels: 0.7252
[2021-03-21 05:01:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:01:57] Found 50000 unlabelled features.
[2021-03-21 05:02:08] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 05:02:08] Creating trainer with model on device: cuda
[2021-03-21 05:02:08] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 05:02:29] Training accuracy: 0.9986
[2021-03-21 05:02:29] Testing on 10000 data points...
[2021-03-21 05:02:31] Test score for 1400 training labels: 0.7476
[2021-03-21 05:02:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:02:31] Found 50000 unlabelled features.
[2021-03-21 05:02:41] Showing labelled: True (1798/50000 visible, 2 redundant)
[2021-03-21 05:02:41] Creating trainer with model on device: cuda
[2021-03-21 05:02:41] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 05:03:09] Training accuracy: 0.9990
[2021-03-21 05:03:09] Testing on 10000 data points...
[2021-03-21 05:03:11] Test score for 1800 training labels: 0.7670
[2021-03-21 05:03:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:03:11] Found 50000 unlabelled features.
[2021-03-21 05:03:21] Showing labelled: True (2198/50000 visible, 2 redundant)
[2021-03-21 05:03:21] Creating trainer with model on device: cuda
[2021-03-21 05:03:21] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 05:03:55] Training accuracy: 0.9999
[2021-03-21 05:03:55] Testing on 10000 data points...
[2021-03-21 05:03:56] Test score for 2200 training labels: 0.7867
[2021-03-21 05:03:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:03:56] Found 50000 unlabelled features.
[2021-03-21 05:04:07] Showing labelled: True (2598/50000 visible, 2 redundant)
[2021-03-21 05:04:07] Creating trainer with model on device: cuda
[2021-03-21 05:04:07] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 05:04:46] Training accuracy: 1.0000
[2021-03-21 05:04:46] Testing on 10000 data points...
[2021-03-21 05:04:48] Test score for 2600 training labels: 0.7978
[2021-03-21 05:04:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:04:48] Found 50000 unlabelled features.
[2021-03-21 05:04:59] Showing labelled: True (2998/50000 visible, 2 redundant)
[2021-03-21 05:04:59] Creating trainer with model on device: cuda
[2021-03-21 05:04:59] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 05:05:44] Training accuracy: 0.9992
[2021-03-21 05:05:44] Testing on 10000 data points...
[2021-03-21 05:05:46] Test score for 3000 training labels: 0.8068
[2021-03-21 05:05:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:05:46] Found 50000 unlabelled features.
[2021-03-21 05:05:56] Showing labelled: True (3398/50000 visible, 2 redundant)
[2021-03-21 05:05:56] Creating trainer with model on device: cuda
[2021-03-21 05:05:56] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 05:06:47] Training accuracy: 1.0000
[2021-03-21 05:06:47] Testing on 10000 data points...
[2021-03-21 05:06:49] Test score for 3400 training labels: 0.8173
[2021-03-21 05:06:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:06:49] Found 50000 unlabelled features.
[2021-03-21 05:07:00] Showing labelled: True (3798/50000 visible, 2 redundant)
[2021-03-21 05:07:00] Creating trainer with model on device: cuda
[2021-03-21 05:07:00] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 05:07:57] Training accuracy: 1.0000
[2021-03-21 05:07:57] Testing on 10000 data points...
[2021-03-21 05:07:59] Test score for 3800 training labels: 0.8193
[2021-03-21 05:07:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:07:59] Found 50000 unlabelled features.
[2021-03-21 05:08:10] Showing labelled: True (4198/50000 visible, 2 redundant)
[2021-03-21 05:08:10] Creating trainer with model on device: cuda
[2021-03-21 05:08:10] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 05:09:13] Training accuracy: 0.9992
[2021-03-21 05:09:13] Testing on 10000 data points...
[2021-03-21 05:09:15] Test score for 4200 training labels: 0.8260
[2021-03-21 05:09:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:09:15] Found 50000 unlabelled features.
[2021-03-21 05:09:24] Showing labelled: True (4597/50000 visible, 3 redundant)
[2021-03-21 05:09:24] Creating trainer with model on device: cuda
[2021-03-21 05:09:24] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 05:10:33] Training accuracy: 0.9995
[2021-03-21 05:10:33] Testing on 10000 data points...
[2021-03-21 05:10:35] Test score for 4600 training labels: 0.8286
[2021-03-21 05:10:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:10:35] Found 50000 unlabelled features.
[2021-03-21 05:10:46] Showing labelled: True (4996/50000 visible, 4 redundant)
[2021-03-21 05:10:46] Creating trainer with model on device: cuda
[2021-03-21 05:10:46] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 05:12:01] Training accuracy: 1.0000
[2021-03-21 05:12:01] Testing on 10000 data points...
[2021-03-21 05:12:03] Test score for 5000 training labels: 0.8338
[2021-03-21 05:12:03] Seeded: 8
[2021-03-21 05:12:03] Running: experiment 3
[2021-03-21 05:12:03] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 05:12:03] Creating pretrained=True VGG16...
[2021-03-21 05:12:05] No parameter reset since we are using a pretrained model.
[2021-03-21 05:12:05] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 05:12:05] Creating trainer with model on device: cuda
[2021-03-21 05:12:05] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 05:13:05] Training accuracy: 0.9999
[2021-03-21 05:13:05] Testing on 10000 data points...
[2021-03-21 05:13:07] Test score for 1000 training labels: 0.7344
[2021-03-21 05:13:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:13:07] Found 50000 unlabelled features.
[2021-03-21 05:13:17] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:13:20] Searching for coresets with 1 beams...
[2021-03-21 05:13:21] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 05:13:21] Creating trainer with model on device: cuda
[2021-03-21 05:13:21] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 05:13:42] Training accuracy: 0.9993
[2021-03-21 05:13:42] Testing on 10000 data points...
[2021-03-21 05:13:44] Test score for 1400 training labels: 0.7447
[2021-03-21 05:13:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:13:44] Found 50000 unlabelled features.
[2021-03-21 05:13:55] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:14:00] Searching for coresets with 1 beams...
[2021-03-21 05:14:01] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 05:14:01] Creating trainer with model on device: cuda
[2021-03-21 05:14:01] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 05:14:28] Training accuracy: 0.9984
[2021-03-21 05:14:28] Testing on 10000 data points...
[2021-03-21 05:14:30] Test score for 1800 training labels: 0.7675
[2021-03-21 05:14:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:14:30] Found 50000 unlabelled features.
[2021-03-21 05:14:40] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:14:46] Searching for coresets with 1 beams...
[2021-03-21 05:14:47] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 05:14:47] Creating trainer with model on device: cuda
[2021-03-21 05:14:47] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 05:15:21] Training accuracy: 0.9990
[2021-03-21 05:15:21] Testing on 10000 data points...
[2021-03-21 05:15:23] Test score for 2200 training labels: 0.7717
[2021-03-21 05:15:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:15:23] Found 50000 unlabelled features.
[2021-03-21 05:15:33] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:15:40] Searching for coresets with 1 beams...
[2021-03-21 05:15:41] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 05:15:41] Creating trainer with model on device: cuda
[2021-03-21 05:15:41] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 05:16:21] Training accuracy: 0.9962
[2021-03-21 05:16:21] Testing on 10000 data points...
[2021-03-21 05:16:22] Test score for 2600 training labels: 0.7774
[2021-03-21 05:16:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:16:22] Found 50000 unlabelled features.
[2021-03-21 05:16:33] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:16:41] Searching for coresets with 1 beams...
[2021-03-21 05:16:42] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 05:16:42] Creating trainer with model on device: cuda
[2021-03-21 05:16:42] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 05:17:27] Training accuracy: 1.0000
[2021-03-21 05:17:27] Testing on 10000 data points...
[2021-03-21 05:17:29] Test score for 3000 training labels: 0.7958
[2021-03-21 05:17:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:17:29] Found 50000 unlabelled features.
[2021-03-21 05:17:41] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:17:49] Searching for coresets with 1 beams...
[2021-03-21 05:17:51] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 05:17:51] Creating trainer with model on device: cuda
[2021-03-21 05:17:51] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 05:18:43] Training accuracy: 0.9994
[2021-03-21 05:18:43] Testing on 10000 data points...
[2021-03-21 05:18:45] Test score for 3400 training labels: 0.8053
[2021-03-21 05:18:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:18:45] Found 50000 unlabelled features.
[2021-03-21 05:18:55] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:19:05] Searching for coresets with 1 beams...
[2021-03-21 05:19:07] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 05:19:07] Creating trainer with model on device: cuda
[2021-03-21 05:19:07] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 05:20:04] Training accuracy: 1.0000
[2021-03-21 05:20:04] Testing on 10000 data points...
[2021-03-21 05:20:06] Test score for 3800 training labels: 0.8168
[2021-03-21 05:20:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:20:06] Found 50000 unlabelled features.
[2021-03-21 05:20:18] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:20:28] Searching for coresets with 1 beams...
[2021-03-21 05:20:30] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 05:20:30] Creating trainer with model on device: cuda
[2021-03-21 05:20:30] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 05:21:33] Training accuracy: 0.9998
[2021-03-21 05:21:33] Testing on 10000 data points...
[2021-03-21 05:21:35] Test score for 4200 training labels: 0.8158
[2021-03-21 05:21:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:21:35] Found 50000 unlabelled features.
[2021-03-21 05:21:45] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:21:58] Searching for coresets with 1 beams...
[2021-03-21 05:22:00] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 05:22:00] Creating trainer with model on device: cuda
[2021-03-21 05:22:00] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 05:23:09] Training accuracy: 1.0000
[2021-03-21 05:23:09] Testing on 10000 data points...
[2021-03-21 05:23:11] Test score for 4600 training labels: 0.8223
[2021-03-21 05:23:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:23:11] Found 50000 unlabelled features.
[2021-03-21 05:23:23] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:23:36] Searching for coresets with 1 beams...
[2021-03-21 05:23:37] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 05:23:37] Creating trainer with model on device: cuda
[2021-03-21 05:23:37] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 05:24:52] Training accuracy: 1.0000
[2021-03-21 05:24:52] Testing on 10000 data points...
[2021-03-21 05:24:54] Test score for 5000 training labels: 0.8274
[2021-03-21 05:24:54] Seeded: 8
[2021-03-21 05:24:54] Running: experiment 4
[2021-03-21 05:24:54] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 05:24:54] Creating pretrained=True VGG16...
[2021-03-21 05:24:56] No parameter reset since we are using a pretrained model.
[2021-03-21 05:24:56] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 05:24:56] Creating trainer with model on device: cuda
[2021-03-21 05:24:56] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 05:25:57] Training accuracy: 0.9992
[2021-03-21 05:25:57] Testing on 10000 data points...
[2021-03-21 05:25:59] Test score for 1000 training labels: 0.7192
[2021-03-21 05:25:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:25:59] Found 50000 unlabelled features.
[2021-03-21 05:26:08] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:26:11] Searching for coresets with 20 beams...
[2021-03-21 05:26:41] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 05:26:41] Creating trainer with model on device: cuda
[2021-03-21 05:26:41] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 05:27:02] Training accuracy: 0.9799
[2021-03-21 05:27:02] Testing on 10000 data points...
[2021-03-21 05:27:04] Test score for 1400 training labels: 0.7403
[2021-03-21 05:27:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:27:04] Found 50000 unlabelled features.
[2021-03-21 05:27:14] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:27:18] Searching for coresets with 20 beams...
[2021-03-21 05:27:46] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 05:27:46] Creating trainer with model on device: cuda
[2021-03-21 05:27:46] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 05:28:14] Training accuracy: 0.9999
[2021-03-21 05:28:14] Testing on 10000 data points...
[2021-03-21 05:28:16] Test score for 1800 training labels: 0.7666
[2021-03-21 05:28:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:28:16] Found 50000 unlabelled features.
[2021-03-21 05:28:27] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:28:32] Searching for coresets with 20 beams...
[2021-03-21 05:29:01] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 05:29:01] Creating trainer with model on device: cuda
[2021-03-21 05:29:01] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 05:29:35] Training accuracy: 1.0000
[2021-03-21 05:29:35] Testing on 10000 data points...
[2021-03-21 05:29:37] Test score for 2200 training labels: 0.7839
[2021-03-21 05:29:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:29:37] Found 50000 unlabelled features.
[2021-03-21 05:29:48] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:29:55] Searching for coresets with 20 beams...
[2021-03-21 05:30:24] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 05:30:24] Creating trainer with model on device: cuda
[2021-03-21 05:30:24] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 05:31:04] Training accuracy: 0.9999
[2021-03-21 05:31:04] Testing on 10000 data points...
[2021-03-21 05:31:06] Test score for 2600 training labels: 0.7953
[2021-03-21 05:31:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:31:06] Found 50000 unlabelled features.
[2021-03-21 05:31:17] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:31:25] Searching for coresets with 20 beams...
[2021-03-21 05:31:54] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 05:31:54] Creating trainer with model on device: cuda
[2021-03-21 05:31:54] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 05:32:39] Training accuracy: 0.9998
[2021-03-21 05:32:39] Testing on 10000 data points...
[2021-03-21 05:32:41] Test score for 3000 training labels: 0.7959
[2021-03-21 05:32:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:32:41] Found 50000 unlabelled features.
[2021-03-21 05:32:52] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:33:01] Searching for coresets with 20 beams...
[2021-03-21 05:33:30] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 05:33:30] Creating trainer with model on device: cuda
[2021-03-21 05:33:30] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 05:34:22] Training accuracy: 1.0000
[2021-03-21 05:34:22] Testing on 10000 data points...
[2021-03-21 05:34:24] Test score for 3400 training labels: 0.8065
[2021-03-21 05:34:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:34:24] Found 50000 unlabelled features.
[2021-03-21 05:34:35] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:34:45] Searching for coresets with 20 beams...
[2021-03-21 05:35:14] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 05:35:14] Creating trainer with model on device: cuda
[2021-03-21 05:35:14] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 05:36:13] Training accuracy: 0.9999
[2021-03-21 05:36:13] Testing on 10000 data points...
[2021-03-21 05:36:15] Test score for 3800 training labels: 0.8079
[2021-03-21 05:36:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:36:15] Found 50000 unlabelled features.
[2021-03-21 05:36:25] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:36:37] Searching for coresets with 20 beams...
[2021-03-21 05:37:06] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 05:37:06] Creating trainer with model on device: cuda
[2021-03-21 05:37:06] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 05:38:09] Training accuracy: 1.0000
[2021-03-21 05:38:09] Testing on 10000 data points...
[2021-03-21 05:38:11] Test score for 4200 training labels: 0.8215
[2021-03-21 05:38:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:38:11] Found 50000 unlabelled features.
[2021-03-21 05:38:23] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:38:35] Searching for coresets with 20 beams...
[2021-03-21 05:39:03] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 05:39:03] Creating trainer with model on device: cuda
[2021-03-21 05:39:03] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 05:40:13] Training accuracy: 1.0000
[2021-03-21 05:40:13] Testing on 10000 data points...
[2021-03-21 05:40:15] Test score for 4600 training labels: 0.8234
[2021-03-21 05:40:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:40:15] Found 50000 unlabelled features.
[2021-03-21 05:40:27] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:40:41] Searching for coresets with 20 beams...
[2021-03-21 05:41:09] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 05:41:09] Creating trainer with model on device: cuda
[2021-03-21 05:41:09] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 05:42:25] Training accuracy: 1.0000
[2021-03-21 05:42:25] Testing on 10000 data points...
[2021-03-21 05:42:27] Test score for 5000 training labels: 0.8313
[2021-03-21 05:42:27] Seeded: 8
[2021-03-21 05:42:27] Running: experiment 5
[2021-03-21 05:42:27] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 05:42:27] Creating pretrained=True VGG16...
[2021-03-21 05:42:29] No parameter reset since we are using a pretrained model.
[2021-03-21 05:42:29] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 05:42:29] Creating trainer with model on device: cuda
[2021-03-21 05:42:29] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 05:43:28] Training accuracy: 0.9997
[2021-03-21 05:43:28] Testing on 10000 data points...
[2021-03-21 05:43:30] Test score for 1000 training labels: 0.7355
[2021-03-21 05:43:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:43:30] Found 50000 unlabelled features.
[2021-03-21 05:43:41] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:43:44] Searching for coresets with 1 beams...
[2021-03-21 05:43:45] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 05:43:45] Creating trainer with model on device: cuda
[2021-03-21 05:43:45] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 05:44:07] Training accuracy: 0.9999
[2021-03-21 05:44:07] Testing on 10000 data points...
[2021-03-21 05:44:09] Test score for 1400 training labels: 0.7567
[2021-03-21 05:44:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:44:09] Found 50000 unlabelled features.
[2021-03-21 05:44:20] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:44:24] Searching for coresets with 1 beams...
[2021-03-21 05:44:25] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 05:44:25] Creating trainer with model on device: cuda
[2021-03-21 05:44:25] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 05:44:53] Training accuracy: 0.9996
[2021-03-21 05:44:53] Testing on 10000 data points...
[2021-03-21 05:44:55] Test score for 1800 training labels: 0.7750
[2021-03-21 05:44:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:44:55] Found 50000 unlabelled features.
[2021-03-21 05:45:06] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:45:11] Searching for coresets with 1 beams...
[2021-03-21 05:45:13] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 05:45:13] Creating trainer with model on device: cuda
[2021-03-21 05:45:13] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 05:45:46] Training accuracy: 0.9996
[2021-03-21 05:45:46] Testing on 10000 data points...
[2021-03-21 05:45:48] Test score for 2200 training labels: 0.7850
[2021-03-21 05:45:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:45:48] Found 50000 unlabelled features.
[2021-03-21 05:45:58] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:46:04] Searching for coresets with 1 beams...
[2021-03-21 05:46:06] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 05:46:06] Creating trainer with model on device: cuda
[2021-03-21 05:46:06] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 05:46:45] Training accuracy: 0.9993
[2021-03-21 05:46:45] Testing on 10000 data points...
[2021-03-21 05:46:47] Test score for 2600 training labels: 0.7974
[2021-03-21 05:46:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:46:47] Found 50000 unlabelled features.
[2021-03-21 05:46:57] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:47:05] Searching for coresets with 1 beams...
[2021-03-21 05:47:07] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 05:47:07] Creating trainer with model on device: cuda
[2021-03-21 05:47:07] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 05:47:52] Training accuracy: 0.9999
[2021-03-21 05:47:52] Testing on 10000 data points...
[2021-03-21 05:47:54] Test score for 3000 training labels: 0.8073
[2021-03-21 05:47:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:47:54] Found 50000 unlabelled features.
[2021-03-21 05:48:05] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:48:14] Searching for coresets with 1 beams...
[2021-03-21 05:48:15] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 05:48:15] Creating trainer with model on device: cuda
[2021-03-21 05:48:15] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 05:49:07] Training accuracy: 0.9998
[2021-03-21 05:49:07] Testing on 10000 data points...
[2021-03-21 05:49:09] Test score for 3400 training labels: 0.8161
[2021-03-21 05:49:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:49:09] Found 50000 unlabelled features.
[2021-03-21 05:49:21] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:49:31] Searching for coresets with 1 beams...
[2021-03-21 05:49:33] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 05:49:33] Creating trainer with model on device: cuda
[2021-03-21 05:49:33] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 05:50:31] Training accuracy: 1.0000
[2021-03-21 05:50:31] Testing on 10000 data points...
[2021-03-21 05:50:33] Test score for 3800 training labels: 0.8194
[2021-03-21 05:50:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:50:33] Found 50000 unlabelled features.
[2021-03-21 05:50:44] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:50:56] Searching for coresets with 1 beams...
[2021-03-21 05:50:57] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 05:50:57] Creating trainer with model on device: cuda
[2021-03-21 05:50:57] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 05:52:00] Training accuracy: 1.0000
[2021-03-21 05:52:00] Testing on 10000 data points...
[2021-03-21 05:52:02] Test score for 4200 training labels: 0.8228
[2021-03-21 05:52:02] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:52:02] Found 50000 unlabelled features.
[2021-03-21 05:52:13] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:52:25] Searching for coresets with 1 beams...
[2021-03-21 05:52:27] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 05:52:27] Creating trainer with model on device: cuda
[2021-03-21 05:52:27] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 05:53:37] Training accuracy: 1.0000
[2021-03-21 05:53:37] Testing on 10000 data points...
[2021-03-21 05:53:39] Test score for 4600 training labels: 0.8319
[2021-03-21 05:53:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:53:39] Found 50000 unlabelled features.
[2021-03-21 05:53:51] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:54:04] Searching for coresets with 1 beams...
[2021-03-21 05:54:05] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 05:54:05] Creating trainer with model on device: cuda
[2021-03-21 05:54:05] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 05:55:21] Training accuracy: 1.0000
[2021-03-21 05:55:21] Testing on 10000 data points...
[2021-03-21 05:55:23] Test score for 5000 training labels: 0.8324
[2021-03-21 05:55:23] Seeded: 8
[2021-03-21 05:55:23] Running: experiment 6
[2021-03-21 05:55:23] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 05:55:23] Creating pretrained=True VGG16...
[2021-03-21 05:55:25] No parameter reset since we are using a pretrained model.
[2021-03-21 05:55:25] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 05:55:25] Creating trainer with model on device: cuda
[2021-03-21 05:55:25] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 05:56:25] Training accuracy: 0.9989
[2021-03-21 05:56:25] Testing on 10000 data points...
[2021-03-21 05:56:28] Test score for 1000 training labels: 0.7363
[2021-03-21 05:56:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:56:28] Found 50000 unlabelled features.
[2021-03-21 05:56:37] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:56:40] Searching for coresets with 20 beams...
[2021-03-21 05:57:09] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 05:57:09] Creating trainer with model on device: cuda
[2021-03-21 05:57:09] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 05:57:31] Training accuracy: 0.9969
[2021-03-21 05:57:31] Testing on 10000 data points...
[2021-03-21 05:57:33] Test score for 1400 training labels: 0.7650
[2021-03-21 05:57:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:57:33] Found 50000 unlabelled features.
[2021-03-21 05:57:44] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:57:48] Searching for coresets with 20 beams...
[2021-03-21 05:58:17] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 05:58:17] Creating trainer with model on device: cuda
[2021-03-21 05:58:17] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 05:58:45] Training accuracy: 0.9994
[2021-03-21 05:58:45] Testing on 10000 data points...
[2021-03-21 05:58:47] Test score for 1800 training labels: 0.7749
[2021-03-21 05:58:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 05:58:47] Found 50000 unlabelled features.
[2021-03-21 05:58:58] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 05:59:04] Searching for coresets with 20 beams...
[2021-03-21 05:59:32] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 05:59:32] Creating trainer with model on device: cuda
[2021-03-21 05:59:32] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 06:00:07] Training accuracy: 0.9997
[2021-03-21 06:00:07] Testing on 10000 data points...
[2021-03-21 06:00:09] Test score for 2200 training labels: 0.7859
[2021-03-21 06:00:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:00:09] Found 50000 unlabelled features.
[2021-03-21 06:00:20] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:00:26] Searching for coresets with 20 beams...
[2021-03-21 06:00:55] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 06:00:55] Creating trainer with model on device: cuda
[2021-03-21 06:00:55] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 06:01:34] Training accuracy: 0.9995
[2021-03-21 06:01:34] Testing on 10000 data points...
[2021-03-21 06:01:36] Test score for 2600 training labels: 0.8007
[2021-03-21 06:01:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:01:36] Found 50000 unlabelled features.
[2021-03-21 06:01:47] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:01:55] Searching for coresets with 20 beams...
[2021-03-21 06:02:24] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 06:02:24] Creating trainer with model on device: cuda
[2021-03-21 06:02:24] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 06:03:09] Training accuracy: 0.9988
[2021-03-21 06:03:09] Testing on 10000 data points...
[2021-03-21 06:03:11] Test score for 3000 training labels: 0.8048
[2021-03-21 06:03:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:03:11] Found 50000 unlabelled features.
[2021-03-21 06:03:21] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:03:30] Searching for coresets with 20 beams...
[2021-03-21 06:03:59] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 06:03:59] Creating trainer with model on device: cuda
[2021-03-21 06:03:59] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 06:04:51] Training accuracy: 0.9999
[2021-03-21 06:04:51] Testing on 10000 data points...
[2021-03-21 06:04:53] Test score for 3400 training labels: 0.8120
[2021-03-21 06:04:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:04:53] Found 50000 unlabelled features.
[2021-03-21 06:05:05] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:05:15] Searching for coresets with 20 beams...
[2021-03-21 06:05:44] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 06:05:44] Creating trainer with model on device: cuda
[2021-03-21 06:05:44] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 06:06:42] Training accuracy: 0.9996
[2021-03-21 06:06:42] Testing on 10000 data points...
[2021-03-21 06:06:44] Test score for 3800 training labels: 0.8193
[2021-03-21 06:06:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:06:44] Found 50000 unlabelled features.
[2021-03-21 06:06:56] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:07:07] Searching for coresets with 20 beams...
[2021-03-21 06:07:36] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 06:07:36] Creating trainer with model on device: cuda
[2021-03-21 06:07:36] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 06:08:39] Training accuracy: 1.0000
[2021-03-21 06:08:39] Testing on 10000 data points...
[2021-03-21 06:08:41] Test score for 4200 training labels: 0.8272
[2021-03-21 06:08:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:08:41] Found 50000 unlabelled features.
[2021-03-21 06:08:53] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:09:06] Searching for coresets with 20 beams...
[2021-03-21 06:09:34] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 06:09:34] Creating trainer with model on device: cuda
[2021-03-21 06:09:34] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 06:10:45] Training accuracy: 0.9993
[2021-03-21 06:10:45] Testing on 10000 data points...
[2021-03-21 06:10:47] Test score for 4600 training labels: 0.8295
[2021-03-21 06:10:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:10:47] Found 50000 unlabelled features.
[2021-03-21 06:10:58] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:11:11] Searching for coresets with 20 beams...
[2021-03-21 06:11:40] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 06:11:40] Creating trainer with model on device: cuda
[2021-03-21 06:11:40] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 06:12:55] Training accuracy: 1.0000
[2021-03-21 06:12:55] Testing on 10000 data points...
[2021-03-21 06:12:57] Test score for 5000 training labels: 0.8377
[2021-03-21 06:12:57] Seeded: 8
[2021-03-21 06:12:57] Running: experiment 7
[2021-03-21 06:12:57] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 06:12:57] Creating pretrained=True VGG16...
[2021-03-21 06:12:59] No parameter reset since we are using a pretrained model.
[2021-03-21 06:12:59] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 06:12:59] Creating trainer with model on device: cuda
[2021-03-21 06:12:59] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 06:13:59] Training accuracy: 0.9988
[2021-03-21 06:13:59] Testing on 10000 data points...
[2021-03-21 06:14:01] Test score for 1000 training labels: 0.7280
[2021-03-21 06:14:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:14:01] Found 50000 unlabelled features.
[2021-03-21 06:14:12] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:14:15] Searching for coresets with 20 beams...
[2021-03-21 06:14:43] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 06:14:43] Creating trainer with model on device: cuda
[2021-03-21 06:14:43] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 06:15:05] Training accuracy: 0.9977
[2021-03-21 06:15:05] Testing on 10000 data points...
[2021-03-21 06:15:07] Test score for 1400 training labels: 0.7578
[2021-03-21 06:15:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:15:07] Found 50000 unlabelled features.
[2021-03-21 06:15:18] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:15:22] Searching for coresets with 20 beams...
[2021-03-21 06:15:50] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 06:15:50] Creating trainer with model on device: cuda
[2021-03-21 06:15:50] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 06:16:18] Training accuracy: 0.9986
[2021-03-21 06:16:18] Testing on 10000 data points...
[2021-03-21 06:16:20] Test score for 1800 training labels: 0.7689
[2021-03-21 06:16:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:16:20] Found 50000 unlabelled features.
[2021-03-21 06:16:31] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:16:37] Searching for coresets with 20 beams...
[2021-03-21 06:17:05] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 06:17:05] Creating trainer with model on device: cuda
[2021-03-21 06:17:05] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 06:17:39] Training accuracy: 0.9991
[2021-03-21 06:17:39] Testing on 10000 data points...
[2021-03-21 06:17:41] Test score for 2200 training labels: 0.7848
[2021-03-21 06:17:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:17:41] Found 50000 unlabelled features.
[2021-03-21 06:17:52] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:17:59] Searching for coresets with 20 beams...
[2021-03-21 06:18:27] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 06:18:27] Creating trainer with model on device: cuda
[2021-03-21 06:18:27] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 06:19:06] Training accuracy: 0.9993
[2021-03-21 06:19:06] Testing on 10000 data points...
[2021-03-21 06:19:09] Test score for 2600 training labels: 0.7935
[2021-03-21 06:19:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:19:09] Found 50000 unlabelled features.
[2021-03-21 06:19:20] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:19:27] Searching for coresets with 20 beams...
[2021-03-21 06:19:55] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 06:19:55] Creating trainer with model on device: cuda
[2021-03-21 06:19:55] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 06:20:41] Training accuracy: 0.9989
[2021-03-21 06:20:41] Testing on 10000 data points...
[2021-03-21 06:20:43] Test score for 3000 training labels: 0.8018
[2021-03-21 06:20:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:20:43] Found 50000 unlabelled features.
[2021-03-21 06:20:54] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:21:04] Searching for coresets with 20 beams...
[2021-03-21 06:21:32] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 06:21:32] Creating trainer with model on device: cuda
[2021-03-21 06:21:32] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 06:22:23] Training accuracy: 1.0000
[2021-03-21 06:22:23] Testing on 10000 data points...
[2021-03-21 06:22:25] Test score for 3400 training labels: 0.8103
[2021-03-21 06:22:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:22:26] Found 50000 unlabelled features.
[2021-03-21 06:22:37] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:22:47] Searching for coresets with 20 beams...
[2021-03-21 06:23:15] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 06:23:15] Creating trainer with model on device: cuda
[2021-03-21 06:23:15] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 06:24:15] Training accuracy: 1.0000
[2021-03-21 06:24:15] Testing on 10000 data points...
[2021-03-21 06:24:17] Test score for 3800 training labels: 0.8140
[2021-03-21 06:24:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:24:17] Found 50000 unlabelled features.
[2021-03-21 06:24:28] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:24:40] Searching for coresets with 20 beams...
[2021-03-21 06:25:08] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 06:25:08] Creating trainer with model on device: cuda
[2021-03-21 06:25:08] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 06:26:12] Training accuracy: 0.9996
[2021-03-21 06:26:12] Testing on 10000 data points...
[2021-03-21 06:26:14] Test score for 4200 training labels: 0.8238
[2021-03-21 06:26:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:26:14] Found 50000 unlabelled features.
[2021-03-21 06:26:25] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:26:37] Searching for coresets with 20 beams...
[2021-03-21 06:27:06] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 06:27:06] Creating trainer with model on device: cuda
[2021-03-21 06:27:06] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 06:28:15] Training accuracy: 0.9999
[2021-03-21 06:28:15] Testing on 10000 data points...
[2021-03-21 06:28:17] Test score for 4600 training labels: 0.8247
[2021-03-21 06:28:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:28:17] Found 50000 unlabelled features.
[2021-03-21 06:28:28] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:28:42] Searching for coresets with 20 beams...
[2021-03-21 06:29:10] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 06:29:10] Creating trainer with model on device: cuda
[2021-03-21 06:29:10] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 06:30:26] Training accuracy: 1.0000
[2021-03-21 06:30:26] Testing on 10000 data points...
[2021-03-21 06:30:28] Test score for 5000 training labels: 0.8378
[2021-03-21 06:30:28] Experiment repeat 4/5
[2021-03-21 06:30:28] Seeded: 8
[2021-03-21 06:30:28] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 06:30:28] Randomly labelled 1000/50000
[2021-03-21 06:30:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 06:30:28] Seeded: 9
[2021-03-21 06:30:28] Running: experiment 0
[2021-03-21 06:30:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 06:30:28] Creating pretrained=True VGG16...
[2021-03-21 06:30:30] No parameter reset since we are using a pretrained model.
[2021-03-21 06:30:30] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 06:30:30] Creating trainer with model on device: cuda
[2021-03-21 06:30:30] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 06:31:29] Training accuracy: 0.9998
[2021-03-21 06:31:29] Testing on 10000 data points...
[2021-03-21 06:31:31] Test score for 1000 training labels: 0.7378
[2021-03-21 06:31:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:31:31] Found 50000 unlabelled features.
[2021-03-21 06:31:37] Showing labelled: True (1389/50000 visible, 11 redundant)
[2021-03-21 06:31:37] Creating trainer with model on device: cuda
[2021-03-21 06:31:37] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 06:31:58] Training accuracy: 0.9993
[2021-03-21 06:31:58] Testing on 10000 data points...
[2021-03-21 06:32:00] Test score for 1400 training labels: 0.7492
[2021-03-21 06:32:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:32:00] Found 50000 unlabelled features.
[2021-03-21 06:32:05] Showing labelled: True (1780/50000 visible, 20 redundant)
[2021-03-21 06:32:05] Creating trainer with model on device: cuda
[2021-03-21 06:32:05] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 06:32:32] Training accuracy: 0.9999
[2021-03-21 06:32:32] Testing on 10000 data points...
[2021-03-21 06:32:35] Test score for 1800 training labels: 0.7610
[2021-03-21 06:32:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:32:35] Found 50000 unlabelled features.
[2021-03-21 06:32:41] Showing labelled: True (2166/50000 visible, 34 redundant)
[2021-03-21 06:32:41] Creating trainer with model on device: cuda
[2021-03-21 06:32:41] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 06:33:15] Training accuracy: 0.9999
[2021-03-21 06:33:15] Testing on 10000 data points...
[2021-03-21 06:33:17] Test score for 2200 training labels: 0.7716
[2021-03-21 06:33:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:33:17] Found 50000 unlabelled features.
[2021-03-21 06:33:22] Showing labelled: True (2553/50000 visible, 47 redundant)
[2021-03-21 06:33:22] Creating trainer with model on device: cuda
[2021-03-21 06:33:22] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 06:34:01] Training accuracy: 0.9992
[2021-03-21 06:34:01] Testing on 10000 data points...
[2021-03-21 06:34:03] Test score for 2600 training labels: 0.7795
[2021-03-21 06:34:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:34:03] Found 50000 unlabelled features.
[2021-03-21 06:34:09] Showing labelled: True (2936/50000 visible, 64 redundant)
[2021-03-21 06:34:09] Creating trainer with model on device: cuda
[2021-03-21 06:34:09] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 06:34:54] Training accuracy: 1.0000
[2021-03-21 06:34:54] Testing on 10000 data points...
[2021-03-21 06:34:56] Test score for 3000 training labels: 0.7834
[2021-03-21 06:34:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:34:56] Found 50000 unlabelled features.
[2021-03-21 06:35:02] Showing labelled: True (3316/50000 visible, 84 redundant)
[2021-03-21 06:35:02] Creating trainer with model on device: cuda
[2021-03-21 06:35:02] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 06:35:53] Training accuracy: 0.9987
[2021-03-21 06:35:53] Testing on 10000 data points...
[2021-03-21 06:35:56] Test score for 3400 training labels: 0.7905
[2021-03-21 06:35:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:35:56] Found 50000 unlabelled features.
[2021-03-21 06:36:02] Showing labelled: True (3695/50000 visible, 105 redundant)
[2021-03-21 06:36:02] Creating trainer with model on device: cuda
[2021-03-21 06:36:02] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 06:37:00] Training accuracy: 1.0000
[2021-03-21 06:37:00] Testing on 10000 data points...
[2021-03-21 06:37:02] Test score for 3800 training labels: 0.7891
[2021-03-21 06:37:02] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:37:02] Found 50000 unlabelled features.
[2021-03-21 06:37:08] Showing labelled: True (4064/50000 visible, 136 redundant)
[2021-03-21 06:37:08] Creating trainer with model on device: cuda
[2021-03-21 06:37:08] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 06:38:12] Training accuracy: 0.9997
[2021-03-21 06:38:12] Testing on 10000 data points...
[2021-03-21 06:38:14] Test score for 4200 training labels: 0.7954
[2021-03-21 06:38:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:38:14] Found 50000 unlabelled features.
[2021-03-21 06:38:19] Showing labelled: True (4436/50000 visible, 164 redundant)
[2021-03-21 06:38:19] Creating trainer with model on device: cuda
[2021-03-21 06:38:19] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 06:39:29] Training accuracy: 0.9999
[2021-03-21 06:39:29] Testing on 10000 data points...
[2021-03-21 06:39:31] Test score for 4600 training labels: 0.8014
[2021-03-21 06:39:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:39:31] Found 50000 unlabelled features.
[2021-03-21 06:39:36] Showing labelled: True (4804/50000 visible, 196 redundant)
[2021-03-21 06:39:36] Creating trainer with model on device: cuda
[2021-03-21 06:39:36] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 06:40:53] Training accuracy: 1.0000
[2021-03-21 06:40:53] Testing on 10000 data points...
[2021-03-21 06:40:55] Test score for 5000 training labels: 0.7989
[2021-03-21 06:40:55] Seeded: 9
[2021-03-21 06:40:55] Running: experiment 1
[2021-03-21 06:40:55] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 06:40:55] Creating pretrained=True VGG16...
[2021-03-21 06:40:56] No parameter reset since we are using a pretrained model.
[2021-03-21 06:40:56] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 06:40:56] Creating trainer with model on device: cuda
[2021-03-21 06:40:56] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 06:41:57] Training accuracy: 1.0000
[2021-03-21 06:41:57] Testing on 10000 data points...
[2021-03-21 06:41:59] Test score for 1000 training labels: 0.7343
[2021-03-21 06:41:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:41:59] Found 50000 unlabelled features.
[2021-03-21 06:42:10] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:42:12] Searching for coresets greedily...
[2021-03-21 06:42:13] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 06:42:13] Creating trainer with model on device: cuda
[2021-03-21 06:42:13] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 06:42:34] Training accuracy: 0.9994
[2021-03-21 06:42:34] Testing on 10000 data points...
[2021-03-21 06:42:36] Test score for 1400 training labels: 0.7493
[2021-03-21 06:42:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:42:36] Found 50000 unlabelled features.
[2021-03-21 06:42:47] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:42:51] Searching for coresets greedily...
[2021-03-21 06:42:52] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 06:42:52] Creating trainer with model on device: cuda
[2021-03-21 06:42:52] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 06:43:19] Training accuracy: 0.9993
[2021-03-21 06:43:19] Testing on 10000 data points...
[2021-03-21 06:43:21] Test score for 1800 training labels: 0.7674
[2021-03-21 06:43:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:43:21] Found 50000 unlabelled features.
[2021-03-21 06:43:32] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:43:38] Searching for coresets greedily...
[2021-03-21 06:43:39] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 06:43:39] Creating trainer with model on device: cuda
[2021-03-21 06:43:39] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 06:44:13] Training accuracy: 0.9993
[2021-03-21 06:44:13] Testing on 10000 data points...
[2021-03-21 06:44:15] Test score for 2200 training labels: 0.7779
[2021-03-21 06:44:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:44:15] Found 50000 unlabelled features.
[2021-03-21 06:44:26] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:44:32] Searching for coresets greedily...
[2021-03-21 06:44:33] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 06:44:33] Creating trainer with model on device: cuda
[2021-03-21 06:44:33] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 06:45:12] Training accuracy: 0.9989
[2021-03-21 06:45:12] Testing on 10000 data points...
[2021-03-21 06:45:14] Test score for 2600 training labels: 0.7911
[2021-03-21 06:45:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:45:14] Found 50000 unlabelled features.
[2021-03-21 06:45:25] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:45:33] Searching for coresets greedily...
[2021-03-21 06:45:34] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 06:45:34] Creating trainer with model on device: cuda
[2021-03-21 06:45:34] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 06:46:19] Training accuracy: 0.9998
[2021-03-21 06:46:19] Testing on 10000 data points...
[2021-03-21 06:46:21] Test score for 3000 training labels: 0.8000
[2021-03-21 06:46:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:46:21] Found 50000 unlabelled features.
[2021-03-21 06:46:32] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:46:41] Searching for coresets greedily...
[2021-03-21 06:46:42] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 06:46:42] Creating trainer with model on device: cuda
[2021-03-21 06:46:42] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 06:47:33] Training accuracy: 0.9997
[2021-03-21 06:47:33] Testing on 10000 data points...
[2021-03-21 06:47:36] Test score for 3400 training labels: 0.8034
[2021-03-21 06:47:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:47:36] Found 50000 unlabelled features.
[2021-03-21 06:47:47] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:47:57] Searching for coresets greedily...
[2021-03-21 06:47:58] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 06:47:58] Creating trainer with model on device: cuda
[2021-03-21 06:47:58] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 06:48:57] Training accuracy: 1.0000
[2021-03-21 06:48:57] Testing on 10000 data points...
[2021-03-21 06:48:58] Test score for 3800 training labels: 0.8028
[2021-03-21 06:48:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:48:58] Found 50000 unlabelled features.
[2021-03-21 06:49:10] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:49:20] Searching for coresets greedily...
[2021-03-21 06:49:21] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 06:49:21] Creating trainer with model on device: cuda
[2021-03-21 06:49:21] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 06:50:25] Training accuracy: 1.0000
[2021-03-21 06:50:25] Testing on 10000 data points...
[2021-03-21 06:50:27] Test score for 4200 training labels: 0.8180
[2021-03-21 06:50:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:50:27] Found 50000 unlabelled features.
[2021-03-21 06:50:39] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:50:51] Searching for coresets greedily...
[2021-03-21 06:50:52] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 06:50:52] Creating trainer with model on device: cuda
[2021-03-21 06:50:52] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 06:52:01] Training accuracy: 0.9997
[2021-03-21 06:52:01] Testing on 10000 data points...
[2021-03-21 06:52:03] Test score for 4600 training labels: 0.8249
[2021-03-21 06:52:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:52:03] Found 50000 unlabelled features.
[2021-03-21 06:52:15] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 06:52:29] Searching for coresets greedily...
[2021-03-21 06:52:30] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 06:52:30] Creating trainer with model on device: cuda
[2021-03-21 06:52:30] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 06:53:44] Training accuracy: 1.0000
[2021-03-21 06:53:44] Testing on 10000 data points...
[2021-03-21 06:53:47] Test score for 5000 training labels: 0.8314
[2021-03-21 06:53:47] Seeded: 9
[2021-03-21 06:53:47] Running: experiment 2
[2021-03-21 06:53:47] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 06:53:47] Creating pretrained=True VGG16...
[2021-03-21 06:53:48] No parameter reset since we are using a pretrained model.
[2021-03-21 06:53:48] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 06:53:48] Creating trainer with model on device: cuda
[2021-03-21 06:53:48] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 06:54:48] Training accuracy: 0.9997
[2021-03-21 06:54:48] Testing on 10000 data points...
[2021-03-21 06:54:50] Test score for 1000 training labels: 0.7402
[2021-03-21 06:54:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:54:50] Found 50000 unlabelled features.
[2021-03-21 06:55:01] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 06:55:01] Creating trainer with model on device: cuda
[2021-03-21 06:55:01] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 06:55:23] Training accuracy: 0.9995
[2021-03-21 06:55:23] Testing on 10000 data points...
[2021-03-21 06:55:25] Test score for 1400 training labels: 0.7597
[2021-03-21 06:55:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:55:25] Found 50000 unlabelled features.
[2021-03-21 06:55:35] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 06:55:35] Creating trainer with model on device: cuda
[2021-03-21 06:55:35] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 06:56:02] Training accuracy: 0.7654
[2021-03-21 06:56:02] Testing on 10000 data points...
[2021-03-21 06:56:04] Test score for 1800 training labels: 0.6774
[2021-03-21 06:56:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:56:04] Found 50000 unlabelled features.
[2021-03-21 06:56:15] Showing labelled: True (2172/50000 visible, 28 redundant)
[2021-03-21 06:56:15] Creating trainer with model on device: cuda
[2021-03-21 06:56:15] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 06:56:49] Training accuracy: 0.9995
[2021-03-21 06:56:49] Testing on 10000 data points...
[2021-03-21 06:56:51] Test score for 2200 training labels: 0.7767
[2021-03-21 06:56:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:56:51] Found 50000 unlabelled features.
[2021-03-21 06:57:02] Showing labelled: True (2571/50000 visible, 29 redundant)
[2021-03-21 06:57:02] Creating trainer with model on device: cuda
[2021-03-21 06:57:02] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 06:57:41] Training accuracy: 1.0000
[2021-03-21 06:57:41] Testing on 10000 data points...
[2021-03-21 06:57:43] Test score for 2600 training labels: 0.7901
[2021-03-21 06:57:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:57:43] Found 50000 unlabelled features.
[2021-03-21 06:57:54] Showing labelled: True (2971/50000 visible, 29 redundant)
[2021-03-21 06:57:54] Creating trainer with model on device: cuda
[2021-03-21 06:57:54] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 06:58:39] Training accuracy: 0.9994
[2021-03-21 06:58:39] Testing on 10000 data points...
[2021-03-21 06:58:41] Test score for 3000 training labels: 0.7951
[2021-03-21 06:58:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:58:41] Found 50000 unlabelled features.
[2021-03-21 06:58:51] Showing labelled: True (3370/50000 visible, 30 redundant)
[2021-03-21 06:58:51] Creating trainer with model on device: cuda
[2021-03-21 06:58:51] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 06:59:42] Training accuracy: 0.9999
[2021-03-21 06:59:42] Testing on 10000 data points...
[2021-03-21 06:59:44] Test score for 3400 training labels: 0.8044
[2021-03-21 06:59:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 06:59:44] Found 50000 unlabelled features.
[2021-03-21 06:59:55] Showing labelled: True (3769/50000 visible, 31 redundant)
[2021-03-21 06:59:55] Creating trainer with model on device: cuda
[2021-03-21 06:59:55] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 07:00:53] Training accuracy: 0.9998
[2021-03-21 07:00:53] Testing on 10000 data points...
[2021-03-21 07:00:55] Test score for 3800 training labels: 0.8154
[2021-03-21 07:00:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:00:55] Found 50000 unlabelled features.
[2021-03-21 07:01:04] Showing labelled: True (4168/50000 visible, 32 redundant)
[2021-03-21 07:01:04] Creating trainer with model on device: cuda
[2021-03-21 07:01:04] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 07:02:08] Training accuracy: 1.0000
[2021-03-21 07:02:08] Testing on 10000 data points...
[2021-03-21 07:02:10] Test score for 4200 training labels: 0.8230
[2021-03-21 07:02:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:02:10] Found 50000 unlabelled features.
[2021-03-21 07:02:20] Showing labelled: True (4568/50000 visible, 32 redundant)
[2021-03-21 07:02:20] Creating trainer with model on device: cuda
[2021-03-21 07:02:20] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 07:03:29] Training accuracy: 0.9999
[2021-03-21 07:03:29] Testing on 10000 data points...
[2021-03-21 07:03:31] Test score for 4600 training labels: 0.8286
[2021-03-21 07:03:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:03:31] Found 50000 unlabelled features.
[2021-03-21 07:03:42] Showing labelled: True (4967/50000 visible, 33 redundant)
[2021-03-21 07:03:42] Creating trainer with model on device: cuda
[2021-03-21 07:03:42] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 07:04:57] Training accuracy: 1.0000
[2021-03-21 07:04:57] Testing on 10000 data points...
[2021-03-21 07:04:59] Test score for 5000 training labels: 0.8326
[2021-03-21 07:04:59] Seeded: 9
[2021-03-21 07:04:59] Running: experiment 3
[2021-03-21 07:04:59] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 07:04:59] Creating pretrained=True VGG16...
[2021-03-21 07:05:01] No parameter reset since we are using a pretrained model.
[2021-03-21 07:05:01] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 07:05:01] Creating trainer with model on device: cuda
[2021-03-21 07:05:01] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 07:06:01] Training accuracy: 1.0000
[2021-03-21 07:06:01] Testing on 10000 data points...
[2021-03-21 07:06:03] Test score for 1000 training labels: 0.7343
[2021-03-21 07:06:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:06:03] Found 50000 unlabelled features.
[2021-03-21 07:06:14] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:06:17] Searching for coresets with 1 beams...
[2021-03-21 07:06:18] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 07:06:18] Creating trainer with model on device: cuda
[2021-03-21 07:06:18] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 07:06:39] Training accuracy: 0.9926
[2021-03-21 07:06:39] Testing on 10000 data points...
[2021-03-21 07:06:41] Test score for 1400 training labels: 0.7415
[2021-03-21 07:06:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:06:41] Found 50000 unlabelled features.
[2021-03-21 07:06:51] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:06:56] Searching for coresets with 1 beams...
[2021-03-21 07:06:57] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 07:06:57] Creating trainer with model on device: cuda
[2021-03-21 07:06:57] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 07:07:24] Training accuracy: 0.9989
[2021-03-21 07:07:24] Testing on 10000 data points...
[2021-03-21 07:07:26] Test score for 1800 training labels: 0.7685
[2021-03-21 07:07:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:07:26] Found 50000 unlabelled features.
[2021-03-21 07:07:37] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:07:43] Searching for coresets with 1 beams...
[2021-03-21 07:07:44] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 07:07:44] Creating trainer with model on device: cuda
[2021-03-21 07:07:44] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 07:08:18] Training accuracy: 0.9998
[2021-03-21 07:08:18] Testing on 10000 data points...
[2021-03-21 07:08:21] Test score for 2200 training labels: 0.7809
[2021-03-21 07:08:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:08:21] Found 50000 unlabelled features.
[2021-03-21 07:08:32] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:08:38] Searching for coresets with 1 beams...
[2021-03-21 07:08:39] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 07:08:39] Creating trainer with model on device: cuda
[2021-03-21 07:08:39] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 07:09:19] Training accuracy: 0.9994
[2021-03-21 07:09:19] Testing on 10000 data points...
[2021-03-21 07:09:21] Test score for 2600 training labels: 0.7942
[2021-03-21 07:09:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:09:21] Found 50000 unlabelled features.
[2021-03-21 07:09:32] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:09:40] Searching for coresets with 1 beams...
[2021-03-21 07:09:41] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 07:09:41] Creating trainer with model on device: cuda
[2021-03-21 07:09:41] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 07:10:27] Training accuracy: 0.9997
[2021-03-21 07:10:27] Testing on 10000 data points...
[2021-03-21 07:10:28] Test score for 3000 training labels: 0.8038
[2021-03-21 07:10:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:10:28] Found 50000 unlabelled features.
[2021-03-21 07:10:40] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:10:49] Searching for coresets with 1 beams...
[2021-03-21 07:10:50] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 07:10:50] Creating trainer with model on device: cuda
[2021-03-21 07:10:50] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 07:11:42] Training accuracy: 1.0000
[2021-03-21 07:11:42] Testing on 10000 data points...
[2021-03-21 07:11:44] Test score for 3400 training labels: 0.8031
[2021-03-21 07:11:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:11:44] Found 50000 unlabelled features.
[2021-03-21 07:11:55] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:12:06] Searching for coresets with 1 beams...
[2021-03-21 07:12:07] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 07:12:07] Creating trainer with model on device: cuda
[2021-03-21 07:12:07] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 07:13:05] Training accuracy: 1.0000
[2021-03-21 07:13:05] Testing on 10000 data points...
[2021-03-21 07:13:07] Test score for 3800 training labels: 0.8100
[2021-03-21 07:13:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:13:07] Found 50000 unlabelled features.
[2021-03-21 07:13:18] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:13:30] Searching for coresets with 1 beams...
[2021-03-21 07:13:31] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 07:13:31] Creating trainer with model on device: cuda
[2021-03-21 07:13:31] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 07:14:34] Training accuracy: 0.9995
[2021-03-21 07:14:34] Testing on 10000 data points...
[2021-03-21 07:14:36] Test score for 4200 training labels: 0.8173
[2021-03-21 07:14:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:14:36] Found 50000 unlabelled features.
[2021-03-21 07:14:48] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:15:01] Searching for coresets with 1 beams...
[2021-03-21 07:15:02] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 07:15:02] Creating trainer with model on device: cuda
[2021-03-21 07:15:02] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 07:16:12] Training accuracy: 0.9999
[2021-03-21 07:16:12] Testing on 10000 data points...
[2021-03-21 07:16:14] Test score for 4600 training labels: 0.8235
[2021-03-21 07:16:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:16:14] Found 50000 unlabelled features.
[2021-03-21 07:16:26] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:16:40] Searching for coresets with 1 beams...
[2021-03-21 07:16:41] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 07:16:41] Creating trainer with model on device: cuda
[2021-03-21 07:16:41] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 07:17:56] Training accuracy: 1.0000
[2021-03-21 07:17:56] Testing on 10000 data points...
[2021-03-21 07:17:59] Test score for 5000 training labels: 0.8249
[2021-03-21 07:17:59] Seeded: 9
[2021-03-21 07:17:59] Running: experiment 4
[2021-03-21 07:17:59] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 07:17:59] Creating pretrained=True VGG16...
[2021-03-21 07:18:00] No parameter reset since we are using a pretrained model.
[2021-03-21 07:18:00] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 07:18:00] Creating trainer with model on device: cuda
[2021-03-21 07:18:00] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 07:19:01] Training accuracy: 1.0000
[2021-03-21 07:19:01] Testing on 10000 data points...
[2021-03-21 07:19:03] Test score for 1000 training labels: 0.7383
[2021-03-21 07:19:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:19:03] Found 50000 unlabelled features.
[2021-03-21 07:19:14] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:19:17] Searching for coresets with 20 beams...
[2021-03-21 07:19:46] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 07:19:46] Creating trainer with model on device: cuda
[2021-03-21 07:19:46] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 07:20:07] Training accuracy: 0.9982
[2021-03-21 07:20:07] Testing on 10000 data points...
[2021-03-21 07:20:09] Test score for 1400 training labels: 0.7507
[2021-03-21 07:20:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:20:09] Found 50000 unlabelled features.
[2021-03-21 07:20:20] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:20:25] Searching for coresets with 20 beams...
[2021-03-21 07:20:53] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 07:20:53] Creating trainer with model on device: cuda
[2021-03-21 07:20:53] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 07:21:21] Training accuracy: 0.9985
[2021-03-21 07:21:21] Testing on 10000 data points...
[2021-03-21 07:21:23] Test score for 1800 training labels: 0.7682
[2021-03-21 07:21:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:21:23] Found 50000 unlabelled features.
[2021-03-21 07:21:34] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:21:39] Searching for coresets with 20 beams...
[2021-03-21 07:22:08] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 07:22:08] Creating trainer with model on device: cuda
[2021-03-21 07:22:08] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 07:22:42] Training accuracy: 0.9993
[2021-03-21 07:22:42] Testing on 10000 data points...
[2021-03-21 07:22:44] Test score for 2200 training labels: 0.7847
[2021-03-21 07:22:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:22:44] Found 50000 unlabelled features.
[2021-03-21 07:22:55] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:23:01] Searching for coresets with 20 beams...
[2021-03-21 07:23:30] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 07:23:30] Creating trainer with model on device: cuda
[2021-03-21 07:23:30] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 07:24:09] Training accuracy: 0.9994
[2021-03-21 07:24:09] Testing on 10000 data points...
[2021-03-21 07:24:11] Test score for 2600 training labels: 0.7958
[2021-03-21 07:24:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:24:11] Found 50000 unlabelled features.
[2021-03-21 07:24:22] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:24:30] Searching for coresets with 20 beams...
[2021-03-21 07:24:59] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 07:24:59] Creating trainer with model on device: cuda
[2021-03-21 07:24:59] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 07:25:45] Training accuracy: 0.9999
[2021-03-21 07:25:45] Testing on 10000 data points...
[2021-03-21 07:25:47] Test score for 3000 training labels: 0.8007
[2021-03-21 07:25:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:25:47] Found 50000 unlabelled features.
[2021-03-21 07:25:58] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:26:07] Searching for coresets with 20 beams...
[2021-03-21 07:26:36] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 07:26:36] Creating trainer with model on device: cuda
[2021-03-21 07:26:36] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 07:27:29] Training accuracy: 1.0000
[2021-03-21 07:27:29] Testing on 10000 data points...
[2021-03-21 07:27:31] Test score for 3400 training labels: 0.8043
[2021-03-21 07:27:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:27:31] Found 50000 unlabelled features.
[2021-03-21 07:27:42] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:27:53] Searching for coresets with 20 beams...
[2021-03-21 07:28:21] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 07:28:21] Creating trainer with model on device: cuda
[2021-03-21 07:28:21] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 07:29:19] Training accuracy: 0.9995
[2021-03-21 07:29:19] Testing on 10000 data points...
[2021-03-21 07:29:21] Test score for 3800 training labels: 0.8145
[2021-03-21 07:29:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:29:21] Found 50000 unlabelled features.
[2021-03-21 07:29:33] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:29:44] Searching for coresets with 20 beams...
[2021-03-21 07:30:13] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 07:30:13] Creating trainer with model on device: cuda
[2021-03-21 07:30:13] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 07:31:17] Training accuracy: 0.9998
[2021-03-21 07:31:17] Testing on 10000 data points...
[2021-03-21 07:31:19] Test score for 4200 training labels: 0.8144
[2021-03-21 07:31:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:31:19] Found 50000 unlabelled features.
[2021-03-21 07:31:30] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:31:43] Searching for coresets with 20 beams...
[2021-03-21 07:32:12] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 07:32:12] Creating trainer with model on device: cuda
[2021-03-21 07:32:12] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 07:33:21] Training accuracy: 0.9998
[2021-03-21 07:33:21] Testing on 10000 data points...
[2021-03-21 07:33:23] Test score for 4600 training labels: 0.8229
[2021-03-21 07:33:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:33:23] Found 50000 unlabelled features.
[2021-03-21 07:33:35] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:33:49] Searching for coresets with 20 beams...
[2021-03-21 07:34:18] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 07:34:18] Creating trainer with model on device: cuda
[2021-03-21 07:34:18] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 07:35:34] Training accuracy: 1.0000
[2021-03-21 07:35:34] Testing on 10000 data points...
[2021-03-21 07:35:37] Test score for 5000 training labels: 0.8302
[2021-03-21 07:35:37] Seeded: 9
[2021-03-21 07:35:37] Running: experiment 5
[2021-03-21 07:35:37] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 07:35:37] Creating pretrained=True VGG16...
[2021-03-21 07:35:38] No parameter reset since we are using a pretrained model.
[2021-03-21 07:35:38] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 07:35:38] Creating trainer with model on device: cuda
[2021-03-21 07:35:38] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 07:36:38] Training accuracy: 0.9999
[2021-03-21 07:36:38] Testing on 10000 data points...
[2021-03-21 07:36:40] Test score for 1000 training labels: 0.7368
[2021-03-21 07:36:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:36:40] Found 50000 unlabelled features.
[2021-03-21 07:36:51] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:36:54] Searching for coresets with 1 beams...
[2021-03-21 07:36:56] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 07:36:56] Creating trainer with model on device: cuda
[2021-03-21 07:36:56] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 07:37:16] Training accuracy: 0.9986
[2021-03-21 07:37:16] Testing on 10000 data points...
[2021-03-21 07:37:18] Test score for 1400 training labels: 0.7592
[2021-03-21 07:37:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:37:18] Found 50000 unlabelled features.
[2021-03-21 07:37:29] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:37:33] Searching for coresets with 1 beams...
[2021-03-21 07:37:34] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 07:37:34] Creating trainer with model on device: cuda
[2021-03-21 07:37:34] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 07:38:01] Training accuracy: 0.9995
[2021-03-21 07:38:01] Testing on 10000 data points...
[2021-03-21 07:38:03] Test score for 1800 training labels: 0.7737
[2021-03-21 07:38:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:38:03] Found 50000 unlabelled features.
[2021-03-21 07:38:14] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:38:20] Searching for coresets with 1 beams...
[2021-03-21 07:38:21] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 07:38:21] Creating trainer with model on device: cuda
[2021-03-21 07:38:21] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 07:38:56] Training accuracy: 0.9999
[2021-03-21 07:38:56] Testing on 10000 data points...
[2021-03-21 07:38:58] Test score for 2200 training labels: 0.7892
[2021-03-21 07:38:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:38:58] Found 50000 unlabelled features.
[2021-03-21 07:39:09] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:39:15] Searching for coresets with 1 beams...
[2021-03-21 07:39:16] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 07:39:16] Creating trainer with model on device: cuda
[2021-03-21 07:39:16] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 07:39:56] Training accuracy: 0.9999
[2021-03-21 07:39:56] Testing on 10000 data points...
[2021-03-21 07:39:58] Test score for 2600 training labels: 0.7977
[2021-03-21 07:39:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:39:58] Found 50000 unlabelled features.
[2021-03-21 07:40:10] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:40:17] Searching for coresets with 1 beams...
[2021-03-21 07:40:19] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 07:40:19] Creating trainer with model on device: cuda
[2021-03-21 07:40:19] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 07:41:03] Training accuracy: 1.0000
[2021-03-21 07:41:03] Testing on 10000 data points...
[2021-03-21 07:41:05] Test score for 3000 training labels: 0.8092
[2021-03-21 07:41:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:41:05] Found 50000 unlabelled features.
[2021-03-21 07:41:16] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:41:24] Searching for coresets with 1 beams...
[2021-03-21 07:41:26] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 07:41:26] Creating trainer with model on device: cuda
[2021-03-21 07:41:26] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 07:42:17] Training accuracy: 0.9997
[2021-03-21 07:42:17] Testing on 10000 data points...
[2021-03-21 07:42:19] Test score for 3400 training labels: 0.8155
[2021-03-21 07:42:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:42:19] Found 50000 unlabelled features.
[2021-03-21 07:42:30] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:42:41] Searching for coresets with 1 beams...
[2021-03-21 07:42:42] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 07:42:42] Creating trainer with model on device: cuda
[2021-03-21 07:42:42] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 07:43:41] Training accuracy: 0.9996
[2021-03-21 07:43:41] Testing on 10000 data points...
[2021-03-21 07:43:43] Test score for 3800 training labels: 0.8209
[2021-03-21 07:43:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:43:43] Found 50000 unlabelled features.
[2021-03-21 07:43:53] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:44:04] Searching for coresets with 1 beams...
[2021-03-21 07:44:05] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 07:44:05] Creating trainer with model on device: cuda
[2021-03-21 07:44:05] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 07:45:10] Training accuracy: 1.0000
[2021-03-21 07:45:10] Testing on 10000 data points...
[2021-03-21 07:45:12] Test score for 4200 training labels: 0.8284
[2021-03-21 07:45:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:45:12] Found 50000 unlabelled features.
[2021-03-21 07:45:22] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:45:34] Searching for coresets with 1 beams...
[2021-03-21 07:45:35] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 07:45:35] Creating trainer with model on device: cuda
[2021-03-21 07:45:35] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 07:46:44] Training accuracy: 0.9261
[2021-03-21 07:46:44] Testing on 10000 data points...
[2021-03-21 07:46:46] Test score for 4600 training labels: 0.7929
[2021-03-21 07:46:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:46:46] Found 50000 unlabelled features.
[2021-03-21 07:46:58] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:47:12] Searching for coresets with 1 beams...
[2021-03-21 07:47:13] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 07:47:13] Creating trainer with model on device: cuda
[2021-03-21 07:47:13] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 07:48:29] Training accuracy: 1.0000
[2021-03-21 07:48:29] Testing on 10000 data points...
[2021-03-21 07:48:31] Test score for 5000 training labels: 0.8395
[2021-03-21 07:48:31] Seeded: 9
[2021-03-21 07:48:31] Running: experiment 6
[2021-03-21 07:48:31] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 07:48:31] Creating pretrained=True VGG16...
[2021-03-21 07:48:32] No parameter reset since we are using a pretrained model.
[2021-03-21 07:48:32] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 07:48:32] Creating trainer with model on device: cuda
[2021-03-21 07:48:32] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 07:49:33] Training accuracy: 0.9631
[2021-03-21 07:49:33] Testing on 10000 data points...
[2021-03-21 07:49:35] Test score for 1000 training labels: 0.6615
[2021-03-21 07:49:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:49:35] Found 50000 unlabelled features.
[2021-03-21 07:49:46] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:49:49] Searching for coresets with 20 beams...
[2021-03-21 07:50:17] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 07:50:17] Creating trainer with model on device: cuda
[2021-03-21 07:50:17] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 07:50:39] Training accuracy: 0.9978
[2021-03-21 07:50:39] Testing on 10000 data points...
[2021-03-21 07:50:41] Test score for 1400 training labels: 0.7159
[2021-03-21 07:50:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:50:41] Found 50000 unlabelled features.
[2021-03-21 07:50:52] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:50:56] Searching for coresets with 20 beams...
[2021-03-21 07:51:25] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 07:51:25] Creating trainer with model on device: cuda
[2021-03-21 07:51:25] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 07:51:52] Training accuracy: 0.9986
[2021-03-21 07:51:52] Testing on 10000 data points...
[2021-03-21 07:51:54] Test score for 1800 training labels: 0.7432
[2021-03-21 07:51:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:51:54] Found 50000 unlabelled features.
[2021-03-21 07:52:05] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:52:10] Searching for coresets with 20 beams...
[2021-03-21 07:52:39] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 07:52:39] Creating trainer with model on device: cuda
[2021-03-21 07:52:39] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 07:53:13] Training accuracy: 0.9959
[2021-03-21 07:53:13] Testing on 10000 data points...
[2021-03-21 07:53:15] Test score for 2200 training labels: 0.7592
[2021-03-21 07:53:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:53:15] Found 50000 unlabelled features.
[2021-03-21 07:53:26] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:53:32] Searching for coresets with 20 beams...
[2021-03-21 07:54:01] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 07:54:01] Creating trainer with model on device: cuda
[2021-03-21 07:54:01] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 07:54:40] Training accuracy: 0.9984
[2021-03-21 07:54:40] Testing on 10000 data points...
[2021-03-21 07:54:42] Test score for 2600 training labels: 0.7711
[2021-03-21 07:54:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:54:42] Found 50000 unlabelled features.
[2021-03-21 07:54:53] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:55:01] Searching for coresets with 20 beams...
[2021-03-21 07:55:30] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 07:55:30] Creating trainer with model on device: cuda
[2021-03-21 07:55:30] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 07:56:15] Training accuracy: 0.9997
[2021-03-21 07:56:15] Testing on 10000 data points...
[2021-03-21 07:56:17] Test score for 3000 training labels: 0.7829
[2021-03-21 07:56:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:56:17] Found 50000 unlabelled features.
[2021-03-21 07:56:28] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:56:37] Searching for coresets with 20 beams...
[2021-03-21 07:57:06] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 07:57:06] Creating trainer with model on device: cuda
[2021-03-21 07:57:06] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 07:57:58] Training accuracy: 0.9999
[2021-03-21 07:57:58] Testing on 10000 data points...
[2021-03-21 07:58:00] Test score for 3400 training labels: 0.7878
[2021-03-21 07:58:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:58:00] Found 50000 unlabelled features.
[2021-03-21 07:58:11] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 07:58:22] Searching for coresets with 20 beams...
[2021-03-21 07:58:50] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 07:58:51] Creating trainer with model on device: cuda
[2021-03-21 07:58:51] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 07:59:48] Training accuracy: 0.9993
[2021-03-21 07:59:48] Testing on 10000 data points...
[2021-03-21 07:59:50] Test score for 3800 training labels: 0.8007
[2021-03-21 07:59:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 07:59:50] Found 50000 unlabelled features.
[2021-03-21 08:00:02] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:00:13] Searching for coresets with 20 beams...
[2021-03-21 08:00:42] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 08:00:42] Creating trainer with model on device: cuda
[2021-03-21 08:00:42] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 08:01:45] Training accuracy: 0.9987
[2021-03-21 08:01:45] Testing on 10000 data points...
[2021-03-21 08:01:47] Test score for 4200 training labels: 0.7990
[2021-03-21 08:01:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:01:47] Found 50000 unlabelled features.
[2021-03-21 08:01:59] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:02:11] Searching for coresets with 20 beams...
[2021-03-21 08:02:39] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 08:02:39] Creating trainer with model on device: cuda
[2021-03-21 08:02:39] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 08:03:49] Training accuracy: 0.9991
[2021-03-21 08:03:49] Testing on 10000 data points...
[2021-03-21 08:03:51] Test score for 4600 training labels: 0.8126
[2021-03-21 08:03:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:03:51] Found 50000 unlabelled features.
[2021-03-21 08:04:03] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:04:17] Searching for coresets with 20 beams...
[2021-03-21 08:04:46] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 08:04:46] Creating trainer with model on device: cuda
[2021-03-21 08:04:46] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 08:06:02] Training accuracy: 1.0000
[2021-03-21 08:06:02] Testing on 10000 data points...
[2021-03-21 08:06:04] Test score for 5000 training labels: 0.8206
[2021-03-21 08:06:04] Seeded: 9
[2021-03-21 08:06:04] Running: experiment 7
[2021-03-21 08:06:04] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 08:06:04] Creating pretrained=True VGG16...
[2021-03-21 08:06:06] No parameter reset since we are using a pretrained model.
[2021-03-21 08:06:06] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 08:06:06] Creating trainer with model on device: cuda
[2021-03-21 08:06:06] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 08:07:06] Training accuracy: 0.9997
[2021-03-21 08:07:06] Testing on 10000 data points...
[2021-03-21 08:07:08] Test score for 1000 training labels: 0.7341
[2021-03-21 08:07:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:07:08] Found 50000 unlabelled features.
[2021-03-21 08:07:19] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:07:22] Searching for coresets with 20 beams...
[2021-03-21 08:07:50] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 08:07:50] Creating trainer with model on device: cuda
[2021-03-21 08:07:50] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 08:08:11] Training accuracy: 0.9693
[2021-03-21 08:08:11] Testing on 10000 data points...
[2021-03-21 08:08:13] Test score for 1400 training labels: 0.7407
[2021-03-21 08:08:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:08:13] Found 50000 unlabelled features.
[2021-03-21 08:08:24] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:08:28] Searching for coresets with 20 beams...
[2021-03-21 08:08:56] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 08:08:56] Creating trainer with model on device: cuda
[2021-03-21 08:08:56] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 08:09:24] Training accuracy: 0.9981
[2021-03-21 08:09:24] Testing on 10000 data points...
[2021-03-21 08:09:26] Test score for 1800 training labels: 0.7774
[2021-03-21 08:09:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:09:26] Found 50000 unlabelled features.
[2021-03-21 08:09:37] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:09:43] Searching for coresets with 20 beams...
[2021-03-21 08:10:11] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 08:10:11] Creating trainer with model on device: cuda
[2021-03-21 08:10:11] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 08:10:45] Training accuracy: 0.9998
[2021-03-21 08:10:45] Testing on 10000 data points...
[2021-03-21 08:10:47] Test score for 2200 training labels: 0.7904
[2021-03-21 08:10:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:10:47] Found 50000 unlabelled features.
[2021-03-21 08:10:58] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:11:04] Searching for coresets with 20 beams...
[2021-03-21 08:11:33] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 08:11:33] Creating trainer with model on device: cuda
[2021-03-21 08:11:33] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 08:12:12] Training accuracy: 0.9582
[2021-03-21 08:12:12] Testing on 10000 data points...
[2021-03-21 08:12:14] Test score for 2600 training labels: 0.7671
[2021-03-21 08:12:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:12:14] Found 50000 unlabelled features.
[2021-03-21 08:12:25] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:12:33] Searching for coresets with 20 beams...
[2021-03-21 08:13:02] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 08:13:02] Creating trainer with model on device: cuda
[2021-03-21 08:13:02] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 08:13:48] Training accuracy: 0.9999
[2021-03-21 08:13:48] Testing on 10000 data points...
[2021-03-21 08:13:50] Test score for 3000 training labels: 0.8048
[2021-03-21 08:13:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:13:50] Found 50000 unlabelled features.
[2021-03-21 08:14:01] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:14:10] Searching for coresets with 20 beams...
[2021-03-21 08:14:38] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 08:14:38] Creating trainer with model on device: cuda
[2021-03-21 08:14:38] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 08:15:30] Training accuracy: 0.9998
[2021-03-21 08:15:30] Testing on 10000 data points...
[2021-03-21 08:15:32] Test score for 3400 training labels: 0.8137
[2021-03-21 08:15:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:15:32] Found 50000 unlabelled features.
[2021-03-21 08:15:44] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:15:54] Searching for coresets with 20 beams...
[2021-03-21 08:16:22] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 08:16:22] Creating trainer with model on device: cuda
[2021-03-21 08:16:22] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 08:17:20] Training accuracy: 1.0000
[2021-03-21 08:17:20] Testing on 10000 data points...
[2021-03-21 08:17:22] Test score for 3800 training labels: 0.8171
[2021-03-21 08:17:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:17:22] Found 50000 unlabelled features.
[2021-03-21 08:17:33] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:17:45] Searching for coresets with 20 beams...
[2021-03-21 08:18:13] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 08:18:13] Creating trainer with model on device: cuda
[2021-03-21 08:18:13] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 08:19:16] Training accuracy: 0.9994
[2021-03-21 08:19:16] Testing on 10000 data points...
[2021-03-21 08:19:18] Test score for 4200 training labels: 0.8235
[2021-03-21 08:19:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:19:18] Found 50000 unlabelled features.
[2021-03-21 08:19:30] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:19:43] Searching for coresets with 20 beams...
[2021-03-21 08:20:11] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 08:20:11] Creating trainer with model on device: cuda
[2021-03-21 08:20:11] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 08:21:20] Training accuracy: 0.9994
[2021-03-21 08:21:20] Testing on 10000 data points...
[2021-03-21 08:21:22] Test score for 4600 training labels: 0.8291
[2021-03-21 08:21:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:21:22] Found 50000 unlabelled features.
[2021-03-21 08:21:33] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:21:46] Searching for coresets with 20 beams...
[2021-03-21 08:22:15] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 08:22:15] Creating trainer with model on device: cuda
[2021-03-21 08:22:15] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 08:23:30] Training accuracy: 1.0000
[2021-03-21 08:23:30] Testing on 10000 data points...
[2021-03-21 08:23:32] Test score for 5000 training labels: 0.8339
[2021-03-21 08:23:32] Experiment repeat 5/5
[2021-03-21 08:23:32] Seeded: 9
[2021-03-21 08:23:32] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 08:23:32] Randomly labelled 1000/50000
[2021-03-21 08:23:32] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 08:23:32] Seeded: 10
[2021-03-21 08:23:32] Running: experiment 0
[2021-03-21 08:23:32] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 08:23:32] Creating pretrained=True VGG16...
[2021-03-21 08:23:34] No parameter reset since we are using a pretrained model.
[2021-03-21 08:23:34] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 08:23:34] Creating trainer with model on device: cuda
[2021-03-21 08:23:34] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 08:24:34] Training accuracy: 1.0000
[2021-03-21 08:24:34] Testing on 10000 data points...
[2021-03-21 08:24:36] Test score for 1000 training labels: 0.7301
[2021-03-21 08:24:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:24:36] Found 50000 unlabelled features.
[2021-03-21 08:24:41] Showing labelled: True (1392/50000 visible, 8 redundant)
[2021-03-21 08:24:41] Creating trainer with model on device: cuda
[2021-03-21 08:24:41] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 08:25:02] Training accuracy: 0.9994
[2021-03-21 08:25:02] Testing on 10000 data points...
[2021-03-21 08:25:05] Test score for 1400 training labels: 0.7522
[2021-03-21 08:25:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:25:05] Found 50000 unlabelled features.
[2021-03-21 08:25:11] Showing labelled: True (1784/50000 visible, 16 redundant)
[2021-03-21 08:25:11] Creating trainer with model on device: cuda
[2021-03-21 08:25:11] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 08:25:38] Training accuracy: 0.9974
[2021-03-21 08:25:38] Testing on 10000 data points...
[2021-03-21 08:25:40] Test score for 1800 training labels: 0.7557
[2021-03-21 08:25:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:25:40] Found 50000 unlabelled features.
[2021-03-21 08:25:45] Showing labelled: True (2171/50000 visible, 29 redundant)
[2021-03-21 08:25:45] Creating trainer with model on device: cuda
[2021-03-21 08:25:45] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 08:26:19] Training accuracy: 1.0000
[2021-03-21 08:26:19] Testing on 10000 data points...
[2021-03-21 08:26:21] Test score for 2200 training labels: 0.7682
[2021-03-21 08:26:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:26:21] Found 50000 unlabelled features.
[2021-03-21 08:26:26] Showing labelled: True (2554/50000 visible, 46 redundant)
[2021-03-21 08:26:26] Creating trainer with model on device: cuda
[2021-03-21 08:26:26] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 08:27:05] Training accuracy: 0.9998
[2021-03-21 08:27:05] Testing on 10000 data points...
[2021-03-21 08:27:07] Test score for 2600 training labels: 0.7752
[2021-03-21 08:27:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:27:07] Found 50000 unlabelled features.
[2021-03-21 08:27:13] Showing labelled: True (2937/50000 visible, 63 redundant)
[2021-03-21 08:27:13] Creating trainer with model on device: cuda
[2021-03-21 08:27:13] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 08:27:59] Training accuracy: 0.9993
[2021-03-21 08:27:59] Testing on 10000 data points...
[2021-03-21 08:28:01] Test score for 3000 training labels: 0.7823
[2021-03-21 08:28:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:28:01] Found 50000 unlabelled features.
[2021-03-21 08:28:07] Showing labelled: True (3319/50000 visible, 81 redundant)
[2021-03-21 08:28:07] Creating trainer with model on device: cuda
[2021-03-21 08:28:07] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 08:28:58] Training accuracy: 1.0000
[2021-03-21 08:28:58] Testing on 10000 data points...
[2021-03-21 08:29:00] Test score for 3400 training labels: 0.7870
[2021-03-21 08:29:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:29:00] Found 50000 unlabelled features.
[2021-03-21 08:29:05] Showing labelled: True (3686/50000 visible, 114 redundant)
[2021-03-21 08:29:05] Creating trainer with model on device: cuda
[2021-03-21 08:29:05] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 08:30:03] Training accuracy: 1.0000
[2021-03-21 08:30:03] Testing on 10000 data points...
[2021-03-21 08:30:05] Test score for 3800 training labels: 0.7943
[2021-03-21 08:30:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:30:05] Found 50000 unlabelled features.
[2021-03-21 08:30:10] Showing labelled: True (4066/50000 visible, 134 redundant)
[2021-03-21 08:30:10] Creating trainer with model on device: cuda
[2021-03-21 08:30:10] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 08:31:13] Training accuracy: 1.0000
[2021-03-21 08:31:13] Testing on 10000 data points...
[2021-03-21 08:31:15] Test score for 4200 training labels: 0.7949
[2021-03-21 08:31:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:31:15] Found 50000 unlabelled features.
[2021-03-21 08:31:22] Showing labelled: True (4427/50000 visible, 173 redundant)
[2021-03-21 08:31:22] Creating trainer with model on device: cuda
[2021-03-21 08:31:22] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 08:32:32] Training accuracy: 0.9999
[2021-03-21 08:32:32] Testing on 10000 data points...
[2021-03-21 08:32:34] Test score for 4600 training labels: 0.7975
[2021-03-21 08:32:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:32:34] Found 50000 unlabelled features.
[2021-03-21 08:32:40] Showing labelled: True (4788/50000 visible, 212 redundant)
[2021-03-21 08:32:40] Creating trainer with model on device: cuda
[2021-03-21 08:32:40] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 08:33:56] Training accuracy: 1.0000
[2021-03-21 08:33:56] Testing on 10000 data points...
[2021-03-21 08:33:58] Test score for 5000 training labels: 0.7993
[2021-03-21 08:33:58] Seeded: 10
[2021-03-21 08:33:58] Running: experiment 1
[2021-03-21 08:33:58] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 08:33:58] Creating pretrained=True VGG16...
[2021-03-21 08:34:00] No parameter reset since we are using a pretrained model.
[2021-03-21 08:34:00] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 08:34:00] Creating trainer with model on device: cuda
[2021-03-21 08:34:00] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 08:35:00] Training accuracy: 1.0000
[2021-03-21 08:35:00] Testing on 10000 data points...
[2021-03-21 08:35:02] Test score for 1000 training labels: 0.7284
[2021-03-21 08:35:02] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:35:02] Found 50000 unlabelled features.
[2021-03-21 08:35:13] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:35:16] Searching for coresets greedily...
[2021-03-21 08:35:16] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 08:35:16] Creating trainer with model on device: cuda
[2021-03-21 08:35:16] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 08:35:38] Training accuracy: 0.9968
[2021-03-21 08:35:38] Testing on 10000 data points...
[2021-03-21 08:35:40] Test score for 1400 training labels: 0.7278
[2021-03-21 08:35:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:35:40] Found 50000 unlabelled features.
[2021-03-21 08:35:51] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:35:55] Searching for coresets greedily...
[2021-03-21 08:35:56] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 08:35:56] Creating trainer with model on device: cuda
[2021-03-21 08:35:56] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 08:36:23] Training accuracy: 0.9866
[2021-03-21 08:36:23] Testing on 10000 data points...
[2021-03-21 08:36:25] Test score for 1800 training labels: 0.7474
[2021-03-21 08:36:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:36:25] Found 50000 unlabelled features.
[2021-03-21 08:36:36] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:36:41] Searching for coresets greedily...
[2021-03-21 08:36:42] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 08:36:42] Creating trainer with model on device: cuda
[2021-03-21 08:36:42] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 08:37:15] Training accuracy: 0.9992
[2021-03-21 08:37:15] Testing on 10000 data points...
[2021-03-21 08:37:17] Test score for 2200 training labels: 0.7716
[2021-03-21 08:37:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:37:17] Found 50000 unlabelled features.
[2021-03-21 08:37:27] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:37:34] Searching for coresets greedily...
[2021-03-21 08:37:35] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 08:37:35] Creating trainer with model on device: cuda
[2021-03-21 08:37:35] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 08:38:15] Training accuracy: 0.9997
[2021-03-21 08:38:15] Testing on 10000 data points...
[2021-03-21 08:38:17] Test score for 2600 training labels: 0.7845
[2021-03-21 08:38:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:38:17] Found 50000 unlabelled features.
[2021-03-21 08:38:28] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:38:36] Searching for coresets greedily...
[2021-03-21 08:38:36] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 08:38:36] Creating trainer with model on device: cuda
[2021-03-21 08:38:36] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 08:39:22] Training accuracy: 0.9991
[2021-03-21 08:39:22] Testing on 10000 data points...
[2021-03-21 08:39:24] Test score for 3000 training labels: 0.7962
[2021-03-21 08:39:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:39:24] Found 50000 unlabelled features.
[2021-03-21 08:39:35] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:39:44] Searching for coresets greedily...
[2021-03-21 08:39:45] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 08:39:45] Creating trainer with model on device: cuda
[2021-03-21 08:39:45] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 08:40:37] Training accuracy: 0.9986
[2021-03-21 08:40:37] Testing on 10000 data points...
[2021-03-21 08:40:39] Test score for 3400 training labels: 0.7986
[2021-03-21 08:40:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:40:39] Found 50000 unlabelled features.
[2021-03-21 08:40:50] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:40:59] Searching for coresets greedily...
[2021-03-21 08:41:00] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 08:41:00] Creating trainer with model on device: cuda
[2021-03-21 08:41:00] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 08:41:59] Training accuracy: 0.9997
[2021-03-21 08:41:59] Testing on 10000 data points...
[2021-03-21 08:42:01] Test score for 3800 training labels: 0.8066
[2021-03-21 08:42:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:42:01] Found 50000 unlabelled features.
[2021-03-21 08:42:11] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:42:22] Searching for coresets greedily...
[2021-03-21 08:42:23] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 08:42:23] Creating trainer with model on device: cuda
[2021-03-21 08:42:23] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 08:43:27] Training accuracy: 1.0000
[2021-03-21 08:43:27] Testing on 10000 data points...
[2021-03-21 08:43:29] Test score for 4200 training labels: 0.8144
[2021-03-21 08:43:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:43:29] Found 50000 unlabelled features.
[2021-03-21 08:43:41] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:43:53] Searching for coresets greedily...
[2021-03-21 08:43:53] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 08:43:53] Creating trainer with model on device: cuda
[2021-03-21 08:43:53] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 08:45:03] Training accuracy: 1.0000
[2021-03-21 08:45:03] Testing on 10000 data points...
[2021-03-21 08:45:05] Test score for 4600 training labels: 0.8232
[2021-03-21 08:45:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:45:05] Found 50000 unlabelled features.
[2021-03-21 08:45:17] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:45:30] Searching for coresets greedily...
[2021-03-21 08:45:31] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 08:45:31] Creating trainer with model on device: cuda
[2021-03-21 08:45:31] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 08:46:46] Training accuracy: 1.0000
[2021-03-21 08:46:46] Testing on 10000 data points...
[2021-03-21 08:46:48] Test score for 5000 training labels: 0.8255
[2021-03-21 08:46:48] Seeded: 10
[2021-03-21 08:46:48] Running: experiment 2
[2021-03-21 08:46:48] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 08:46:48] Creating pretrained=True VGG16...
[2021-03-21 08:46:50] No parameter reset since we are using a pretrained model.
[2021-03-21 08:46:50] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 08:46:50] Creating trainer with model on device: cuda
[2021-03-21 08:46:50] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 08:47:51] Training accuracy: 0.9993
[2021-03-21 08:47:51] Testing on 10000 data points...
[2021-03-21 08:47:53] Test score for 1000 training labels: 0.7332
[2021-03-21 08:47:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:47:53] Found 50000 unlabelled features.
[2021-03-21 08:48:03] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 08:48:03] Creating trainer with model on device: cuda
[2021-03-21 08:48:03] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 08:48:24] Training accuracy: 0.9991
[2021-03-21 08:48:24] Testing on 10000 data points...
[2021-03-21 08:48:27] Test score for 1400 training labels: 0.7595
[2021-03-21 08:48:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:48:27] Found 50000 unlabelled features.
[2021-03-21 08:48:36] Showing labelled: True (1799/50000 visible, 1 redundant)
[2021-03-21 08:48:36] Creating trainer with model on device: cuda
[2021-03-21 08:48:36] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 08:49:03] Training accuracy: 0.9978
[2021-03-21 08:49:03] Testing on 10000 data points...
[2021-03-21 08:49:05] Test score for 1800 training labels: 0.7694
[2021-03-21 08:49:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:49:05] Found 50000 unlabelled features.
[2021-03-21 08:49:15] Showing labelled: True (2197/50000 visible, 3 redundant)
[2021-03-21 08:49:15] Creating trainer with model on device: cuda
[2021-03-21 08:49:15] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 08:49:48] Training accuracy: 0.9977
[2021-03-21 08:49:48] Testing on 10000 data points...
[2021-03-21 08:49:50] Test score for 2200 training labels: 0.7829
[2021-03-21 08:49:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:49:50] Found 50000 unlabelled features.
[2021-03-21 08:50:01] Showing labelled: True (2597/50000 visible, 3 redundant)
[2021-03-21 08:50:01] Creating trainer with model on device: cuda
[2021-03-21 08:50:01] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 08:50:41] Training accuracy: 1.0000
[2021-03-21 08:50:41] Testing on 10000 data points...
[2021-03-21 08:50:43] Test score for 2600 training labels: 0.7980
[2021-03-21 08:50:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:50:43] Found 50000 unlabelled features.
[2021-03-21 08:50:53] Showing labelled: True (2997/50000 visible, 3 redundant)
[2021-03-21 08:50:53] Creating trainer with model on device: cuda
[2021-03-21 08:50:53] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 08:51:38] Training accuracy: 0.9992
[2021-03-21 08:51:38] Testing on 10000 data points...
[2021-03-21 08:51:40] Test score for 3000 training labels: 0.8060
[2021-03-21 08:51:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:51:40] Found 50000 unlabelled features.
[2021-03-21 08:51:51] Showing labelled: True (3396/50000 visible, 4 redundant)
[2021-03-21 08:51:51] Creating trainer with model on device: cuda
[2021-03-21 08:51:51] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 08:52:42] Training accuracy: 0.9999
[2021-03-21 08:52:42] Testing on 10000 data points...
[2021-03-21 08:52:44] Test score for 3400 training labels: 0.8106
[2021-03-21 08:52:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:52:44] Found 50000 unlabelled features.
[2021-03-21 08:52:54] Showing labelled: True (3796/50000 visible, 4 redundant)
[2021-03-21 08:52:54] Creating trainer with model on device: cuda
[2021-03-21 08:52:54] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 08:53:52] Training accuracy: 1.0000
[2021-03-21 08:53:52] Testing on 10000 data points...
[2021-03-21 08:53:54] Test score for 3800 training labels: 0.8198
[2021-03-21 08:53:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:53:54] Found 50000 unlabelled features.
[2021-03-21 08:54:05] Showing labelled: True (4196/50000 visible, 4 redundant)
[2021-03-21 08:54:05] Creating trainer with model on device: cuda
[2021-03-21 08:54:05] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 08:55:08] Training accuracy: 0.9996
[2021-03-21 08:55:08] Testing on 10000 data points...
[2021-03-21 08:55:09] Test score for 4200 training labels: 0.8275
[2021-03-21 08:55:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:55:10] Found 50000 unlabelled features.
[2021-03-21 08:55:20] Showing labelled: True (4595/50000 visible, 5 redundant)
[2021-03-21 08:55:20] Creating trainer with model on device: cuda
[2021-03-21 08:55:20] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 08:56:30] Training accuracy: 0.9992
[2021-03-21 08:56:30] Testing on 10000 data points...
[2021-03-21 08:56:32] Test score for 4600 training labels: 0.8330
[2021-03-21 08:56:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:56:32] Found 50000 unlabelled features.
[2021-03-21 08:56:42] Showing labelled: True (4993/50000 visible, 7 redundant)
[2021-03-21 08:56:42] Creating trainer with model on device: cuda
[2021-03-21 08:56:42] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 08:57:57] Training accuracy: 1.0000
[2021-03-21 08:57:57] Testing on 10000 data points...
[2021-03-21 08:57:59] Test score for 5000 training labels: 0.8370
[2021-03-21 08:57:59] Seeded: 10
[2021-03-21 08:57:59] Running: experiment 3
[2021-03-21 08:57:59] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 08:57:59] Creating pretrained=True VGG16...
[2021-03-21 08:58:01] No parameter reset since we are using a pretrained model.
[2021-03-21 08:58:01] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 08:58:01] Creating trainer with model on device: cuda
[2021-03-21 08:58:01] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 08:59:02] Training accuracy: 1.0000
[2021-03-21 08:59:02] Testing on 10000 data points...
[2021-03-21 08:59:04] Test score for 1000 training labels: 0.7314
[2021-03-21 08:59:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:59:04] Found 50000 unlabelled features.
[2021-03-21 08:59:14] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:59:16] Searching for coresets with 1 beams...
[2021-03-21 08:59:18] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 08:59:18] Creating trainer with model on device: cuda
[2021-03-21 08:59:18] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 08:59:39] Training accuracy: 0.9982
[2021-03-21 08:59:39] Testing on 10000 data points...
[2021-03-21 08:59:41] Test score for 1400 training labels: 0.7427
[2021-03-21 08:59:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 08:59:41] Found 50000 unlabelled features.
[2021-03-21 08:59:52] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 08:59:56] Searching for coresets with 1 beams...
[2021-03-21 08:59:57] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 08:59:57] Creating trainer with model on device: cuda
[2021-03-21 08:59:57] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 09:00:25] Training accuracy: 0.9981
[2021-03-21 09:00:25] Testing on 10000 data points...
[2021-03-21 09:00:27] Test score for 1800 training labels: 0.7568
[2021-03-21 09:00:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:00:27] Found 50000 unlabelled features.
[2021-03-21 09:00:38] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:00:44] Searching for coresets with 1 beams...
[2021-03-21 09:00:45] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 09:00:45] Creating trainer with model on device: cuda
[2021-03-21 09:00:45] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 09:01:19] Training accuracy: 0.9988
[2021-03-21 09:01:19] Testing on 10000 data points...
[2021-03-21 09:01:21] Test score for 2200 training labels: 0.7726
[2021-03-21 09:01:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:01:21] Found 50000 unlabelled features.
[2021-03-21 09:01:32] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:01:39] Searching for coresets with 1 beams...
[2021-03-21 09:01:40] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 09:01:40] Creating trainer with model on device: cuda
[2021-03-21 09:01:40] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 09:02:20] Training accuracy: 0.9996
[2021-03-21 09:02:20] Testing on 10000 data points...
[2021-03-21 09:02:22] Test score for 2600 training labels: 0.7813
[2021-03-21 09:02:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:02:22] Found 50000 unlabelled features.
[2021-03-21 09:02:33] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:02:41] Searching for coresets with 1 beams...
[2021-03-21 09:02:42] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 09:02:42] Creating trainer with model on device: cuda
[2021-03-21 09:02:42] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 09:03:27] Training accuracy: 1.0000
[2021-03-21 09:03:27] Testing on 10000 data points...
[2021-03-21 09:03:29] Test score for 3000 training labels: 0.7893
[2021-03-21 09:03:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:03:29] Found 50000 unlabelled features.
[2021-03-21 09:03:40] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:03:48] Searching for coresets with 1 beams...
[2021-03-21 09:03:50] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 09:03:50] Creating trainer with model on device: cuda
[2021-03-21 09:03:50] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 09:04:40] Training accuracy: 1.0000
[2021-03-21 09:04:40] Testing on 10000 data points...
[2021-03-21 09:04:43] Test score for 3400 training labels: 0.8063
[2021-03-21 09:04:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:04:43] Found 50000 unlabelled features.
[2021-03-21 09:04:53] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:05:03] Searching for coresets with 1 beams...
[2021-03-21 09:05:04] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 09:05:04] Creating trainer with model on device: cuda
[2021-03-21 09:05:04] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 09:06:02] Training accuracy: 1.0000
[2021-03-21 09:06:02] Testing on 10000 data points...
[2021-03-21 09:06:04] Test score for 3800 training labels: 0.8061
[2021-03-21 09:06:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:06:04] Found 50000 unlabelled features.
[2021-03-21 09:06:14] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:06:26] Searching for coresets with 1 beams...
[2021-03-21 09:06:27] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 09:06:27] Creating trainer with model on device: cuda
[2021-03-21 09:06:27] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 09:07:32] Training accuracy: 1.0000
[2021-03-21 09:07:32] Testing on 10000 data points...
[2021-03-21 09:07:34] Test score for 4200 training labels: 0.8174
[2021-03-21 09:07:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:07:34] Found 50000 unlabelled features.
[2021-03-21 09:07:45] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:07:58] Searching for coresets with 1 beams...
[2021-03-21 09:07:59] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 09:07:59] Creating trainer with model on device: cuda
[2021-03-21 09:07:59] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 09:09:09] Training accuracy: 0.9994
[2021-03-21 09:09:09] Testing on 10000 data points...
[2021-03-21 09:09:11] Test score for 4600 training labels: 0.8274
[2021-03-21 09:09:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:09:11] Found 50000 unlabelled features.
[2021-03-21 09:09:23] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:09:36] Searching for coresets with 1 beams...
[2021-03-21 09:09:37] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 09:09:37] Creating trainer with model on device: cuda
[2021-03-21 09:09:37] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 09:10:52] Training accuracy: 0.9994
[2021-03-21 09:10:52] Testing on 10000 data points...
[2021-03-21 09:10:54] Test score for 5000 training labels: 0.8316
[2021-03-21 09:10:54] Seeded: 10
[2021-03-21 09:10:54] Running: experiment 4
[2021-03-21 09:10:54] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 09:10:54] Creating pretrained=True VGG16...
[2021-03-21 09:10:56] No parameter reset since we are using a pretrained model.
[2021-03-21 09:10:56] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 09:10:56] Creating trainer with model on device: cuda
[2021-03-21 09:10:56] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 09:11:56] Training accuracy: 0.9995
[2021-03-21 09:11:56] Testing on 10000 data points...
[2021-03-21 09:11:58] Test score for 1000 training labels: 0.7370
[2021-03-21 09:11:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:11:58] Found 50000 unlabelled features.
[2021-03-21 09:12:09] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:12:12] Searching for coresets with 20 beams...
[2021-03-21 09:12:41] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 09:12:41] Creating trainer with model on device: cuda
[2021-03-21 09:12:41] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 09:13:02] Training accuracy: 0.9981
[2021-03-21 09:13:02] Testing on 10000 data points...
[2021-03-21 09:13:04] Test score for 1400 training labels: 0.7482
[2021-03-21 09:13:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:13:04] Found 50000 unlabelled features.
[2021-03-21 09:13:14] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:13:18] Searching for coresets with 20 beams...
[2021-03-21 09:13:47] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 09:13:47] Creating trainer with model on device: cuda
[2021-03-21 09:13:47] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 09:14:15] Training accuracy: 0.9995
[2021-03-21 09:14:15] Testing on 10000 data points...
[2021-03-21 09:14:17] Test score for 1800 training labels: 0.7671
[2021-03-21 09:14:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:14:17] Found 50000 unlabelled features.
[2021-03-21 09:14:27] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:14:33] Searching for coresets with 20 beams...
[2021-03-21 09:15:02] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 09:15:02] Creating trainer with model on device: cuda
[2021-03-21 09:15:02] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 09:15:35] Training accuracy: 1.0000
[2021-03-21 09:15:35] Testing on 10000 data points...
[2021-03-21 09:15:38] Test score for 2200 training labels: 0.7759
[2021-03-21 09:15:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:15:38] Found 50000 unlabelled features.
[2021-03-21 09:15:49] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:15:55] Searching for coresets with 20 beams...
[2021-03-21 09:16:24] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 09:16:24] Creating trainer with model on device: cuda
[2021-03-21 09:16:24] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 09:17:03] Training accuracy: 0.9990
[2021-03-21 09:17:03] Testing on 10000 data points...
[2021-03-21 09:17:05] Test score for 2600 training labels: 0.7873
[2021-03-21 09:17:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:17:05] Found 50000 unlabelled features.
[2021-03-21 09:17:16] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:17:24] Searching for coresets with 20 beams...
[2021-03-21 09:17:53] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 09:17:53] Creating trainer with model on device: cuda
[2021-03-21 09:17:53] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 09:18:39] Training accuracy: 0.9998
[2021-03-21 09:18:39] Testing on 10000 data points...
[2021-03-21 09:18:40] Test score for 3000 training labels: 0.7973
[2021-03-21 09:18:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:18:41] Found 50000 unlabelled features.
[2021-03-21 09:18:52] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:19:01] Searching for coresets with 20 beams...
[2021-03-21 09:19:30] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 09:19:30] Creating trainer with model on device: cuda
[2021-03-21 09:19:30] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 09:20:22] Training accuracy: 1.0000
[2021-03-21 09:20:22] Testing on 10000 data points...
[2021-03-21 09:20:23] Test score for 3400 training labels: 0.8040
[2021-03-21 09:20:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:20:23] Found 50000 unlabelled features.
[2021-03-21 09:20:34] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:20:44] Searching for coresets with 20 beams...
[2021-03-21 09:21:13] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 09:21:13] Creating trainer with model on device: cuda
[2021-03-21 09:21:13] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 09:22:12] Training accuracy: 0.9997
[2021-03-21 09:22:12] Testing on 10000 data points...
[2021-03-21 09:22:14] Test score for 3800 training labels: 0.8115
[2021-03-21 09:22:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:22:14] Found 50000 unlabelled features.
[2021-03-21 09:22:24] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:22:35] Searching for coresets with 20 beams...
[2021-03-21 09:23:04] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 09:23:04] Creating trainer with model on device: cuda
[2021-03-21 09:23:04] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 09:24:08] Training accuracy: 1.0000
[2021-03-21 09:24:08] Testing on 10000 data points...
[2021-03-21 09:24:10] Test score for 4200 training labels: 0.8153
[2021-03-21 09:24:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:24:10] Found 50000 unlabelled features.
[2021-03-21 09:24:22] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:24:34] Searching for coresets with 20 beams...
[2021-03-21 09:25:03] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 09:25:03] Creating trainer with model on device: cuda
[2021-03-21 09:25:03] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 09:26:13] Training accuracy: 1.0000
[2021-03-21 09:26:13] Testing on 10000 data points...
[2021-03-21 09:26:15] Test score for 4600 training labels: 0.8206
[2021-03-21 09:26:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:26:15] Found 50000 unlabelled features.
[2021-03-21 09:26:26] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:26:39] Searching for coresets with 20 beams...
[2021-03-21 09:27:08] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 09:27:08] Creating trainer with model on device: cuda
[2021-03-21 09:27:08] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 09:28:23] Training accuracy: 1.0000
[2021-03-21 09:28:23] Testing on 10000 data points...
[2021-03-21 09:28:25] Test score for 5000 training labels: 0.8299
[2021-03-21 09:28:25] Seeded: 10
[2021-03-21 09:28:25] Running: experiment 5
[2021-03-21 09:28:25] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 09:28:25] Creating pretrained=True VGG16...
[2021-03-21 09:28:27] No parameter reset since we are using a pretrained model.
[2021-03-21 09:28:27] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 09:28:27] Creating trainer with model on device: cuda
[2021-03-21 09:28:27] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 09:29:27] Training accuracy: 0.9995
[2021-03-21 09:29:27] Testing on 10000 data points...
[2021-03-21 09:29:29] Test score for 1000 training labels: 0.7293
[2021-03-21 09:29:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:29:29] Found 50000 unlabelled features.
[2021-03-21 09:29:40] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:29:43] Searching for coresets with 1 beams...
[2021-03-21 09:29:45] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 09:29:45] Creating trainer with model on device: cuda
[2021-03-21 09:29:45] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 09:30:06] Training accuracy: 0.9985
[2021-03-21 09:30:06] Testing on 10000 data points...
[2021-03-21 09:30:08] Test score for 1400 training labels: 0.7539
[2021-03-21 09:30:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:30:08] Found 50000 unlabelled features.
[2021-03-21 09:30:19] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:30:23] Searching for coresets with 1 beams...
[2021-03-21 09:30:25] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 09:30:25] Creating trainer with model on device: cuda
[2021-03-21 09:30:25] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 09:30:52] Training accuracy: 0.9984
[2021-03-21 09:30:52] Testing on 10000 data points...
[2021-03-21 09:30:54] Test score for 1800 training labels: 0.7791
[2021-03-21 09:30:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:30:54] Found 50000 unlabelled features.
[2021-03-21 09:31:05] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:31:10] Searching for coresets with 1 beams...
[2021-03-21 09:31:11] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 09:31:11] Creating trainer with model on device: cuda
[2021-03-21 09:31:11] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 09:31:45] Training accuracy: 0.9994
[2021-03-21 09:31:45] Testing on 10000 data points...
[2021-03-21 09:31:47] Test score for 2200 training labels: 0.7893
[2021-03-21 09:31:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:31:47] Found 50000 unlabelled features.
[2021-03-21 09:31:59] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:32:05] Searching for coresets with 1 beams...
[2021-03-21 09:32:06] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 09:32:06] Creating trainer with model on device: cuda
[2021-03-21 09:32:06] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 09:32:46] Training accuracy: 1.0000
[2021-03-21 09:32:46] Testing on 10000 data points...
[2021-03-21 09:32:48] Test score for 2600 training labels: 0.8019
[2021-03-21 09:32:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:32:48] Found 50000 unlabelled features.
[2021-03-21 09:32:59] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:33:07] Searching for coresets with 1 beams...
[2021-03-21 09:33:08] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 09:33:08] Creating trainer with model on device: cuda
[2021-03-21 09:33:08] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 09:33:54] Training accuracy: 0.9982
[2021-03-21 09:33:54] Testing on 10000 data points...
[2021-03-21 09:33:56] Test score for 3000 training labels: 0.7814
[2021-03-21 09:33:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:33:56] Found 50000 unlabelled features.
[2021-03-21 09:34:07] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:34:16] Searching for coresets with 1 beams...
[2021-03-21 09:34:17] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 09:34:17] Creating trainer with model on device: cuda
[2021-03-21 09:34:17] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 09:35:08] Training accuracy: 1.0000
[2021-03-21 09:35:08] Testing on 10000 data points...
[2021-03-21 09:35:10] Test score for 3400 training labels: 0.8012
[2021-03-21 09:35:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:35:10] Found 50000 unlabelled features.
[2021-03-21 09:35:22] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:35:31] Searching for coresets with 1 beams...
[2021-03-21 09:35:33] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 09:35:33] Creating trainer with model on device: cuda
[2021-03-21 09:35:33] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 09:36:30] Training accuracy: 1.0000
[2021-03-21 09:36:30] Testing on 10000 data points...
[2021-03-21 09:36:32] Test score for 3800 training labels: 0.8109
[2021-03-21 09:36:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:36:33] Found 50000 unlabelled features.
[2021-03-21 09:36:43] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:36:54] Searching for coresets with 1 beams...
[2021-03-21 09:36:56] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 09:36:56] Creating trainer with model on device: cuda
[2021-03-21 09:36:56] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 09:37:59] Training accuracy: 0.9994
[2021-03-21 09:37:59] Testing on 10000 data points...
[2021-03-21 09:38:01] Test score for 4200 training labels: 0.8180
[2021-03-21 09:38:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:38:01] Found 50000 unlabelled features.
[2021-03-21 09:38:12] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:38:24] Searching for coresets with 1 beams...
[2021-03-21 09:38:26] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 09:38:26] Creating trainer with model on device: cuda
[2021-03-21 09:38:26] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 09:39:35] Training accuracy: 0.9998
[2021-03-21 09:39:35] Testing on 10000 data points...
[2021-03-21 09:39:37] Test score for 4600 training labels: 0.8258
[2021-03-21 09:39:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:39:37] Found 50000 unlabelled features.
[2021-03-21 09:39:49] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:40:04] Searching for coresets with 1 beams...
[2021-03-21 09:40:05] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 09:40:05] Creating trainer with model on device: cuda
[2021-03-21 09:40:05] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 09:41:21] Training accuracy: 0.9998
[2021-03-21 09:41:21] Testing on 10000 data points...
[2021-03-21 09:41:23] Test score for 5000 training labels: 0.8289
[2021-03-21 09:41:23] Seeded: 10
[2021-03-21 09:41:23] Running: experiment 6
[2021-03-21 09:41:23] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 09:41:23] Creating pretrained=True VGG16...
[2021-03-21 09:41:24] No parameter reset since we are using a pretrained model.
[2021-03-21 09:41:24] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 09:41:24] Creating trainer with model on device: cuda
[2021-03-21 09:41:24] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 09:42:26] Training accuracy: 0.9991
[2021-03-21 09:42:26] Testing on 10000 data points...
[2021-03-21 09:42:28] Test score for 1000 training labels: 0.7310
[2021-03-21 09:42:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:42:28] Found 50000 unlabelled features.
[2021-03-21 09:42:38] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:42:41] Searching for coresets with 20 beams...
[2021-03-21 09:43:10] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 09:43:10] Creating trainer with model on device: cuda
[2021-03-21 09:43:10] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 09:43:31] Training accuracy: 0.9969
[2021-03-21 09:43:31] Testing on 10000 data points...
[2021-03-21 09:43:33] Test score for 1400 training labels: 0.7598
[2021-03-21 09:43:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:43:33] Found 50000 unlabelled features.
[2021-03-21 09:43:44] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:43:49] Searching for coresets with 20 beams...
[2021-03-21 09:44:17] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 09:44:17] Creating trainer with model on device: cuda
[2021-03-21 09:44:17] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 09:44:45] Training accuracy: 0.9976
[2021-03-21 09:44:45] Testing on 10000 data points...
[2021-03-21 09:44:46] Test score for 1800 training labels: 0.7756
[2021-03-21 09:44:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:44:46] Found 50000 unlabelled features.
[2021-03-21 09:44:57] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:45:03] Searching for coresets with 20 beams...
[2021-03-21 09:45:32] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 09:45:32] Creating trainer with model on device: cuda
[2021-03-21 09:45:32] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 09:46:06] Training accuracy: 0.9990
[2021-03-21 09:46:06] Testing on 10000 data points...
[2021-03-21 09:46:08] Test score for 2200 training labels: 0.7832
[2021-03-21 09:46:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:46:08] Found 50000 unlabelled features.
[2021-03-21 09:46:19] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:46:25] Searching for coresets with 20 beams...
[2021-03-21 09:46:54] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 09:46:54] Creating trainer with model on device: cuda
[2021-03-21 09:46:54] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 09:47:34] Training accuracy: 1.0000
[2021-03-21 09:47:34] Testing on 10000 data points...
[2021-03-21 09:47:36] Test score for 2600 training labels: 0.7944
[2021-03-21 09:47:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:47:36] Found 50000 unlabelled features.
[2021-03-21 09:47:47] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:47:55] Searching for coresets with 20 beams...
[2021-03-21 09:48:24] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 09:48:24] Creating trainer with model on device: cuda
[2021-03-21 09:48:24] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 09:49:09] Training accuracy: 1.0000
[2021-03-21 09:49:09] Testing on 10000 data points...
[2021-03-21 09:49:11] Test score for 3000 training labels: 0.8083
[2021-03-21 09:49:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:49:11] Found 50000 unlabelled features.
[2021-03-21 09:49:23] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:49:31] Searching for coresets with 20 beams...
[2021-03-21 09:50:00] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 09:50:00] Creating trainer with model on device: cuda
[2021-03-21 09:50:00] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 09:50:51] Training accuracy: 1.0000
[2021-03-21 09:50:51] Testing on 10000 data points...
[2021-03-21 09:50:53] Test score for 3400 training labels: 0.8129
[2021-03-21 09:50:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:50:53] Found 50000 unlabelled features.
[2021-03-21 09:51:05] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:51:15] Searching for coresets with 20 beams...
[2021-03-21 09:51:44] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 09:51:44] Creating trainer with model on device: cuda
[2021-03-21 09:51:44] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 09:52:43] Training accuracy: 1.0000
[2021-03-21 09:52:43] Testing on 10000 data points...
[2021-03-21 09:52:45] Test score for 3800 training labels: 0.8193
[2021-03-21 09:52:45] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:52:45] Found 50000 unlabelled features.
[2021-03-21 09:52:57] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:53:08] Searching for coresets with 20 beams...
[2021-03-21 09:53:37] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 09:53:37] Creating trainer with model on device: cuda
[2021-03-21 09:53:37] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 09:54:41] Training accuracy: 0.9997
[2021-03-21 09:54:41] Testing on 10000 data points...
[2021-03-21 09:54:43] Test score for 4200 training labels: 0.8238
[2021-03-21 09:54:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:54:43] Found 50000 unlabelled features.
[2021-03-21 09:54:55] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:55:08] Searching for coresets with 20 beams...
[2021-03-21 09:55:37] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 09:55:37] Creating trainer with model on device: cuda
[2021-03-21 09:55:37] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 09:56:47] Training accuracy: 1.0000
[2021-03-21 09:56:47] Testing on 10000 data points...
[2021-03-21 09:56:49] Test score for 4600 training labels: 0.8265
[2021-03-21 09:56:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 09:56:49] Found 50000 unlabelled features.
[2021-03-21 09:57:00] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 09:57:14] Searching for coresets with 20 beams...
[2021-03-21 09:57:43] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 09:57:43] Creating trainer with model on device: cuda
[2021-03-21 09:57:43] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 09:59:00] Training accuracy: 1.0000
[2021-03-21 09:59:00] Testing on 10000 data points...
[2021-03-21 09:59:02] Test score for 5000 training labels: 0.8354
[2021-03-21 09:59:02] Seeded: 10
[2021-03-21 09:59:02] Running: experiment 7
[2021-03-21 09:59:02] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 09:59:02] Creating pretrained=True VGG16...
[2021-03-21 09:59:04] No parameter reset since we are using a pretrained model.
[2021-03-21 09:59:04] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 09:59:04] Creating trainer with model on device: cuda
[2021-03-21 09:59:04] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 10:00:05] Training accuracy: 0.9944
[2021-03-21 10:00:05] Testing on 10000 data points...
[2021-03-21 10:00:07] Test score for 1000 training labels: 0.7127
[2021-03-21 10:00:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:00:07] Found 50000 unlabelled features.
[2021-03-21 10:00:17] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:00:20] Searching for coresets with 20 beams...
[2021-03-21 10:00:49] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 10:00:49] Creating trainer with model on device: cuda
[2021-03-21 10:00:49] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 10:01:10] Training accuracy: 0.9981
[2021-03-21 10:01:10] Testing on 10000 data points...
[2021-03-21 10:01:12] Test score for 1400 training labels: 0.7457
[2021-03-21 10:01:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:01:12] Found 50000 unlabelled features.
[2021-03-21 10:01:23] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:01:27] Searching for coresets with 20 beams...
[2021-03-21 10:01:56] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 10:01:56] Creating trainer with model on device: cuda
[2021-03-21 10:01:56] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 10:02:23] Training accuracy: 0.9979
[2021-03-21 10:02:23] Testing on 10000 data points...
[2021-03-21 10:02:25] Test score for 1800 training labels: 0.7618
[2021-03-21 10:02:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:02:25] Found 50000 unlabelled features.
[2021-03-21 10:02:36] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:02:41] Searching for coresets with 20 beams...
[2021-03-21 10:03:09] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 10:03:09] Creating trainer with model on device: cuda
[2021-03-21 10:03:09] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 10:03:43] Training accuracy: 0.9999
[2021-03-21 10:03:43] Testing on 10000 data points...
[2021-03-21 10:03:46] Test score for 2200 training labels: 0.7811
[2021-03-21 10:03:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:03:46] Found 50000 unlabelled features.
[2021-03-21 10:03:57] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:04:03] Searching for coresets with 20 beams...
[2021-03-21 10:04:31] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 10:04:31] Creating trainer with model on device: cuda
[2021-03-21 10:04:31] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 10:05:11] Training accuracy: 0.9998
[2021-03-21 10:05:11] Testing on 10000 data points...
[2021-03-21 10:05:13] Test score for 2600 training labels: 0.7939
[2021-03-21 10:05:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:05:13] Found 50000 unlabelled features.
[2021-03-21 10:05:24] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:05:32] Searching for coresets with 20 beams...
[2021-03-21 10:06:00] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 10:06:00] Creating trainer with model on device: cuda
[2021-03-21 10:06:00] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 10:06:45] Training accuracy: 0.9994
[2021-03-21 10:06:45] Testing on 10000 data points...
[2021-03-21 10:06:47] Test score for 3000 training labels: 0.8024
[2021-03-21 10:06:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:06:47] Found 50000 unlabelled features.
[2021-03-21 10:06:58] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:07:07] Searching for coresets with 20 beams...
[2021-03-21 10:07:35] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 10:07:35] Creating trainer with model on device: cuda
[2021-03-21 10:07:35] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 10:08:26] Training accuracy: 0.9998
[2021-03-21 10:08:26] Testing on 10000 data points...
[2021-03-21 10:08:29] Test score for 3400 training labels: 0.8074
[2021-03-21 10:08:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:08:29] Found 50000 unlabelled features.
[2021-03-21 10:08:40] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:08:50] Searching for coresets with 20 beams...
[2021-03-21 10:09:18] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 10:09:18] Creating trainer with model on device: cuda
[2021-03-21 10:09:18] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 10:10:17] Training accuracy: 0.9999
[2021-03-21 10:10:17] Testing on 10000 data points...
[2021-03-21 10:10:19] Test score for 3800 training labels: 0.8200
[2021-03-21 10:10:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:10:19] Found 50000 unlabelled features.
[2021-03-21 10:10:31] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:10:42] Searching for coresets with 20 beams...
[2021-03-21 10:11:10] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 10:11:10] Creating trainer with model on device: cuda
[2021-03-21 10:11:10] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 10:12:15] Training accuracy: 1.0000
[2021-03-21 10:12:15] Testing on 10000 data points...
[2021-03-21 10:12:17] Test score for 4200 training labels: 0.8202
[2021-03-21 10:12:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:12:17] Found 50000 unlabelled features.
[2021-03-21 10:12:27] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:12:40] Searching for coresets with 20 beams...
[2021-03-21 10:13:08] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 10:13:08] Creating trainer with model on device: cuda
[2021-03-21 10:13:08] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 10:14:18] Training accuracy: 1.0000
[2021-03-21 10:14:18] Testing on 10000 data points...
[2021-03-21 10:14:20] Test score for 4600 training labels: 0.8297
[2021-03-21 10:14:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 10:14:20] Found 50000 unlabelled features.
[2021-03-21 10:14:33] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 10:14:46] Searching for coresets with 20 beams...
[2021-03-21 10:15:14] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 10:15:14] Creating trainer with model on device: cuda
[2021-03-21 10:15:14] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 10:16:30] Training accuracy: 1.0000
[2021-03-21 10:16:30] Testing on 10000 data points...
[2021-03-21 10:16:32] Test score for 5000 training labels: 0.8299
[2021-03-21 10:16:32] Updated results: ../results/cifar10/main/results.json
