[2021-03-20 03:02:14] Created experiment 0:
[2021-03-20 03:02:14]  - Model: vgg-pretrained
[2021-03-20 03:02:14]  - Acquisition function: random
[2021-03-20 03:02:14] Created experiment 1:
[2021-03-20 03:02:14]  - Model: vgg-pretrained
[2021-03-20 03:02:14]  - Acquisition function: top2entropy-beam-coreset
[2021-03-20 03:02:14] Created experiment 2:
[2021-03-20 03:02:14]  - Model: vgg-pretrained
[2021-03-20 03:02:14]  - Acquisition function: lc-beam-coreset
[2021-03-20 03:02:14] Created experiment 3:
[2021-03-20 03:02:14]  - Model: vgg-pretrained
[2021-03-20 03:02:14]  - Acquisition function: greedy-coreset
[2021-03-20 03:02:14] Created experiment 4:
[2021-03-20 03:02:14]  - Model: vgg-pretrained
[2021-03-20 03:02:14]  - Acquisition function: least-confidence
[2021-03-20 03:02:14] Loading cifar10-augmented test set...
[2021-03-20 03:02:15] Experiment repeat 1/4
[2021-03-20 03:02:15] Seeded: 5
[2021-03-20 03:02:15] Using 2.00% labels of the dataset (1000/50000)
[2021-03-20 03:02:15] Randomly labelled 1000/50000
[2021-03-20 03:02:15] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 03:02:15] Seeded: 6
[2021-03-20 03:02:15] Running: experiment 0
[2021-03-20 03:02:15] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 03:02:15] Creating pretrained=True VGG16...
[2021-03-20 03:02:17] No parameter reset since we are using a pretrained model.
[2021-03-20 03:02:17] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 03:02:17] Creating trainer with model on device: cuda
[2021-03-20 03:02:21] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 03:03:24] Training accuracy: 0.8885
[2021-03-20 03:03:24] Testing on 10000 data points...
[2021-03-20 03:03:26] Test score for 1000 training labels: 0.7395
[2021-03-20 03:03:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:03:26] Found 50000 unlabelled features.
[2021-03-20 03:03:34] Showing labelled: True (1382/50000 visible, 18 redundant)
[2021-03-20 03:03:34] Creating trainer with model on device: cuda
[2021-03-20 03:03:34] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 03:03:56] Training accuracy: 0.8788
[2021-03-20 03:03:56] Testing on 10000 data points...
[2021-03-20 03:03:57] Test score for 1400 training labels: 0.7544
[2021-03-20 03:03:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:03:57] Found 50000 unlabelled features.
[2021-03-20 03:04:05] Showing labelled: True (1769/50000 visible, 31 redundant)
[2021-03-20 03:04:05] Creating trainer with model on device: cuda
[2021-03-20 03:04:05] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 03:04:33] Training accuracy: 0.8926
[2021-03-20 03:04:33] Testing on 10000 data points...
[2021-03-20 03:04:35] Test score for 1800 training labels: 0.7682
[2021-03-20 03:04:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:04:35] Found 50000 unlabelled features.
[2021-03-20 03:04:43] Showing labelled: True (2150/50000 visible, 50 redundant)
[2021-03-20 03:04:43] Creating trainer with model on device: cuda
[2021-03-20 03:04:43] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 03:05:18] Training accuracy: 0.8932
[2021-03-20 03:05:18] Testing on 10000 data points...
[2021-03-20 03:05:20] Test score for 2200 training labels: 0.7836
[2021-03-20 03:05:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:05:20] Found 50000 unlabelled features.
[2021-03-20 03:05:28] Showing labelled: True (2531/50000 visible, 69 redundant)
[2021-03-20 03:05:28] Creating trainer with model on device: cuda
[2021-03-20 03:05:28] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 03:06:08] Training accuracy: 0.8870
[2021-03-20 03:06:08] Testing on 10000 data points...
[2021-03-20 03:06:10] Test score for 2600 training labels: 0.7913
[2021-03-20 03:06:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:06:10] Found 50000 unlabelled features.
[2021-03-20 03:06:18] Showing labelled: True (2910/50000 visible, 90 redundant)
[2021-03-20 03:06:18] Creating trainer with model on device: cuda
[2021-03-20 03:06:18] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 03:07:05] Training accuracy: 0.8933
[2021-03-20 03:07:05] Testing on 10000 data points...
[2021-03-20 03:07:07] Test score for 3000 training labels: 0.8038
[2021-03-20 03:07:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:07:07] Found 50000 unlabelled features.
[2021-03-20 03:07:14] Showing labelled: True (3286/50000 visible, 114 redundant)
[2021-03-20 03:07:14] Creating trainer with model on device: cuda
[2021-03-20 03:07:14] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 03:08:08] Training accuracy: 0.9004
[2021-03-20 03:08:08] Testing on 10000 data points...
[2021-03-20 03:08:10] Test score for 3400 training labels: 0.8069
[2021-03-20 03:08:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:08:10] Found 50000 unlabelled features.
[2021-03-20 03:08:18] Showing labelled: True (3662/50000 visible, 138 redundant)
[2021-03-20 03:08:18] Creating trainer with model on device: cuda
[2021-03-20 03:08:18] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 03:09:18] Training accuracy: 0.9079
[2021-03-20 03:09:18] Testing on 10000 data points...
[2021-03-20 03:09:20] Test score for 3800 training labels: 0.8153
[2021-03-20 03:09:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:09:20] Found 50000 unlabelled features.
[2021-03-20 03:09:28] Showing labelled: True (4033/50000 visible, 167 redundant)
[2021-03-20 03:09:28] Creating trainer with model on device: cuda
[2021-03-20 03:09:28] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 03:10:34] Training accuracy: 0.8984
[2021-03-20 03:10:34] Testing on 10000 data points...
[2021-03-20 03:10:36] Test score for 4200 training labels: 0.8206
[2021-03-20 03:10:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:10:36] Found 50000 unlabelled features.
[2021-03-20 03:10:43] Showing labelled: True (4398/50000 visible, 202 redundant)
[2021-03-20 03:10:43] Creating trainer with model on device: cuda
[2021-03-20 03:10:43] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 03:11:56] Training accuracy: 0.8970
[2021-03-20 03:11:56] Testing on 10000 data points...
[2021-03-20 03:11:58] Test score for 4600 training labels: 0.8206
[2021-03-20 03:11:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:11:58] Found 50000 unlabelled features.
[2021-03-20 03:12:05] Showing labelled: True (4760/50000 visible, 240 redundant)
[2021-03-20 03:12:05] Creating trainer with model on device: cuda
[2021-03-20 03:12:05] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 03:13:24] Training accuracy: 0.9158
[2021-03-20 03:13:24] Testing on 10000 data points...
[2021-03-20 03:13:26] Test score for 5000 training labels: 0.8248
[2021-03-20 03:13:26] Seeded: 6
[2021-03-20 03:13:26] Running: experiment 1
[2021-03-20 03:13:26] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 03:13:26] Creating pretrained=True VGG16...
[2021-03-20 03:13:28] No parameter reset since we are using a pretrained model.
[2021-03-20 03:13:28] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 03:13:28] Creating trainer with model on device: cuda
[2021-03-20 03:13:28] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 03:14:31] Training accuracy: 0.8877
[2021-03-20 03:14:31] Testing on 10000 data points...
[2021-03-20 03:14:33] Test score for 1000 training labels: 0.7375
[2021-03-20 03:14:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:14:33] Found 50000 unlabelled features.
[2021-03-20 03:14:46] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:14:48] Searching for coresets with 20 beams...
[2021-03-20 03:15:17] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 03:15:17] Creating trainer with model on device: cuda
[2021-03-20 03:15:17] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 03:15:40] Training accuracy: 0.8799
[2021-03-20 03:15:40] Testing on 10000 data points...
[2021-03-20 03:15:42] Test score for 1400 training labels: 0.7555
[2021-03-20 03:15:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:15:42] Found 50000 unlabelled features.
[2021-03-20 03:15:54] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:15:58] Searching for coresets with 20 beams...
[2021-03-20 03:16:27] Showing labelled: True (1790/50000 visible, 10 redundant)
[2021-03-20 03:16:27] Creating trainer with model on device: cuda
[2021-03-20 03:16:27] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 03:16:56] Training accuracy: 0.8879
[2021-03-20 03:16:56] Testing on 10000 data points...
[2021-03-20 03:16:57] Test score for 1800 training labels: 0.7713
[2021-03-20 03:16:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:16:58] Found 50000 unlabelled features.
[2021-03-20 03:17:10] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:17:15] Searching for coresets with 20 beams...
[2021-03-20 03:17:44] Showing labelled: True (2167/50000 visible, 33 redundant)
[2021-03-20 03:17:44] Creating trainer with model on device: cuda
[2021-03-20 03:17:44] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 03:18:20] Training accuracy: 0.8774
[2021-03-20 03:18:20] Testing on 10000 data points...
[2021-03-20 03:18:22] Test score for 2200 training labels: 0.7879
[2021-03-20 03:18:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:18:22] Found 50000 unlabelled features.
[2021-03-20 03:18:34] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:18:40] Searching for coresets with 20 beams...
[2021-03-20 03:19:09] Showing labelled: True (2551/50000 visible, 49 redundant)
[2021-03-20 03:19:09] Creating trainer with model on device: cuda
[2021-03-20 03:19:09] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 03:19:51] Training accuracy: 0.8829
[2021-03-20 03:19:51] Testing on 10000 data points...
[2021-03-20 03:19:53] Test score for 2600 training labels: 0.7963
[2021-03-20 03:19:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:19:53] Found 50000 unlabelled features.
[2021-03-20 03:20:06] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:20:13] Searching for coresets with 20 beams...
[2021-03-20 03:20:42] Showing labelled: True (2930/50000 visible, 70 redundant)
[2021-03-20 03:20:42] Creating trainer with model on device: cuda
[2021-03-20 03:20:42] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 03:21:30] Training accuracy: 0.9031
[2021-03-20 03:21:30] Testing on 10000 data points...
[2021-03-20 03:21:32] Test score for 3000 training labels: 0.8068
[2021-03-20 03:21:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:21:32] Found 50000 unlabelled features.
[2021-03-20 03:21:44] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:21:53] Searching for coresets with 20 beams...
[2021-03-20 03:22:21] Showing labelled: True (3308/50000 visible, 92 redundant)
[2021-03-20 03:22:21] Creating trainer with model on device: cuda
[2021-03-20 03:22:21] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 03:23:15] Training accuracy: 0.8919
[2021-03-20 03:23:15] Testing on 10000 data points...
[2021-03-20 03:23:17] Test score for 3400 training labels: 0.8173
[2021-03-20 03:23:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:23:17] Found 50000 unlabelled features.
[2021-03-20 03:23:30] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:23:39] Searching for coresets with 20 beams...
[2021-03-20 03:24:08] Showing labelled: True (3680/50000 visible, 120 redundant)
[2021-03-20 03:24:08] Creating trainer with model on device: cuda
[2021-03-20 03:24:08] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 03:25:09] Training accuracy: 0.9031
[2021-03-20 03:25:09] Testing on 10000 data points...
[2021-03-20 03:25:11] Test score for 3800 training labels: 0.8313
[2021-03-20 03:25:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:25:11] Found 50000 unlabelled features.
[2021-03-20 03:25:24] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:25:34] Searching for coresets with 20 beams...
[2021-03-20 03:26:03] Showing labelled: True (4058/50000 visible, 142 redundant)
[2021-03-20 03:26:03] Creating trainer with model on device: cuda
[2021-03-20 03:26:03] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 03:27:10] Training accuracy: 0.9056
[2021-03-20 03:27:10] Testing on 10000 data points...
[2021-03-20 03:27:12] Test score for 4200 training labels: 0.8337
[2021-03-20 03:27:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:27:12] Found 50000 unlabelled features.
[2021-03-20 03:27:25] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:27:37] Searching for coresets with 20 beams...
[2021-03-20 03:28:05] Showing labelled: True (4429/50000 visible, 171 redundant)
[2021-03-20 03:28:05] Creating trainer with model on device: cuda
[2021-03-20 03:28:05] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 03:29:19] Training accuracy: 0.8935
[2021-03-20 03:29:19] Testing on 10000 data points...
[2021-03-20 03:29:21] Test score for 4600 training labels: 0.8406
[2021-03-20 03:29:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:29:21] Found 50000 unlabelled features.
[2021-03-20 03:29:34] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:29:47] Searching for coresets with 20 beams...
[2021-03-20 03:30:15] Showing labelled: True (4799/50000 visible, 201 redundant)
[2021-03-20 03:30:15] Creating trainer with model on device: cuda
[2021-03-20 03:30:15] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 03:31:35] Training accuracy: 0.8963
[2021-03-20 03:31:35] Testing on 10000 data points...
[2021-03-20 03:31:37] Test score for 5000 training labels: 0.8490
[2021-03-20 03:31:37] Seeded: 6
[2021-03-20 03:31:37] Running: experiment 2
[2021-03-20 03:31:37] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 03:31:37] Creating pretrained=True VGG16...
[2021-03-20 03:31:39] No parameter reset since we are using a pretrained model.
[2021-03-20 03:31:39] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 03:31:39] Creating trainer with model on device: cuda
[2021-03-20 03:31:39] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 03:32:42] Training accuracy: 0.8889
[2021-03-20 03:32:42] Testing on 10000 data points...
[2021-03-20 03:32:44] Test score for 1000 training labels: 0.7376
[2021-03-20 03:32:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:32:44] Found 50000 unlabelled features.
[2021-03-20 03:32:57] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:33:00] Searching for coresets with 20 beams...
[2021-03-20 03:33:28] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 03:33:28] Creating trainer with model on device: cuda
[2021-03-20 03:33:28] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 03:33:51] Training accuracy: 0.8815
[2021-03-20 03:33:51] Testing on 10000 data points...
[2021-03-20 03:33:53] Test score for 1400 training labels: 0.7547
[2021-03-20 03:33:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:33:53] Found 50000 unlabelled features.
[2021-03-20 03:34:05] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:34:09] Searching for coresets with 20 beams...
[2021-03-20 03:34:38] Showing labelled: True (1790/50000 visible, 10 redundant)
[2021-03-20 03:34:38] Creating trainer with model on device: cuda
[2021-03-20 03:34:38] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 03:35:07] Training accuracy: 0.8950
[2021-03-20 03:35:07] Testing on 10000 data points...
[2021-03-20 03:35:09] Test score for 1800 training labels: 0.7728
[2021-03-20 03:35:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:35:09] Found 50000 unlabelled features.
[2021-03-20 03:35:21] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:35:26] Searching for coresets with 20 beams...
[2021-03-20 03:35:55] Showing labelled: True (2172/50000 visible, 28 redundant)
[2021-03-20 03:35:55] Creating trainer with model on device: cuda
[2021-03-20 03:35:55] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 03:36:31] Training accuracy: 0.8702
[2021-03-20 03:36:31] Testing on 10000 data points...
[2021-03-20 03:36:33] Test score for 2200 training labels: 0.7869
[2021-03-20 03:36:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:36:33] Found 50000 unlabelled features.
[2021-03-20 03:36:45] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:36:52] Searching for coresets with 20 beams...
[2021-03-20 03:37:20] Showing labelled: True (2546/50000 visible, 54 redundant)
[2021-03-20 03:37:20] Creating trainer with model on device: cuda
[2021-03-20 03:37:20] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 03:38:02] Training accuracy: 0.8863
[2021-03-20 03:38:02] Testing on 10000 data points...
[2021-03-20 03:38:04] Test score for 2600 training labels: 0.7985
[2021-03-20 03:38:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:38:04] Found 50000 unlabelled features.
[2021-03-20 03:38:17] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:38:24] Searching for coresets with 20 beams...
[2021-03-20 03:38:52] Showing labelled: True (2923/50000 visible, 77 redundant)
[2021-03-20 03:38:52] Creating trainer with model on device: cuda
[2021-03-20 03:38:52] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 03:39:40] Training accuracy: 0.9013
[2021-03-20 03:39:40] Testing on 10000 data points...
[2021-03-20 03:39:42] Test score for 3000 training labels: 0.8094
[2021-03-20 03:39:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:39:42] Found 50000 unlabelled features.
[2021-03-20 03:39:55] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:40:03] Searching for coresets with 20 beams...
[2021-03-20 03:40:32] Showing labelled: True (3309/50000 visible, 91 redundant)
[2021-03-20 03:40:32] Creating trainer with model on device: cuda
[2021-03-20 03:40:32] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 03:41:26] Training accuracy: 0.8989
[2021-03-20 03:41:26] Testing on 10000 data points...
[2021-03-20 03:41:28] Test score for 3400 training labels: 0.8128
[2021-03-20 03:41:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:41:28] Found 50000 unlabelled features.
[2021-03-20 03:41:41] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:41:51] Searching for coresets with 20 beams...
[2021-03-20 03:42:19] Showing labelled: True (3678/50000 visible, 122 redundant)
[2021-03-20 03:42:19] Creating trainer with model on device: cuda
[2021-03-20 03:42:19] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 03:43:20] Training accuracy: 0.9012
[2021-03-20 03:43:20] Testing on 10000 data points...
[2021-03-20 03:43:22] Test score for 3800 training labels: 0.8222
[2021-03-20 03:43:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:43:22] Found 50000 unlabelled features.
[2021-03-20 03:43:35] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:43:46] Searching for coresets with 20 beams...
[2021-03-20 03:44:15] Showing labelled: True (4046/50000 visible, 154 redundant)
[2021-03-20 03:44:15] Creating trainer with model on device: cuda
[2021-03-20 03:44:15] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 03:45:21] Training accuracy: 0.9006
[2021-03-20 03:45:21] Testing on 10000 data points...
[2021-03-20 03:45:23] Test score for 4200 training labels: 0.8302
[2021-03-20 03:45:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:45:23] Found 50000 unlabelled features.
[2021-03-20 03:45:36] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:45:48] Searching for coresets with 20 beams...
[2021-03-20 03:46:17] Showing labelled: True (4434/50000 visible, 166 redundant)
[2021-03-20 03:46:17] Creating trainer with model on device: cuda
[2021-03-20 03:46:17] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 03:47:30] Training accuracy: 0.9000
[2021-03-20 03:47:30] Testing on 10000 data points...
[2021-03-20 03:47:32] Test score for 4600 training labels: 0.8399
[2021-03-20 03:47:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:47:32] Found 50000 unlabelled features.
[2021-03-20 03:47:45] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:47:58] Searching for coresets with 20 beams...
[2021-03-20 03:48:27] Showing labelled: True (4810/50000 visible, 190 redundant)
[2021-03-20 03:48:27] Creating trainer with model on device: cuda
[2021-03-20 03:48:27] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 03:49:46] Training accuracy: 0.9059
[2021-03-20 03:49:46] Testing on 10000 data points...
[2021-03-20 03:49:48] Test score for 5000 training labels: 0.8450
[2021-03-20 03:49:48] Seeded: 6
[2021-03-20 03:49:48] Running: experiment 3
[2021-03-20 03:49:48] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 03:49:48] Creating pretrained=True VGG16...
[2021-03-20 03:49:50] No parameter reset since we are using a pretrained model.
[2021-03-20 03:49:50] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 03:49:50] Creating trainer with model on device: cuda
[2021-03-20 03:49:50] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 03:50:53] Training accuracy: 0.8891
[2021-03-20 03:50:53] Testing on 10000 data points...
[2021-03-20 03:50:55] Test score for 1000 training labels: 0.7363
[2021-03-20 03:50:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:50:55] Found 50000 unlabelled features.
[2021-03-20 03:51:08] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:51:11] Searching for coresets greedily...
[2021-03-20 03:51:11] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 03:51:11] Creating trainer with model on device: cuda
[2021-03-20 03:51:11] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 03:51:33] Training accuracy: 0.8678
[2021-03-20 03:51:33] Testing on 10000 data points...
[2021-03-20 03:51:35] Test score for 1400 training labels: 0.7499
[2021-03-20 03:51:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:51:35] Found 50000 unlabelled features.
[2021-03-20 03:51:48] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:51:52] Searching for coresets greedily...
[2021-03-20 03:51:52] Showing labelled: True (1792/50000 visible, 8 redundant)
[2021-03-20 03:51:52] Creating trainer with model on device: cuda
[2021-03-20 03:51:52] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 03:52:21] Training accuracy: 0.8941
[2021-03-20 03:52:21] Testing on 10000 data points...
[2021-03-20 03:52:23] Test score for 1800 training labels: 0.7754
[2021-03-20 03:52:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:52:23] Found 50000 unlabelled features.
[2021-03-20 03:52:35] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:52:40] Searching for coresets greedily...
[2021-03-20 03:52:41] Showing labelled: True (2175/50000 visible, 25 redundant)
[2021-03-20 03:52:41] Creating trainer with model on device: cuda
[2021-03-20 03:52:41] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 03:53:17] Training accuracy: 0.8761
[2021-03-20 03:53:17] Testing on 10000 data points...
[2021-03-20 03:53:18] Test score for 2200 training labels: 0.7837
[2021-03-20 03:53:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:53:19] Found 50000 unlabelled features.
[2021-03-20 03:53:31] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:53:37] Searching for coresets greedily...
[2021-03-20 03:53:38] Showing labelled: True (2553/50000 visible, 47 redundant)
[2021-03-20 03:53:38] Creating trainer with model on device: cuda
[2021-03-20 03:53:38] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 03:54:19] Training accuracy: 0.8900
[2021-03-20 03:54:19] Testing on 10000 data points...
[2021-03-20 03:54:21] Test score for 2600 training labels: 0.8043
[2021-03-20 03:54:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:54:21] Found 50000 unlabelled features.
[2021-03-20 03:54:34] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:54:41] Searching for coresets greedily...
[2021-03-20 03:54:42] Showing labelled: True (2933/50000 visible, 67 redundant)
[2021-03-20 03:54:42] Creating trainer with model on device: cuda
[2021-03-20 03:54:42] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 03:55:29] Training accuracy: 0.8997
[2021-03-20 03:55:29] Testing on 10000 data points...
[2021-03-20 03:55:31] Test score for 3000 training labels: 0.8093
[2021-03-20 03:55:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:55:31] Found 50000 unlabelled features.
[2021-03-20 03:55:44] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:55:52] Searching for coresets greedily...
[2021-03-20 03:55:53] Showing labelled: True (3305/50000 visible, 95 redundant)
[2021-03-20 03:55:53] Creating trainer with model on device: cuda
[2021-03-20 03:55:53] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 03:56:47] Training accuracy: 0.8943
[2021-03-20 03:56:47] Testing on 10000 data points...
[2021-03-20 03:56:49] Test score for 3400 training labels: 0.8160
[2021-03-20 03:56:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:56:49] Found 50000 unlabelled features.
[2021-03-20 03:57:02] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:57:11] Searching for coresets greedily...
[2021-03-20 03:57:12] Showing labelled: True (3670/50000 visible, 130 redundant)
[2021-03-20 03:57:12] Creating trainer with model on device: cuda
[2021-03-20 03:57:12] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 03:58:13] Training accuracy: 0.8992
[2021-03-20 03:58:13] Testing on 10000 data points...
[2021-03-20 03:58:15] Test score for 3800 training labels: 0.8267
[2021-03-20 03:58:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:58:15] Found 50000 unlabelled features.
[2021-03-20 03:58:28] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 03:58:38] Searching for coresets greedily...
[2021-03-20 03:58:39] Showing labelled: True (4048/50000 visible, 152 redundant)
[2021-03-20 03:58:39] Creating trainer with model on device: cuda
[2021-03-20 03:58:39] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 03:59:46] Training accuracy: 0.9050
[2021-03-20 03:59:46] Testing on 10000 data points...
[2021-03-20 03:59:48] Test score for 4200 training labels: 0.8300
[2021-03-20 03:59:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 03:59:48] Found 50000 unlabelled features.
[2021-03-20 04:00:01] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:00:13] Searching for coresets greedily...
[2021-03-20 04:00:14] Showing labelled: True (4423/50000 visible, 177 redundant)
[2021-03-20 04:00:14] Creating trainer with model on device: cuda
[2021-03-20 04:00:14] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 04:01:27] Training accuracy: 0.8983
[2021-03-20 04:01:27] Testing on 10000 data points...
[2021-03-20 04:01:29] Test score for 4600 training labels: 0.8379
[2021-03-20 04:01:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:01:29] Found 50000 unlabelled features.
[2021-03-20 04:01:42] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:01:54] Searching for coresets greedily...
[2021-03-20 04:01:55] Showing labelled: True (4797/50000 visible, 203 redundant)
[2021-03-20 04:01:55] Creating trainer with model on device: cuda
[2021-03-20 04:01:55] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 04:03:15] Training accuracy: 0.8983
[2021-03-20 04:03:15] Testing on 10000 data points...
[2021-03-20 04:03:17] Test score for 5000 training labels: 0.8430
[2021-03-20 04:03:17] Seeded: 6
[2021-03-20 04:03:17] Running: experiment 4
[2021-03-20 04:03:17] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 04:03:17] Creating pretrained=True VGG16...
[2021-03-20 04:03:19] No parameter reset since we are using a pretrained model.
[2021-03-20 04:03:19] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 04:03:19] Creating trainer with model on device: cuda
[2021-03-20 04:03:19] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 04:04:23] Training accuracy: 0.8886
[2021-03-20 04:04:23] Testing on 10000 data points...
[2021-03-20 04:04:25] Test score for 1000 training labels: 0.7369
[2021-03-20 04:04:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:04:25] Found 50000 unlabelled features.
[2021-03-20 04:04:37] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 04:04:37] Creating trainer with model on device: cuda
[2021-03-20 04:04:37] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 04:04:59] Training accuracy: 0.8535
[2021-03-20 04:04:59] Testing on 10000 data points...
[2021-03-20 04:05:01] Test score for 1400 training labels: 0.7595
[2021-03-20 04:05:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:05:01] Found 50000 unlabelled features.
[2021-03-20 04:05:13] Showing labelled: True (1787/50000 visible, 13 redundant)
[2021-03-20 04:05:13] Creating trainer with model on device: cuda
[2021-03-20 04:05:13] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 04:05:41] Training accuracy: 0.8614
[2021-03-20 04:05:41] Testing on 10000 data points...
[2021-03-20 04:05:43] Test score for 1800 training labels: 0.7823
[2021-03-20 04:05:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:05:43] Found 50000 unlabelled features.
[2021-03-20 04:05:55] Showing labelled: True (2163/50000 visible, 37 redundant)
[2021-03-20 04:05:55] Creating trainer with model on device: cuda
[2021-03-20 04:05:55] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 04:06:31] Training accuracy: 0.8717
[2021-03-20 04:06:31] Testing on 10000 data points...
[2021-03-20 04:06:32] Test score for 2200 training labels: 0.7984
[2021-03-20 04:06:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:06:33] Found 50000 unlabelled features.
[2021-03-20 04:06:45] Showing labelled: True (2539/50000 visible, 61 redundant)
[2021-03-20 04:06:45] Creating trainer with model on device: cuda
[2021-03-20 04:06:45] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 04:07:26] Training accuracy: 0.8697
[2021-03-20 04:07:26] Testing on 10000 data points...
[2021-03-20 04:07:28] Test score for 2600 training labels: 0.8150
[2021-03-20 04:07:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:07:28] Found 50000 unlabelled features.
[2021-03-20 04:07:39] Showing labelled: True (2918/50000 visible, 82 redundant)
[2021-03-20 04:07:39] Creating trainer with model on device: cuda
[2021-03-20 04:07:39] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 04:08:27] Training accuracy: 0.8871
[2021-03-20 04:08:27] Testing on 10000 data points...
[2021-03-20 04:08:29] Test score for 3000 training labels: 0.8182
[2021-03-20 04:08:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:08:29] Found 50000 unlabelled features.
[2021-03-20 04:08:41] Showing labelled: True (3299/50000 visible, 101 redundant)
[2021-03-20 04:08:41] Creating trainer with model on device: cuda
[2021-03-20 04:08:41] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 04:09:35] Training accuracy: 0.8891
[2021-03-20 04:09:35] Testing on 10000 data points...
[2021-03-20 04:09:36] Test score for 3400 training labels: 0.8193
[2021-03-20 04:09:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:09:36] Found 50000 unlabelled features.
[2021-03-20 04:09:49] Showing labelled: True (3684/50000 visible, 116 redundant)
[2021-03-20 04:09:49] Creating trainer with model on device: cuda
[2021-03-20 04:09:49] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 04:10:49] Training accuracy: 0.9019
[2021-03-20 04:10:49] Testing on 10000 data points...
[2021-03-20 04:10:51] Test score for 3800 training labels: 0.8350
[2021-03-20 04:10:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:10:51] Found 50000 unlabelled features.
[2021-03-20 04:11:03] Showing labelled: True (4069/50000 visible, 131 redundant)
[2021-03-20 04:11:03] Creating trainer with model on device: cuda
[2021-03-20 04:11:03] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 04:12:10] Training accuracy: 0.8873
[2021-03-20 04:12:10] Testing on 10000 data points...
[2021-03-20 04:12:12] Test score for 4200 training labels: 0.8435
[2021-03-20 04:12:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:12:12] Found 50000 unlabelled features.
[2021-03-20 04:12:24] Showing labelled: True (4452/50000 visible, 148 redundant)
[2021-03-20 04:12:24] Creating trainer with model on device: cuda
[2021-03-20 04:12:24] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 04:13:37] Training accuracy: 0.8965
[2021-03-20 04:13:37] Testing on 10000 data points...
[2021-03-20 04:13:39] Test score for 4600 training labels: 0.8489
[2021-03-20 04:13:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:13:39] Found 50000 unlabelled features.
[2021-03-20 04:13:50] Showing labelled: True (4835/50000 visible, 165 redundant)
[2021-03-20 04:13:50] Creating trainer with model on device: cuda
[2021-03-20 04:13:50] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 04:15:10] Training accuracy: 0.9102
[2021-03-20 04:15:10] Testing on 10000 data points...
[2021-03-20 04:15:12] Test score for 5000 training labels: 0.8522
[2021-03-20 04:15:12] Experiment repeat 2/4
[2021-03-20 04:15:12] Seeded: 6
[2021-03-20 04:15:12] Using 2.00% labels of the dataset (1000/50000)
[2021-03-20 04:15:12] Randomly labelled 1000/50000
[2021-03-20 04:15:12] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 04:15:12] Seeded: 7
[2021-03-20 04:15:12] Running: experiment 0
[2021-03-20 04:15:12] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 04:15:12] Creating pretrained=True VGG16...
[2021-03-20 04:15:13] No parameter reset since we are using a pretrained model.
[2021-03-20 04:15:13] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 04:15:13] Creating trainer with model on device: cuda
[2021-03-20 04:15:13] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 04:16:17] Training accuracy: 0.9011
[2021-03-20 04:16:17] Testing on 10000 data points...
[2021-03-20 04:16:19] Test score for 1000 training labels: 0.7494
[2021-03-20 04:16:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:16:19] Found 50000 unlabelled features.
[2021-03-20 04:16:27] Showing labelled: True (1389/50000 visible, 11 redundant)
[2021-03-20 04:16:27] Creating trainer with model on device: cuda
[2021-03-20 04:16:27] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 04:16:49] Training accuracy: 0.8818
[2021-03-20 04:16:49] Testing on 10000 data points...
[2021-03-20 04:16:51] Test score for 1400 training labels: 0.7618
[2021-03-20 04:16:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:16:51] Found 50000 unlabelled features.
[2021-03-20 04:16:58] Showing labelled: True (1779/50000 visible, 21 redundant)
[2021-03-20 04:16:58] Creating trainer with model on device: cuda
[2021-03-20 04:16:58] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 04:17:27] Training accuracy: 0.8921
[2021-03-20 04:17:27] Testing on 10000 data points...
[2021-03-20 04:17:28] Test score for 1800 training labels: 0.7704
[2021-03-20 04:17:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:17:29] Found 50000 unlabelled features.
[2021-03-20 04:17:36] Showing labelled: True (2163/50000 visible, 37 redundant)
[2021-03-20 04:17:36] Creating trainer with model on device: cuda
[2021-03-20 04:17:36] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 04:18:11] Training accuracy: 0.8978
[2021-03-20 04:18:11] Testing on 10000 data points...
[2021-03-20 04:18:13] Test score for 2200 training labels: 0.7885
[2021-03-20 04:18:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:18:13] Found 50000 unlabelled features.
[2021-03-20 04:18:21] Showing labelled: True (2547/50000 visible, 53 redundant)
[2021-03-20 04:18:21] Creating trainer with model on device: cuda
[2021-03-20 04:18:21] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 04:19:02] Training accuracy: 0.8862
[2021-03-20 04:19:02] Testing on 10000 data points...
[2021-03-20 04:19:04] Test score for 2600 training labels: 0.7918
[2021-03-20 04:19:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:19:04] Found 50000 unlabelled features.
[2021-03-20 04:19:12] Showing labelled: True (2917/50000 visible, 83 redundant)
[2021-03-20 04:19:12] Creating trainer with model on device: cuda
[2021-03-20 04:19:12] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 04:19:59] Training accuracy: 0.8968
[2021-03-20 04:19:59] Testing on 10000 data points...
[2021-03-20 04:20:01] Test score for 3000 training labels: 0.8054
[2021-03-20 04:20:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:20:01] Found 50000 unlabelled features.
[2021-03-20 04:20:09] Showing labelled: True (3299/50000 visible, 101 redundant)
[2021-03-20 04:20:09] Creating trainer with model on device: cuda
[2021-03-20 04:20:09] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 04:21:02] Training accuracy: 0.9052
[2021-03-20 04:21:02] Testing on 10000 data points...
[2021-03-20 04:21:04] Test score for 3400 training labels: 0.8115
[2021-03-20 04:21:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:21:04] Found 50000 unlabelled features.
[2021-03-20 04:21:12] Showing labelled: True (3672/50000 visible, 128 redundant)
[2021-03-20 04:21:12] Creating trainer with model on device: cuda
[2021-03-20 04:21:12] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 04:22:13] Training accuracy: 0.9095
[2021-03-20 04:22:13] Testing on 10000 data points...
[2021-03-20 04:22:15] Test score for 3800 training labels: 0.8176
[2021-03-20 04:22:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:22:15] Found 50000 unlabelled features.
[2021-03-20 04:22:22] Showing labelled: True (4044/50000 visible, 156 redundant)
[2021-03-20 04:22:22] Creating trainer with model on device: cuda
[2021-03-20 04:22:22] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 04:23:29] Training accuracy: 0.8935
[2021-03-20 04:23:29] Testing on 10000 data points...
[2021-03-20 04:23:31] Test score for 4200 training labels: 0.8219
[2021-03-20 04:23:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:23:31] Found 50000 unlabelled features.
[2021-03-20 04:23:38] Showing labelled: True (4415/50000 visible, 185 redundant)
[2021-03-20 04:23:38] Creating trainer with model on device: cuda
[2021-03-20 04:23:38] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 04:24:51] Training accuracy: 0.9038
[2021-03-20 04:24:51] Testing on 10000 data points...
[2021-03-20 04:24:53] Test score for 4600 training labels: 0.8244
[2021-03-20 04:24:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:24:53] Found 50000 unlabelled features.
[2021-03-20 04:25:01] Showing labelled: True (4769/50000 visible, 231 redundant)
[2021-03-20 04:25:01] Creating trainer with model on device: cuda
[2021-03-20 04:25:01] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 04:26:20] Training accuracy: 0.9195
[2021-03-20 04:26:20] Testing on 10000 data points...
[2021-03-20 04:26:22] Test score for 5000 training labels: 0.8275
[2021-03-20 04:26:22] Seeded: 7
[2021-03-20 04:26:22] Running: experiment 1
[2021-03-20 04:26:22] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 04:26:22] Creating pretrained=True VGG16...
[2021-03-20 04:26:24] No parameter reset since we are using a pretrained model.
[2021-03-20 04:26:24] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 04:26:24] Creating trainer with model on device: cuda
[2021-03-20 04:26:24] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 04:27:27] Training accuracy: 0.9024
[2021-03-20 04:27:27] Testing on 10000 data points...
[2021-03-20 04:27:29] Test score for 1000 training labels: 0.7494
[2021-03-20 04:27:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:27:29] Found 50000 unlabelled features.
[2021-03-20 04:27:42] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:27:45] Searching for coresets with 20 beams...
[2021-03-20 04:28:14] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-20 04:28:14] Creating trainer with model on device: cuda
[2021-03-20 04:28:14] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 04:28:36] Training accuracy: 0.8705
[2021-03-20 04:28:36] Testing on 10000 data points...
[2021-03-20 04:28:38] Test score for 1400 training labels: 0.7668
[2021-03-20 04:28:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:28:38] Found 50000 unlabelled features.
[2021-03-20 04:28:50] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:28:54] Searching for coresets with 20 beams...
[2021-03-20 04:29:23] Showing labelled: True (1787/50000 visible, 13 redundant)
[2021-03-20 04:29:23] Creating trainer with model on device: cuda
[2021-03-20 04:29:23] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 04:29:52] Training accuracy: 0.8891
[2021-03-20 04:29:52] Testing on 10000 data points...
[2021-03-20 04:29:54] Test score for 1800 training labels: 0.7768
[2021-03-20 04:29:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:29:54] Found 50000 unlabelled features.
[2021-03-20 04:30:06] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:30:11] Searching for coresets with 20 beams...
[2021-03-20 04:30:40] Showing labelled: True (2166/50000 visible, 34 redundant)
[2021-03-20 04:30:40] Creating trainer with model on device: cuda
[2021-03-20 04:30:40] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 04:31:16] Training accuracy: 0.8862
[2021-03-20 04:31:16] Testing on 10000 data points...
[2021-03-20 04:31:18] Test score for 2200 training labels: 0.7945
[2021-03-20 04:31:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:31:18] Found 50000 unlabelled features.
[2021-03-20 04:31:31] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:31:37] Searching for coresets with 20 beams...
[2021-03-20 04:32:06] Showing labelled: True (2546/50000 visible, 54 redundant)
[2021-03-20 04:32:06] Creating trainer with model on device: cuda
[2021-03-20 04:32:06] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 04:32:47] Training accuracy: 0.8828
[2021-03-20 04:32:47] Testing on 10000 data points...
[2021-03-20 04:32:49] Test score for 2600 training labels: 0.8029
[2021-03-20 04:32:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:32:49] Found 50000 unlabelled features.
[2021-03-20 04:33:02] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:33:09] Searching for coresets with 20 beams...
[2021-03-20 04:33:38] Showing labelled: True (2926/50000 visible, 74 redundant)
[2021-03-20 04:33:38] Creating trainer with model on device: cuda
[2021-03-20 04:33:38] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 04:34:26] Training accuracy: 0.8917
[2021-03-20 04:34:26] Testing on 10000 data points...
[2021-03-20 04:34:28] Test score for 3000 training labels: 0.8114
[2021-03-20 04:34:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:34:28] Found 50000 unlabelled features.
[2021-03-20 04:34:40] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:34:49] Searching for coresets with 20 beams...
[2021-03-20 04:35:18] Showing labelled: True (3309/50000 visible, 91 redundant)
[2021-03-20 04:35:18] Creating trainer with model on device: cuda
[2021-03-20 04:35:18] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 04:36:12] Training accuracy: 0.8955
[2021-03-20 04:36:12] Testing on 10000 data points...
[2021-03-20 04:36:14] Test score for 3400 training labels: 0.8246
[2021-03-20 04:36:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:36:14] Found 50000 unlabelled features.
[2021-03-20 04:36:27] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:36:36] Searching for coresets with 20 beams...
[2021-03-20 04:37:05] Showing labelled: True (3693/50000 visible, 107 redundant)
[2021-03-20 04:37:05] Creating trainer with model on device: cuda
[2021-03-20 04:37:05] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 04:38:07] Training accuracy: 0.9010
[2021-03-20 04:38:07] Testing on 10000 data points...
[2021-03-20 04:38:09] Test score for 3800 training labels: 0.8307
[2021-03-20 04:38:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:38:09] Found 50000 unlabelled features.
[2021-03-20 04:38:22] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:38:32] Searching for coresets with 20 beams...
[2021-03-20 04:39:01] Showing labelled: True (4077/50000 visible, 123 redundant)
[2021-03-20 04:39:01] Creating trainer with model on device: cuda
[2021-03-20 04:39:01] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 04:40:09] Training accuracy: 0.9002
[2021-03-20 04:40:09] Testing on 10000 data points...
[2021-03-20 04:40:11] Test score for 4200 training labels: 0.8372
[2021-03-20 04:40:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:40:11] Found 50000 unlabelled features.
[2021-03-20 04:40:24] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:40:35] Searching for coresets with 20 beams...
[2021-03-20 04:41:04] Showing labelled: True (4454/50000 visible, 146 redundant)
[2021-03-20 04:41:04] Creating trainer with model on device: cuda
[2021-03-20 04:41:04] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 04:42:17] Training accuracy: 0.8985
[2021-03-20 04:42:17] Testing on 10000 data points...
[2021-03-20 04:42:19] Test score for 4600 training labels: 0.8418
[2021-03-20 04:42:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:42:19] Found 50000 unlabelled features.
[2021-03-20 04:42:32] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:42:45] Searching for coresets with 20 beams...
[2021-03-20 04:43:14] Showing labelled: True (4820/50000 visible, 180 redundant)
[2021-03-20 04:43:14] Creating trainer with model on device: cuda
[2021-03-20 04:43:14] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 04:44:33] Training accuracy: 0.9025
[2021-03-20 04:44:33] Testing on 10000 data points...
[2021-03-20 04:44:35] Test score for 5000 training labels: 0.8472
[2021-03-20 04:44:35] Seeded: 7
[2021-03-20 04:44:35] Running: experiment 2
[2021-03-20 04:44:35] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 04:44:35] Creating pretrained=True VGG16...
[2021-03-20 04:44:37] No parameter reset since we are using a pretrained model.
[2021-03-20 04:44:37] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 04:44:37] Creating trainer with model on device: cuda
[2021-03-20 04:44:37] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 04:45:41] Training accuracy: 0.9040
[2021-03-20 04:45:41] Testing on 10000 data points...
[2021-03-20 04:45:43] Test score for 1000 training labels: 0.7493
[2021-03-20 04:45:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:45:43] Found 50000 unlabelled features.
[2021-03-20 04:45:55] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:45:58] Searching for coresets with 20 beams...
[2021-03-20 04:46:26] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 04:46:26] Creating trainer with model on device: cuda
[2021-03-20 04:46:26] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 04:46:49] Training accuracy: 0.8708
[2021-03-20 04:46:49] Testing on 10000 data points...
[2021-03-20 04:46:51] Test score for 1400 training labels: 0.7606
[2021-03-20 04:46:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:46:51] Found 50000 unlabelled features.
[2021-03-20 04:47:03] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:47:07] Searching for coresets with 20 beams...
[2021-03-20 04:47:36] Showing labelled: True (1789/50000 visible, 11 redundant)
[2021-03-20 04:47:36] Creating trainer with model on device: cuda
[2021-03-20 04:47:36] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 04:48:05] Training accuracy: 0.8838
[2021-03-20 04:48:05] Testing on 10000 data points...
[2021-03-20 04:48:07] Test score for 1800 training labels: 0.7802
[2021-03-20 04:48:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:48:07] Found 50000 unlabelled features.
[2021-03-20 04:48:20] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:48:25] Searching for coresets with 20 beams...
[2021-03-20 04:48:53] Showing labelled: True (2157/50000 visible, 43 redundant)
[2021-03-20 04:48:53] Creating trainer with model on device: cuda
[2021-03-20 04:48:53] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 04:49:29] Training accuracy: 0.8839
[2021-03-20 04:49:29] Testing on 10000 data points...
[2021-03-20 04:49:31] Test score for 2200 training labels: 0.7911
[2021-03-20 04:49:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:49:31] Found 50000 unlabelled features.
[2021-03-20 04:49:44] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:49:50] Searching for coresets with 20 beams...
[2021-03-20 04:50:19] Showing labelled: True (2535/50000 visible, 65 redundant)
[2021-03-20 04:50:19] Creating trainer with model on device: cuda
[2021-03-20 04:50:19] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 04:51:00] Training accuracy: 0.8882
[2021-03-20 04:51:00] Testing on 10000 data points...
[2021-03-20 04:51:02] Test score for 2600 training labels: 0.8051
[2021-03-20 04:51:02] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:51:02] Found 50000 unlabelled features.
[2021-03-20 04:51:15] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:51:22] Searching for coresets with 20 beams...
[2021-03-20 04:51:51] Showing labelled: True (2917/50000 visible, 83 redundant)
[2021-03-20 04:51:51] Creating trainer with model on device: cuda
[2021-03-20 04:51:51] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 04:52:39] Training accuracy: 0.8886
[2021-03-20 04:52:39] Testing on 10000 data points...
[2021-03-20 04:52:40] Test score for 3000 training labels: 0.8104
[2021-03-20 04:52:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:52:41] Found 50000 unlabelled features.
[2021-03-20 04:52:53] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:53:01] Searching for coresets with 20 beams...
[2021-03-20 04:53:30] Showing labelled: True (3299/50000 visible, 101 redundant)
[2021-03-20 04:53:30] Creating trainer with model on device: cuda
[2021-03-20 04:53:30] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 04:54:24] Training accuracy: 0.8934
[2021-03-20 04:54:24] Testing on 10000 data points...
[2021-03-20 04:54:26] Test score for 3400 training labels: 0.8236
[2021-03-20 04:54:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:54:26] Found 50000 unlabelled features.
[2021-03-20 04:54:39] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:54:49] Searching for coresets with 20 beams...
[2021-03-20 04:55:17] Showing labelled: True (3680/50000 visible, 120 redundant)
[2021-03-20 04:55:17] Creating trainer with model on device: cuda
[2021-03-20 04:55:17] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 04:56:19] Training accuracy: 0.9009
[2021-03-20 04:56:19] Testing on 10000 data points...
[2021-03-20 04:56:21] Test score for 3800 training labels: 0.8283
[2021-03-20 04:56:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:56:21] Found 50000 unlabelled features.
[2021-03-20 04:56:34] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:56:44] Searching for coresets with 20 beams...
[2021-03-20 04:57:13] Showing labelled: True (4062/50000 visible, 138 redundant)
[2021-03-20 04:57:13] Creating trainer with model on device: cuda
[2021-03-20 04:57:13] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 04:58:20] Training accuracy: 0.9047
[2021-03-20 04:58:20] Testing on 10000 data points...
[2021-03-20 04:58:22] Test score for 4200 training labels: 0.8361
[2021-03-20 04:58:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 04:58:22] Found 50000 unlabelled features.
[2021-03-20 04:58:35] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 04:58:47] Searching for coresets with 20 beams...
[2021-03-20 04:59:15] Showing labelled: True (4441/50000 visible, 159 redundant)
[2021-03-20 04:59:15] Creating trainer with model on device: cuda
[2021-03-20 04:59:15] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 05:00:29] Training accuracy: 0.8996
[2021-03-20 05:00:29] Testing on 10000 data points...
[2021-03-20 05:00:31] Test score for 4600 training labels: 0.8463
[2021-03-20 05:00:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:00:31] Found 50000 unlabelled features.
[2021-03-20 05:00:44] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:00:56] Searching for coresets with 20 beams...
[2021-03-20 05:01:25] Showing labelled: True (4821/50000 visible, 179 redundant)
[2021-03-20 05:01:25] Creating trainer with model on device: cuda
[2021-03-20 05:01:25] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 05:02:45] Training accuracy: 0.8880
[2021-03-20 05:02:45] Testing on 10000 data points...
[2021-03-20 05:02:47] Test score for 5000 training labels: 0.8488
[2021-03-20 05:02:47] Seeded: 7
[2021-03-20 05:02:47] Running: experiment 3
[2021-03-20 05:02:47] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 05:02:47] Creating pretrained=True VGG16...
[2021-03-20 05:02:48] No parameter reset since we are using a pretrained model.
[2021-03-20 05:02:48] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 05:02:48] Creating trainer with model on device: cuda
[2021-03-20 05:02:48] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 05:03:52] Training accuracy: 0.9021
[2021-03-20 05:03:52] Testing on 10000 data points...
[2021-03-20 05:03:53] Test score for 1000 training labels: 0.7475
[2021-03-20 05:03:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:03:54] Found 50000 unlabelled features.
[2021-03-20 05:04:06] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:04:08] Searching for coresets greedily...
[2021-03-20 05:04:09] Showing labelled: True (1398/50000 visible, 2 redundant)
[2021-03-20 05:04:09] Creating trainer with model on device: cuda
[2021-03-20 05:04:09] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 05:04:31] Training accuracy: 0.8760
[2021-03-20 05:04:31] Testing on 10000 data points...
[2021-03-20 05:04:33] Test score for 1400 training labels: 0.7646
[2021-03-20 05:04:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:04:33] Found 50000 unlabelled features.
[2021-03-20 05:04:46] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:04:50] Searching for coresets greedily...
[2021-03-20 05:04:50] Showing labelled: True (1789/50000 visible, 11 redundant)
[2021-03-20 05:04:50] Creating trainer with model on device: cuda
[2021-03-20 05:04:50] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 05:05:19] Training accuracy: 0.8818
[2021-03-20 05:05:19] Testing on 10000 data points...
[2021-03-20 05:05:21] Test score for 1800 training labels: 0.7836
[2021-03-20 05:05:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:05:21] Found 50000 unlabelled features.
[2021-03-20 05:05:33] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:05:38] Searching for coresets greedily...
[2021-03-20 05:05:39] Showing labelled: True (2167/50000 visible, 33 redundant)
[2021-03-20 05:05:39] Creating trainer with model on device: cuda
[2021-03-20 05:05:39] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 05:06:15] Training accuracy: 0.8808
[2021-03-20 05:06:15] Testing on 10000 data points...
[2021-03-20 05:06:16] Test score for 2200 training labels: 0.7994
[2021-03-20 05:06:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:06:16] Found 50000 unlabelled features.
[2021-03-20 05:06:29] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:06:35] Searching for coresets greedily...
[2021-03-20 05:06:36] Showing labelled: True (2542/50000 visible, 58 redundant)
[2021-03-20 05:06:36] Creating trainer with model on device: cuda
[2021-03-20 05:06:36] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 05:07:17] Training accuracy: 0.8823
[2021-03-20 05:07:17] Testing on 10000 data points...
[2021-03-20 05:07:19] Test score for 2600 training labels: 0.8079
[2021-03-20 05:07:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:07:19] Found 50000 unlabelled features.
[2021-03-20 05:07:32] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:07:39] Searching for coresets greedily...
[2021-03-20 05:07:40] Showing labelled: True (2923/50000 visible, 77 redundant)
[2021-03-20 05:07:40] Creating trainer with model on device: cuda
[2021-03-20 05:07:40] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 05:08:27] Training accuracy: 0.8905
[2021-03-20 05:08:27] Testing on 10000 data points...
[2021-03-20 05:08:29] Test score for 3000 training labels: 0.8188
[2021-03-20 05:08:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:08:29] Found 50000 unlabelled features.
[2021-03-20 05:08:42] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:08:50] Searching for coresets greedily...
[2021-03-20 05:08:51] Showing labelled: True (3304/50000 visible, 96 redundant)
[2021-03-20 05:08:51] Creating trainer with model on device: cuda
[2021-03-20 05:08:51] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 05:09:45] Training accuracy: 0.8848
[2021-03-20 05:09:45] Testing on 10000 data points...
[2021-03-20 05:09:47] Test score for 3400 training labels: 0.8182
[2021-03-20 05:09:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:09:47] Found 50000 unlabelled features.
[2021-03-20 05:10:00] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:10:09] Searching for coresets greedily...
[2021-03-20 05:10:10] Showing labelled: True (3680/50000 visible, 120 redundant)
[2021-03-20 05:10:10] Creating trainer with model on device: cuda
[2021-03-20 05:10:10] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 05:11:11] Training accuracy: 0.9026
[2021-03-20 05:11:11] Testing on 10000 data points...
[2021-03-20 05:11:13] Test score for 3800 training labels: 0.8278
[2021-03-20 05:11:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:11:13] Found 50000 unlabelled features.
[2021-03-20 05:11:26] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:11:37] Searching for coresets greedily...
[2021-03-20 05:11:38] Showing labelled: True (4052/50000 visible, 148 redundant)
[2021-03-20 05:11:38] Creating trainer with model on device: cuda
[2021-03-20 05:11:38] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 05:12:44] Training accuracy: 0.9018
[2021-03-20 05:12:44] Testing on 10000 data points...
[2021-03-20 05:12:46] Test score for 4200 training labels: 0.8328
[2021-03-20 05:12:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:12:46] Found 50000 unlabelled features.
[2021-03-20 05:13:00] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:13:11] Searching for coresets greedily...
[2021-03-20 05:13:12] Showing labelled: True (4427/50000 visible, 173 redundant)
[2021-03-20 05:13:12] Creating trainer with model on device: cuda
[2021-03-20 05:13:12] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 05:14:25] Training accuracy: 0.8982
[2021-03-20 05:14:25] Testing on 10000 data points...
[2021-03-20 05:14:27] Test score for 4600 training labels: 0.8387
[2021-03-20 05:14:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:14:27] Found 50000 unlabelled features.
[2021-03-20 05:14:41] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:14:53] Searching for coresets greedily...
[2021-03-20 05:14:54] Showing labelled: True (4794/50000 visible, 206 redundant)
[2021-03-20 05:14:54] Creating trainer with model on device: cuda
[2021-03-20 05:14:54] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 05:16:14] Training accuracy: 0.9020
[2021-03-20 05:16:14] Testing on 10000 data points...
[2021-03-20 05:16:16] Test score for 5000 training labels: 0.8414
[2021-03-20 05:16:16] Seeded: 7
[2021-03-20 05:16:16] Running: experiment 4
[2021-03-20 05:16:16] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 05:16:16] Creating pretrained=True VGG16...
[2021-03-20 05:16:17] No parameter reset since we are using a pretrained model.
[2021-03-20 05:16:17] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 05:16:17] Creating trainer with model on device: cuda
[2021-03-20 05:16:17] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 05:17:21] Training accuracy: 0.8993
[2021-03-20 05:17:21] Testing on 10000 data points...
[2021-03-20 05:17:23] Test score for 1000 training labels: 0.7484
[2021-03-20 05:17:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:17:23] Found 50000 unlabelled features.
[2021-03-20 05:17:35] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 05:17:35] Creating trainer with model on device: cuda
[2021-03-20 05:17:35] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 05:17:57] Training accuracy: 0.8515
[2021-03-20 05:17:57] Testing on 10000 data points...
[2021-03-20 05:17:59] Test score for 1400 training labels: 0.7689
[2021-03-20 05:17:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:17:59] Found 50000 unlabelled features.
[2021-03-20 05:18:11] Showing labelled: True (1779/50000 visible, 21 redundant)
[2021-03-20 05:18:11] Creating trainer with model on device: cuda
[2021-03-20 05:18:11] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 05:18:40] Training accuracy: 0.8492
[2021-03-20 05:18:40] Testing on 10000 data points...
[2021-03-20 05:18:41] Test score for 1800 training labels: 0.7899
[2021-03-20 05:18:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:18:42] Found 50000 unlabelled features.
[2021-03-20 05:18:54] Showing labelled: True (2159/50000 visible, 41 redundant)
[2021-03-20 05:18:54] Creating trainer with model on device: cuda
[2021-03-20 05:18:54] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 05:19:29] Training accuracy: 0.8659
[2021-03-20 05:19:29] Testing on 10000 data points...
[2021-03-20 05:19:31] Test score for 2200 training labels: 0.8018
[2021-03-20 05:19:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:19:31] Found 50000 unlabelled features.
[2021-03-20 05:19:43] Showing labelled: True (2542/50000 visible, 58 redundant)
[2021-03-20 05:19:43] Creating trainer with model on device: cuda
[2021-03-20 05:19:43] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 05:20:24] Training accuracy: 0.8716
[2021-03-20 05:20:24] Testing on 10000 data points...
[2021-03-20 05:20:26] Test score for 2600 training labels: 0.8119
[2021-03-20 05:20:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:20:26] Found 50000 unlabelled features.
[2021-03-20 05:20:38] Showing labelled: True (2918/50000 visible, 82 redundant)
[2021-03-20 05:20:38] Creating trainer with model on device: cuda
[2021-03-20 05:20:38] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 05:21:25] Training accuracy: 0.8901
[2021-03-20 05:21:25] Testing on 10000 data points...
[2021-03-20 05:21:27] Test score for 3000 training labels: 0.8232
[2021-03-20 05:21:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:21:27] Found 50000 unlabelled features.
[2021-03-20 05:21:39] Showing labelled: True (3306/50000 visible, 94 redundant)
[2021-03-20 05:21:39] Creating trainer with model on device: cuda
[2021-03-20 05:21:39] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 05:22:33] Training accuracy: 0.8898
[2021-03-20 05:22:33] Testing on 10000 data points...
[2021-03-20 05:22:35] Test score for 3400 training labels: 0.8303
[2021-03-20 05:22:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:22:35] Found 50000 unlabelled features.
[2021-03-20 05:22:47] Showing labelled: True (3688/50000 visible, 112 redundant)
[2021-03-20 05:22:47] Creating trainer with model on device: cuda
[2021-03-20 05:22:47] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 05:23:48] Training accuracy: 0.9059
[2021-03-20 05:23:48] Testing on 10000 data points...
[2021-03-20 05:23:50] Test score for 3800 training labels: 0.8421
[2021-03-20 05:23:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:23:50] Found 50000 unlabelled features.
[2021-03-20 05:24:02] Showing labelled: True (4073/50000 visible, 127 redundant)
[2021-03-20 05:24:02] Creating trainer with model on device: cuda
[2021-03-20 05:24:02] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 05:25:09] Training accuracy: 0.8903
[2021-03-20 05:25:09] Testing on 10000 data points...
[2021-03-20 05:25:10] Test score for 4200 training labels: 0.8481
[2021-03-20 05:25:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:25:11] Found 50000 unlabelled features.
[2021-03-20 05:25:23] Showing labelled: True (4458/50000 visible, 142 redundant)
[2021-03-20 05:25:23] Creating trainer with model on device: cuda
[2021-03-20 05:25:23] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 05:26:35] Training accuracy: 0.9007
[2021-03-20 05:26:35] Testing on 10000 data points...
[2021-03-20 05:26:37] Test score for 4600 training labels: 0.8531
[2021-03-20 05:26:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:26:37] Found 50000 unlabelled features.
[2021-03-20 05:26:49] Showing labelled: True (4840/50000 visible, 160 redundant)
[2021-03-20 05:26:49] Creating trainer with model on device: cuda
[2021-03-20 05:26:49] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 05:28:09] Training accuracy: 0.9113
[2021-03-20 05:28:09] Testing on 10000 data points...
[2021-03-20 05:28:11] Test score for 5000 training labels: 0.8550
[2021-03-20 05:28:11] Experiment repeat 3/4
[2021-03-20 05:28:11] Seeded: 7
[2021-03-20 05:28:11] Using 2.00% labels of the dataset (1000/50000)
[2021-03-20 05:28:11] Randomly labelled 1000/50000
[2021-03-20 05:28:11] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 05:28:11] Seeded: 8
[2021-03-20 05:28:11] Running: experiment 0
[2021-03-20 05:28:11] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 05:28:11] Creating pretrained=True VGG16...
[2021-03-20 05:28:12] No parameter reset since we are using a pretrained model.
[2021-03-20 05:28:12] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 05:28:12] Creating trainer with model on device: cuda
[2021-03-20 05:28:12] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 05:29:16] Training accuracy: 0.8882
[2021-03-20 05:29:16] Testing on 10000 data points...
[2021-03-20 05:29:18] Test score for 1000 training labels: 0.7407
[2021-03-20 05:29:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:29:18] Found 50000 unlabelled features.
[2021-03-20 05:29:26] Showing labelled: True (1392/50000 visible, 8 redundant)
[2021-03-20 05:29:26] Creating trainer with model on device: cuda
[2021-03-20 05:29:26] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 05:29:48] Training accuracy: 0.8751
[2021-03-20 05:29:48] Testing on 10000 data points...
[2021-03-20 05:29:50] Test score for 1400 training labels: 0.7573
[2021-03-20 05:29:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:29:50] Found 50000 unlabelled features.
[2021-03-20 05:29:57] Showing labelled: True (1783/50000 visible, 17 redundant)
[2021-03-20 05:29:57] Creating trainer with model on device: cuda
[2021-03-20 05:29:57] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 05:30:26] Training accuracy: 0.8756
[2021-03-20 05:30:26] Testing on 10000 data points...
[2021-03-20 05:30:28] Test score for 1800 training labels: 0.7648
[2021-03-20 05:30:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:30:28] Found 50000 unlabelled features.
[2021-03-20 05:30:35] Showing labelled: True (2170/50000 visible, 30 redundant)
[2021-03-20 05:30:35] Creating trainer with model on device: cuda
[2021-03-20 05:30:35] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 05:31:11] Training accuracy: 0.8908
[2021-03-20 05:31:11] Testing on 10000 data points...
[2021-03-20 05:31:13] Test score for 2200 training labels: 0.7821
[2021-03-20 05:31:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:31:13] Found 50000 unlabelled features.
[2021-03-20 05:31:20] Showing labelled: True (2549/50000 visible, 51 redundant)
[2021-03-20 05:31:20] Creating trainer with model on device: cuda
[2021-03-20 05:31:20] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 05:32:01] Training accuracy: 0.8890
[2021-03-20 05:32:01] Testing on 10000 data points...
[2021-03-20 05:32:03] Test score for 2600 training labels: 0.7909
[2021-03-20 05:32:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:32:03] Found 50000 unlabelled features.
[2021-03-20 05:32:11] Showing labelled: True (2927/50000 visible, 73 redundant)
[2021-03-20 05:32:11] Creating trainer with model on device: cuda
[2021-03-20 05:32:11] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 05:32:58] Training accuracy: 0.9096
[2021-03-20 05:32:58] Testing on 10000 data points...
[2021-03-20 05:33:00] Test score for 3000 training labels: 0.8009
[2021-03-20 05:33:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:33:00] Found 50000 unlabelled features.
[2021-03-20 05:33:08] Showing labelled: True (3306/50000 visible, 94 redundant)
[2021-03-20 05:33:08] Creating trainer with model on device: cuda
[2021-03-20 05:33:08] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 05:34:02] Training accuracy: 0.9045
[2021-03-20 05:34:02] Testing on 10000 data points...
[2021-03-20 05:34:04] Test score for 3400 training labels: 0.8053
[2021-03-20 05:34:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:34:04] Found 50000 unlabelled features.
[2021-03-20 05:34:12] Showing labelled: True (3681/50000 visible, 119 redundant)
[2021-03-20 05:34:12] Creating trainer with model on device: cuda
[2021-03-20 05:34:12] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 05:35:13] Training accuracy: 0.9035
[2021-03-20 05:35:13] Testing on 10000 data points...
[2021-03-20 05:35:14] Test score for 3800 training labels: 0.8137
[2021-03-20 05:35:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:35:15] Found 50000 unlabelled features.
[2021-03-20 05:35:22] Showing labelled: True (4049/50000 visible, 151 redundant)
[2021-03-20 05:35:22] Creating trainer with model on device: cuda
[2021-03-20 05:35:22] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 05:36:29] Training accuracy: 0.9021
[2021-03-20 05:36:29] Testing on 10000 data points...
[2021-03-20 05:36:31] Test score for 4200 training labels: 0.8158
[2021-03-20 05:36:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:36:31] Found 50000 unlabelled features.
[2021-03-20 05:36:38] Showing labelled: True (4412/50000 visible, 188 redundant)
[2021-03-20 05:36:38] Creating trainer with model on device: cuda
[2021-03-20 05:36:38] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 05:37:51] Training accuracy: 0.9090
[2021-03-20 05:37:51] Testing on 10000 data points...
[2021-03-20 05:37:53] Test score for 4600 training labels: 0.8232
[2021-03-20 05:37:53] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:37:53] Found 50000 unlabelled features.
[2021-03-20 05:38:01] Showing labelled: True (4781/50000 visible, 219 redundant)
[2021-03-20 05:38:01] Creating trainer with model on device: cuda
[2021-03-20 05:38:01] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 05:39:20] Training accuracy: 0.9162
[2021-03-20 05:39:20] Testing on 10000 data points...
[2021-03-20 05:39:22] Test score for 5000 training labels: 0.8242
[2021-03-20 05:39:22] Seeded: 8
[2021-03-20 05:39:22] Running: experiment 1
[2021-03-20 05:39:22] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 05:39:22] Creating pretrained=True VGG16...
[2021-03-20 05:39:24] No parameter reset since we are using a pretrained model.
[2021-03-20 05:39:24] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 05:39:24] Creating trainer with model on device: cuda
[2021-03-20 05:39:24] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 05:40:28] Training accuracy: 0.8835
[2021-03-20 05:40:28] Testing on 10000 data points...
[2021-03-20 05:40:30] Test score for 1000 training labels: 0.7355
[2021-03-20 05:40:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:40:30] Found 50000 unlabelled features.
[2021-03-20 05:40:42] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:40:45] Searching for coresets with 20 beams...
[2021-03-20 05:41:14] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-20 05:41:14] Creating trainer with model on device: cuda
[2021-03-20 05:41:14] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 05:41:37] Training accuracy: 0.8751
[2021-03-20 05:41:37] Testing on 10000 data points...
[2021-03-20 05:41:38] Test score for 1400 training labels: 0.7505
[2021-03-20 05:41:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:41:39] Found 50000 unlabelled features.
[2021-03-20 05:41:51] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:41:55] Searching for coresets with 20 beams...
[2021-03-20 05:42:24] Showing labelled: True (1791/50000 visible, 9 redundant)
[2021-03-20 05:42:24] Creating trainer with model on device: cuda
[2021-03-20 05:42:24] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 05:42:53] Training accuracy: 0.8737
[2021-03-20 05:42:53] Testing on 10000 data points...
[2021-03-20 05:42:54] Test score for 1800 training labels: 0.7713
[2021-03-20 05:42:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:42:54] Found 50000 unlabelled features.
[2021-03-20 05:43:07] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:43:12] Searching for coresets with 20 beams...
[2021-03-20 05:43:41] Showing labelled: True (2175/50000 visible, 25 redundant)
[2021-03-20 05:43:41] Creating trainer with model on device: cuda
[2021-03-20 05:43:41] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 05:44:17] Training accuracy: 0.8815
[2021-03-20 05:44:17] Testing on 10000 data points...
[2021-03-20 05:44:19] Test score for 2200 training labels: 0.7909
[2021-03-20 05:44:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:44:19] Found 50000 unlabelled features.
[2021-03-20 05:44:31] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:44:38] Searching for coresets with 20 beams...
[2021-03-20 05:45:06] Showing labelled: True (2561/50000 visible, 39 redundant)
[2021-03-20 05:45:06] Creating trainer with model on device: cuda
[2021-03-20 05:45:06] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 05:45:48] Training accuracy: 0.8791
[2021-03-20 05:45:48] Testing on 10000 data points...
[2021-03-20 05:45:50] Test score for 2600 training labels: 0.8028
[2021-03-20 05:45:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:45:50] Found 50000 unlabelled features.
[2021-03-20 05:46:03] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:46:10] Searching for coresets with 20 beams...
[2021-03-20 05:46:38] Showing labelled: True (2946/50000 visible, 54 redundant)
[2021-03-20 05:46:38] Creating trainer with model on device: cuda
[2021-03-20 05:46:38] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 05:47:27] Training accuracy: 0.8926
[2021-03-20 05:47:27] Testing on 10000 data points...
[2021-03-20 05:47:28] Test score for 3000 training labels: 0.8125
[2021-03-20 05:47:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:47:28] Found 50000 unlabelled features.
[2021-03-20 05:47:41] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:47:49] Searching for coresets with 20 beams...
[2021-03-20 05:48:18] Showing labelled: True (3325/50000 visible, 75 redundant)
[2021-03-20 05:48:18] Creating trainer with model on device: cuda
[2021-03-20 05:48:18] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 05:49:12] Training accuracy: 0.8815
[2021-03-20 05:49:12] Testing on 10000 data points...
[2021-03-20 05:49:14] Test score for 3400 training labels: 0.8241
[2021-03-20 05:49:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:49:14] Found 50000 unlabelled features.
[2021-03-20 05:49:27] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:49:36] Searching for coresets with 20 beams...
[2021-03-20 05:50:05] Showing labelled: True (3695/50000 visible, 105 redundant)
[2021-03-20 05:50:05] Creating trainer with model on device: cuda
[2021-03-20 05:50:05] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 05:51:06] Training accuracy: 0.8822
[2021-03-20 05:51:06] Testing on 10000 data points...
[2021-03-20 05:51:08] Test score for 3800 training labels: 0.8320
[2021-03-20 05:51:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:51:08] Found 50000 unlabelled features.
[2021-03-20 05:51:21] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:51:32] Searching for coresets with 20 beams...
[2021-03-20 05:52:00] Showing labelled: True (4068/50000 visible, 132 redundant)
[2021-03-20 05:52:00] Creating trainer with model on device: cuda
[2021-03-20 05:52:00] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 05:53:07] Training accuracy: 0.8938
[2021-03-20 05:53:07] Testing on 10000 data points...
[2021-03-20 05:53:09] Test score for 4200 training labels: 0.8387
[2021-03-20 05:53:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:53:09] Found 50000 unlabelled features.
[2021-03-20 05:53:22] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:53:34] Searching for coresets with 20 beams...
[2021-03-20 05:54:03] Showing labelled: True (4442/50000 visible, 158 redundant)
[2021-03-20 05:54:03] Creating trainer with model on device: cuda
[2021-03-20 05:54:03] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 05:55:16] Training accuracy: 0.8899
[2021-03-20 05:55:16] Testing on 10000 data points...
[2021-03-20 05:55:18] Test score for 4600 training labels: 0.8423
[2021-03-20 05:55:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:55:18] Found 50000 unlabelled features.
[2021-03-20 05:55:32] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:55:44] Searching for coresets with 20 beams...
[2021-03-20 05:56:13] Showing labelled: True (4820/50000 visible, 180 redundant)
[2021-03-20 05:56:13] Creating trainer with model on device: cuda
[2021-03-20 05:56:13] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 05:57:33] Training accuracy: 0.8978
[2021-03-20 05:57:33] Testing on 10000 data points...
[2021-03-20 05:57:35] Test score for 5000 training labels: 0.8493
[2021-03-20 05:57:35] Seeded: 8
[2021-03-20 05:57:35] Running: experiment 2
[2021-03-20 05:57:35] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 05:57:35] Creating pretrained=True VGG16...
[2021-03-20 05:57:37] No parameter reset since we are using a pretrained model.
[2021-03-20 05:57:37] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 05:57:37] Creating trainer with model on device: cuda
[2021-03-20 05:57:37] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 05:58:41] Training accuracy: 0.8834
[2021-03-20 05:58:41] Testing on 10000 data points...
[2021-03-20 05:58:43] Test score for 1000 training labels: 0.7361
[2021-03-20 05:58:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:58:43] Found 50000 unlabelled features.
[2021-03-20 05:58:56] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 05:58:59] Searching for coresets with 20 beams...
[2021-03-20 05:59:27] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-20 05:59:27] Creating trainer with model on device: cuda
[2021-03-20 05:59:27] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 05:59:50] Training accuracy: 0.8757
[2021-03-20 05:59:50] Testing on 10000 data points...
[2021-03-20 05:59:52] Test score for 1400 training labels: 0.7490
[2021-03-20 05:59:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 05:59:52] Found 50000 unlabelled features.
[2021-03-20 06:00:04] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:00:08] Searching for coresets with 20 beams...
[2021-03-20 06:00:37] Showing labelled: True (1792/50000 visible, 8 redundant)
[2021-03-20 06:00:37] Creating trainer with model on device: cuda
[2021-03-20 06:00:37] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 06:01:06] Training accuracy: 0.8737
[2021-03-20 06:01:06] Testing on 10000 data points...
[2021-03-20 06:01:08] Test score for 1800 training labels: 0.7694
[2021-03-20 06:01:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:01:08] Found 50000 unlabelled features.
[2021-03-20 06:01:21] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:01:26] Searching for coresets with 20 beams...
[2021-03-20 06:01:54] Showing labelled: True (2178/50000 visible, 22 redundant)
[2021-03-20 06:01:54] Creating trainer with model on device: cuda
[2021-03-20 06:01:54] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 06:02:30] Training accuracy: 0.8752
[2021-03-20 06:02:30] Testing on 10000 data points...
[2021-03-20 06:02:32] Test score for 2200 training labels: 0.7871
[2021-03-20 06:02:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:02:32] Found 50000 unlabelled features.
[2021-03-20 06:02:45] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:02:51] Searching for coresets with 20 beams...
[2021-03-20 06:03:20] Showing labelled: True (2558/50000 visible, 42 redundant)
[2021-03-20 06:03:20] Creating trainer with model on device: cuda
[2021-03-20 06:03:20] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 06:04:02] Training accuracy: 0.8776
[2021-03-20 06:04:02] Testing on 10000 data points...
[2021-03-20 06:04:04] Test score for 2600 training labels: 0.8003
[2021-03-20 06:04:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:04:04] Found 50000 unlabelled features.
[2021-03-20 06:04:17] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:04:24] Searching for coresets with 20 beams...
[2021-03-20 06:04:52] Showing labelled: True (2942/50000 visible, 58 redundant)
[2021-03-20 06:04:52] Creating trainer with model on device: cuda
[2021-03-20 06:04:52] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 06:05:40] Training accuracy: 0.8937
[2021-03-20 06:05:40] Testing on 10000 data points...
[2021-03-20 06:05:42] Test score for 3000 training labels: 0.8131
[2021-03-20 06:05:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:05:42] Found 50000 unlabelled features.
[2021-03-20 06:05:55] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:06:04] Searching for coresets with 20 beams...
[2021-03-20 06:06:32] Showing labelled: True (3313/50000 visible, 87 redundant)
[2021-03-20 06:06:32] Creating trainer with model on device: cuda
[2021-03-20 06:06:32] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 06:07:27] Training accuracy: 0.8755
[2021-03-20 06:07:27] Testing on 10000 data points...
[2021-03-20 06:07:29] Test score for 3400 training labels: 0.8218
[2021-03-20 06:07:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:07:29] Found 50000 unlabelled features.
[2021-03-20 06:07:42] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:07:51] Searching for coresets with 20 beams...
[2021-03-20 06:08:20] Showing labelled: True (3683/50000 visible, 117 redundant)
[2021-03-20 06:08:20] Creating trainer with model on device: cuda
[2021-03-20 06:08:20] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 06:09:22] Training accuracy: 0.8845
[2021-03-20 06:09:22] Testing on 10000 data points...
[2021-03-20 06:09:24] Test score for 3800 training labels: 0.8280
[2021-03-20 06:09:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:09:24] Found 50000 unlabelled features.
[2021-03-20 06:09:37] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:09:47] Searching for coresets with 20 beams...
[2021-03-20 06:10:16] Showing labelled: True (4057/50000 visible, 143 redundant)
[2021-03-20 06:10:16] Creating trainer with model on device: cuda
[2021-03-20 06:10:16] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 06:11:23] Training accuracy: 0.8967
[2021-03-20 06:11:23] Testing on 10000 data points...
[2021-03-20 06:11:25] Test score for 4200 training labels: 0.8335
[2021-03-20 06:11:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:11:25] Found 50000 unlabelled features.
[2021-03-20 06:11:38] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:11:50] Searching for coresets with 20 beams...
[2021-03-20 06:12:19] Showing labelled: True (4427/50000 visible, 173 redundant)
[2021-03-20 06:12:19] Creating trainer with model on device: cuda
[2021-03-20 06:12:19] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 06:13:32] Training accuracy: 0.8899
[2021-03-20 06:13:32] Testing on 10000 data points...
[2021-03-20 06:13:34] Test score for 4600 training labels: 0.8395
[2021-03-20 06:13:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:13:34] Found 50000 unlabelled features.
[2021-03-20 06:13:47] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:14:00] Searching for coresets with 20 beams...
[2021-03-20 06:14:29] Showing labelled: True (4795/50000 visible, 205 redundant)
[2021-03-20 06:14:29] Creating trainer with model on device: cuda
[2021-03-20 06:14:29] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 06:15:49] Training accuracy: 0.8946
[2021-03-20 06:15:49] Testing on 10000 data points...
[2021-03-20 06:15:51] Test score for 5000 training labels: 0.8467
[2021-03-20 06:15:51] Seeded: 8
[2021-03-20 06:15:51] Running: experiment 3
[2021-03-20 06:15:51] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 06:15:51] Creating pretrained=True VGG16...
[2021-03-20 06:15:52] No parameter reset since we are using a pretrained model.
[2021-03-20 06:15:52] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 06:15:52] Creating trainer with model on device: cuda
[2021-03-20 06:15:52] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 06:16:56] Training accuracy: 0.8859
[2021-03-20 06:16:56] Testing on 10000 data points...
[2021-03-20 06:16:58] Test score for 1000 training labels: 0.7393
[2021-03-20 06:16:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:16:58] Found 50000 unlabelled features.
[2021-03-20 06:17:10] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:17:13] Searching for coresets greedily...
[2021-03-20 06:17:14] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 06:17:14] Creating trainer with model on device: cuda
[2021-03-20 06:17:14] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 06:17:36] Training accuracy: 0.8768
[2021-03-20 06:17:36] Testing on 10000 data points...
[2021-03-20 06:17:38] Test score for 1400 training labels: 0.7556
[2021-03-20 06:17:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:17:38] Found 50000 unlabelled features.
[2021-03-20 06:17:50] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:17:54] Searching for coresets greedily...
[2021-03-20 06:17:55] Showing labelled: True (1792/50000 visible, 8 redundant)
[2021-03-20 06:17:55] Creating trainer with model on device: cuda
[2021-03-20 06:17:55] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 06:18:24] Training accuracy: 0.8625
[2021-03-20 06:18:24] Testing on 10000 data points...
[2021-03-20 06:18:26] Test score for 1800 training labels: 0.7688
[2021-03-20 06:18:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:18:26] Found 50000 unlabelled features.
[2021-03-20 06:18:38] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:18:43] Searching for coresets greedily...
[2021-03-20 06:18:44] Showing labelled: True (2174/50000 visible, 26 redundant)
[2021-03-20 06:18:44] Creating trainer with model on device: cuda
[2021-03-20 06:18:44] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 06:19:19] Training accuracy: 0.8876
[2021-03-20 06:19:19] Testing on 10000 data points...
[2021-03-20 06:19:21] Test score for 2200 training labels: 0.7869
[2021-03-20 06:19:21] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:19:21] Found 50000 unlabelled features.
[2021-03-20 06:19:34] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:19:40] Searching for coresets greedily...
[2021-03-20 06:19:41] Showing labelled: True (2549/50000 visible, 51 redundant)
[2021-03-20 06:19:41] Creating trainer with model on device: cuda
[2021-03-20 06:19:41] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 06:20:22] Training accuracy: 0.8821
[2021-03-20 06:20:22] Testing on 10000 data points...
[2021-03-20 06:20:24] Test score for 2600 training labels: 0.7994
[2021-03-20 06:20:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:20:24] Found 50000 unlabelled features.
[2021-03-20 06:20:37] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:20:44] Searching for coresets greedily...
[2021-03-20 06:20:45] Showing labelled: True (2922/50000 visible, 78 redundant)
[2021-03-20 06:20:45] Creating trainer with model on device: cuda
[2021-03-20 06:20:45] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 06:21:32] Training accuracy: 0.8931
[2021-03-20 06:21:32] Testing on 10000 data points...
[2021-03-20 06:21:34] Test score for 3000 training labels: 0.8130
[2021-03-20 06:21:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:21:34] Found 50000 unlabelled features.
[2021-03-20 06:21:47] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:21:55] Searching for coresets greedily...
[2021-03-20 06:21:56] Showing labelled: True (3297/50000 visible, 103 redundant)
[2021-03-20 06:21:56] Creating trainer with model on device: cuda
[2021-03-20 06:21:56] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 06:22:50] Training accuracy: 0.8777
[2021-03-20 06:22:50] Testing on 10000 data points...
[2021-03-20 06:22:52] Test score for 3400 training labels: 0.8209
[2021-03-20 06:22:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:22:52] Found 50000 unlabelled features.
[2021-03-20 06:23:05] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:23:15] Searching for coresets greedily...
[2021-03-20 06:23:15] Showing labelled: True (3674/50000 visible, 126 redundant)
[2021-03-20 06:23:15] Creating trainer with model on device: cuda
[2021-03-20 06:23:15] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 06:24:16] Training accuracy: 0.8821
[2021-03-20 06:24:16] Testing on 10000 data points...
[2021-03-20 06:24:18] Test score for 3800 training labels: 0.8271
[2021-03-20 06:24:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:24:18] Found 50000 unlabelled features.
[2021-03-20 06:24:32] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:24:42] Searching for coresets greedily...
[2021-03-20 06:24:43] Showing labelled: True (4045/50000 visible, 155 redundant)
[2021-03-20 06:24:43] Creating trainer with model on device: cuda
[2021-03-20 06:24:43] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 06:25:49] Training accuracy: 0.8951
[2021-03-20 06:25:49] Testing on 10000 data points...
[2021-03-20 06:25:51] Test score for 4200 training labels: 0.8350
[2021-03-20 06:25:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:25:51] Found 50000 unlabelled features.
[2021-03-20 06:26:05] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:26:16] Searching for coresets greedily...
[2021-03-20 06:26:17] Showing labelled: True (4418/50000 visible, 182 redundant)
[2021-03-20 06:26:17] Creating trainer with model on device: cuda
[2021-03-20 06:26:17] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 06:27:30] Training accuracy: 0.8861
[2021-03-20 06:27:30] Testing on 10000 data points...
[2021-03-20 06:27:32] Test score for 4600 training labels: 0.8407
[2021-03-20 06:27:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:27:32] Found 50000 unlabelled features.
[2021-03-20 06:27:46] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:27:58] Searching for coresets greedily...
[2021-03-20 06:27:59] Showing labelled: True (4781/50000 visible, 219 redundant)
[2021-03-20 06:27:59] Creating trainer with model on device: cuda
[2021-03-20 06:27:59] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 06:29:19] Training accuracy: 0.8989
[2021-03-20 06:29:19] Testing on 10000 data points...
[2021-03-20 06:29:21] Test score for 5000 training labels: 0.8477
[2021-03-20 06:29:21] Seeded: 8
[2021-03-20 06:29:21] Running: experiment 4
[2021-03-20 06:29:21] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 06:29:21] Creating pretrained=True VGG16...
[2021-03-20 06:29:22] No parameter reset since we are using a pretrained model.
[2021-03-20 06:29:22] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 06:29:22] Creating trainer with model on device: cuda
[2021-03-20 06:29:22] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 06:30:26] Training accuracy: 0.8874
[2021-03-20 06:30:26] Testing on 10000 data points...
[2021-03-20 06:30:28] Test score for 1000 training labels: 0.7392
[2021-03-20 06:30:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:30:28] Found 50000 unlabelled features.
[2021-03-20 06:30:40] Showing labelled: True (1397/50000 visible, 3 redundant)
[2021-03-20 06:30:40] Creating trainer with model on device: cuda
[2021-03-20 06:30:40] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 06:31:02] Training accuracy: 0.8496
[2021-03-20 06:31:02] Testing on 10000 data points...
[2021-03-20 06:31:04] Test score for 1400 training labels: 0.7672
[2021-03-20 06:31:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:31:04] Found 50000 unlabelled features.
[2021-03-20 06:31:16] Showing labelled: True (1781/50000 visible, 19 redundant)
[2021-03-20 06:31:16] Creating trainer with model on device: cuda
[2021-03-20 06:31:16] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 06:31:44] Training accuracy: 0.8557
[2021-03-20 06:31:44] Testing on 10000 data points...
[2021-03-20 06:31:46] Test score for 1800 training labels: 0.7834
[2021-03-20 06:31:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:31:46] Found 50000 unlabelled features.
[2021-03-20 06:31:59] Showing labelled: True (2167/50000 visible, 33 redundant)
[2021-03-20 06:31:59] Creating trainer with model on device: cuda
[2021-03-20 06:31:59] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 06:32:34] Training accuracy: 0.8670
[2021-03-20 06:32:34] Testing on 10000 data points...
[2021-03-20 06:32:36] Test score for 2200 training labels: 0.7993
[2021-03-20 06:32:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:32:36] Found 50000 unlabelled features.
[2021-03-20 06:32:48] Showing labelled: True (2541/50000 visible, 59 redundant)
[2021-03-20 06:32:48] Creating trainer with model on device: cuda
[2021-03-20 06:32:48] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 06:33:29] Training accuracy: 0.8777
[2021-03-20 06:33:29] Testing on 10000 data points...
[2021-03-20 06:33:31] Test score for 2600 training labels: 0.8156
[2021-03-20 06:33:31] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:33:31] Found 50000 unlabelled features.
[2021-03-20 06:33:44] Showing labelled: True (2916/50000 visible, 84 redundant)
[2021-03-20 06:33:44] Creating trainer with model on device: cuda
[2021-03-20 06:33:44] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 06:34:31] Training accuracy: 0.8932
[2021-03-20 06:34:31] Testing on 10000 data points...
[2021-03-20 06:34:33] Test score for 3000 training labels: 0.8247
[2021-03-20 06:34:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:34:33] Found 50000 unlabelled features.
[2021-03-20 06:34:45] Showing labelled: True (3300/50000 visible, 100 redundant)
[2021-03-20 06:34:45] Creating trainer with model on device: cuda
[2021-03-20 06:34:45] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 06:35:39] Training accuracy: 0.8973
[2021-03-20 06:35:39] Testing on 10000 data points...
[2021-03-20 06:35:41] Test score for 3400 training labels: 0.8325
[2021-03-20 06:35:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:35:41] Found 50000 unlabelled features.
[2021-03-20 06:35:53] Showing labelled: True (3680/50000 visible, 120 redundant)
[2021-03-20 06:35:53] Creating trainer with model on device: cuda
[2021-03-20 06:35:53] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 06:36:54] Training accuracy: 0.8890
[2021-03-20 06:36:54] Testing on 10000 data points...
[2021-03-20 06:36:56] Test score for 3800 training labels: 0.8397
[2021-03-20 06:36:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:36:56] Found 50000 unlabelled features.
[2021-03-20 06:37:08] Showing labelled: True (4060/50000 visible, 140 redundant)
[2021-03-20 06:37:08] Creating trainer with model on device: cuda
[2021-03-20 06:37:08] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 06:38:16] Training accuracy: 0.8951
[2021-03-20 06:38:16] Testing on 10000 data points...
[2021-03-20 06:38:17] Test score for 4200 training labels: 0.8452
[2021-03-20 06:38:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:38:18] Found 50000 unlabelled features.
[2021-03-20 06:38:29] Showing labelled: True (4450/50000 visible, 150 redundant)
[2021-03-20 06:38:29] Creating trainer with model on device: cuda
[2021-03-20 06:38:29] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 06:39:42] Training accuracy: 0.9015
[2021-03-20 06:39:42] Testing on 10000 data points...
[2021-03-20 06:39:44] Test score for 4600 training labels: 0.8498
[2021-03-20 06:39:44] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:39:44] Found 50000 unlabelled features.
[2021-03-20 06:39:56] Showing labelled: True (4831/50000 visible, 169 redundant)
[2021-03-20 06:39:56] Creating trainer with model on device: cuda
[2021-03-20 06:39:56] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 06:41:16] Training accuracy: 0.9048
[2021-03-20 06:41:16] Testing on 10000 data points...
[2021-03-20 06:41:18] Test score for 5000 training labels: 0.8545
[2021-03-20 06:41:18] Experiment repeat 4/4
[2021-03-20 06:41:18] Seeded: 8
[2021-03-20 06:41:18] Using 2.00% labels of the dataset (1000/50000)
[2021-03-20 06:41:18] Randomly labelled 1000/50000
[2021-03-20 06:41:18] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 06:41:18] Seeded: 9
[2021-03-20 06:41:18] Running: experiment 0
[2021-03-20 06:41:18] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 06:41:18] Creating pretrained=True VGG16...
[2021-03-20 06:41:20] No parameter reset since we are using a pretrained model.
[2021-03-20 06:41:20] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 06:41:20] Creating trainer with model on device: cuda
[2021-03-20 06:41:20] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 06:42:23] Training accuracy: 0.8939
[2021-03-20 06:42:23] Testing on 10000 data points...
[2021-03-20 06:42:25] Test score for 1000 training labels: 0.7394
[2021-03-20 06:42:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:42:25] Found 50000 unlabelled features.
[2021-03-20 06:42:33] Showing labelled: True (1389/50000 visible, 11 redundant)
[2021-03-20 06:42:33] Creating trainer with model on device: cuda
[2021-03-20 06:42:33] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 06:42:55] Training accuracy: 0.8810
[2021-03-20 06:42:55] Testing on 10000 data points...
[2021-03-20 06:42:57] Test score for 1400 training labels: 0.7598
[2021-03-20 06:42:57] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:42:57] Found 50000 unlabelled features.
[2021-03-20 06:43:05] Showing labelled: True (1780/50000 visible, 20 redundant)
[2021-03-20 06:43:05] Creating trainer with model on device: cuda
[2021-03-20 06:43:05] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 06:43:33] Training accuracy: 0.8865
[2021-03-20 06:43:33] Testing on 10000 data points...
[2021-03-20 06:43:35] Test score for 1800 training labels: 0.7709
[2021-03-20 06:43:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:43:35] Found 50000 unlabelled features.
[2021-03-20 06:43:43] Showing labelled: True (2166/50000 visible, 34 redundant)
[2021-03-20 06:43:43] Creating trainer with model on device: cuda
[2021-03-20 06:43:43] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 06:44:18] Training accuracy: 0.8955
[2021-03-20 06:44:18] Testing on 10000 data points...
[2021-03-20 06:44:20] Test score for 2200 training labels: 0.7900
[2021-03-20 06:44:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:44:20] Found 50000 unlabelled features.
[2021-03-20 06:44:28] Showing labelled: True (2553/50000 visible, 47 redundant)
[2021-03-20 06:44:28] Creating trainer with model on device: cuda
[2021-03-20 06:44:28] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 06:45:09] Training accuracy: 0.8935
[2021-03-20 06:45:09] Testing on 10000 data points...
[2021-03-20 06:45:11] Test score for 2600 training labels: 0.7896
[2021-03-20 06:45:11] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:45:11] Found 50000 unlabelled features.
[2021-03-20 06:45:19] Showing labelled: True (2936/50000 visible, 64 redundant)
[2021-03-20 06:45:19] Creating trainer with model on device: cuda
[2021-03-20 06:45:19] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 06:46:07] Training accuracy: 0.9031
[2021-03-20 06:46:07] Testing on 10000 data points...
[2021-03-20 06:46:08] Test score for 3000 training labels: 0.8019
[2021-03-20 06:46:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:46:08] Found 50000 unlabelled features.
[2021-03-20 06:46:16] Showing labelled: True (3316/50000 visible, 84 redundant)
[2021-03-20 06:46:16] Creating trainer with model on device: cuda
[2021-03-20 06:46:16] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 06:47:10] Training accuracy: 0.9039
[2021-03-20 06:47:10] Testing on 10000 data points...
[2021-03-20 06:47:12] Test score for 3400 training labels: 0.8035
[2021-03-20 06:47:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:47:12] Found 50000 unlabelled features.
[2021-03-20 06:47:20] Showing labelled: True (3695/50000 visible, 105 redundant)
[2021-03-20 06:47:20] Creating trainer with model on device: cuda
[2021-03-20 06:47:20] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 06:48:20] Training accuracy: 0.9014
[2021-03-20 06:48:20] Testing on 10000 data points...
[2021-03-20 06:48:22] Test score for 3800 training labels: 0.8100
[2021-03-20 06:48:22] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:48:22] Found 50000 unlabelled features.
[2021-03-20 06:48:30] Showing labelled: True (4064/50000 visible, 136 redundant)
[2021-03-20 06:48:30] Creating trainer with model on device: cuda
[2021-03-20 06:48:30] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 06:49:37] Training accuracy: 0.9092
[2021-03-20 06:49:37] Testing on 10000 data points...
[2021-03-20 06:49:39] Test score for 4200 training labels: 0.8149
[2021-03-20 06:49:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:49:39] Found 50000 unlabelled features.
[2021-03-20 06:49:46] Showing labelled: True (4436/50000 visible, 164 redundant)
[2021-03-20 06:49:46] Creating trainer with model on device: cuda
[2021-03-20 06:49:46] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 06:51:00] Training accuracy: 0.9068
[2021-03-20 06:51:00] Testing on 10000 data points...
[2021-03-20 06:51:01] Test score for 4600 training labels: 0.8207
[2021-03-20 06:51:02] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:51:02] Found 50000 unlabelled features.
[2021-03-20 06:51:09] Showing labelled: True (4804/50000 visible, 196 redundant)
[2021-03-20 06:51:09] Creating trainer with model on device: cuda
[2021-03-20 06:51:09] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 06:52:29] Training accuracy: 0.9072
[2021-03-20 06:52:29] Testing on 10000 data points...
[2021-03-20 06:52:31] Test score for 5000 training labels: 0.8259
[2021-03-20 06:52:31] Seeded: 9
[2021-03-20 06:52:31] Running: experiment 1
[2021-03-20 06:52:31] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 06:52:31] Creating pretrained=True VGG16...
[2021-03-20 06:52:33] No parameter reset since we are using a pretrained model.
[2021-03-20 06:52:33] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 06:52:33] Creating trainer with model on device: cuda
[2021-03-20 06:52:33] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 06:53:37] Training accuracy: 0.8953
[2021-03-20 06:53:37] Testing on 10000 data points...
[2021-03-20 06:53:39] Test score for 1000 training labels: 0.7398
[2021-03-20 06:53:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:53:39] Found 50000 unlabelled features.
[2021-03-20 06:53:51] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:53:54] Searching for coresets with 20 beams...
[2021-03-20 06:54:23] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 06:54:23] Creating trainer with model on device: cuda
[2021-03-20 06:54:23] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 06:54:45] Training accuracy: 0.8889
[2021-03-20 06:54:45] Testing on 10000 data points...
[2021-03-20 06:54:47] Test score for 1400 training labels: 0.7618
[2021-03-20 06:54:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:54:47] Found 50000 unlabelled features.
[2021-03-20 06:55:00] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:55:04] Searching for coresets with 20 beams...
[2021-03-20 06:55:32] Showing labelled: True (1787/50000 visible, 13 redundant)
[2021-03-20 06:55:32] Creating trainer with model on device: cuda
[2021-03-20 06:55:32] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 06:56:01] Training accuracy: 0.8733
[2021-03-20 06:56:01] Testing on 10000 data points...
[2021-03-20 06:56:03] Test score for 1800 training labels: 0.7747
[2021-03-20 06:56:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:56:03] Found 50000 unlabelled features.
[2021-03-20 06:56:16] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:56:21] Searching for coresets with 20 beams...
[2021-03-20 06:56:49] Showing labelled: True (2164/50000 visible, 36 redundant)
[2021-03-20 06:56:49] Creating trainer with model on device: cuda
[2021-03-20 06:56:49] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 06:57:25] Training accuracy: 0.8872
[2021-03-20 06:57:25] Testing on 10000 data points...
[2021-03-20 06:57:27] Test score for 2200 training labels: 0.7885
[2021-03-20 06:57:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:57:27] Found 50000 unlabelled features.
[2021-03-20 06:57:40] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:57:46] Searching for coresets with 20 beams...
[2021-03-20 06:58:15] Showing labelled: True (2542/50000 visible, 58 redundant)
[2021-03-20 06:58:15] Creating trainer with model on device: cuda
[2021-03-20 06:58:15] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 06:58:57] Training accuracy: 0.8876
[2021-03-20 06:58:57] Testing on 10000 data points...
[2021-03-20 06:58:59] Test score for 2600 training labels: 0.8014
[2021-03-20 06:58:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 06:58:59] Found 50000 unlabelled features.
[2021-03-20 06:59:11] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 06:59:19] Searching for coresets with 20 beams...
[2021-03-20 06:59:47] Showing labelled: True (2917/50000 visible, 83 redundant)
[2021-03-20 06:59:47] Creating trainer with model on device: cuda
[2021-03-20 06:59:47] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 07:00:36] Training accuracy: 0.9009
[2021-03-20 07:00:36] Testing on 10000 data points...
[2021-03-20 07:00:38] Test score for 3000 training labels: 0.8063
[2021-03-20 07:00:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:00:38] Found 50000 unlabelled features.
[2021-03-20 07:00:50] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:00:59] Searching for coresets with 20 beams...
[2021-03-20 07:01:27] Showing labelled: True (3296/50000 visible, 104 redundant)
[2021-03-20 07:01:27] Creating trainer with model on device: cuda
[2021-03-20 07:01:27] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 07:02:22] Training accuracy: 0.9031
[2021-03-20 07:02:22] Testing on 10000 data points...
[2021-03-20 07:02:24] Test score for 3400 training labels: 0.8147
[2021-03-20 07:02:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:02:24] Found 50000 unlabelled features.
[2021-03-20 07:02:36] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:02:46] Searching for coresets with 20 beams...
[2021-03-20 07:03:15] Showing labelled: True (3663/50000 visible, 137 redundant)
[2021-03-20 07:03:15] Creating trainer with model on device: cuda
[2021-03-20 07:03:15] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 07:04:16] Training accuracy: 0.8977
[2021-03-20 07:04:16] Testing on 10000 data points...
[2021-03-20 07:04:18] Test score for 3800 training labels: 0.8242
[2021-03-20 07:04:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:04:18] Found 50000 unlabelled features.
[2021-03-20 07:04:31] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:04:42] Searching for coresets with 20 beams...
[2021-03-20 07:05:10] Showing labelled: True (4034/50000 visible, 166 redundant)
[2021-03-20 07:05:10] Creating trainer with model on device: cuda
[2021-03-20 07:05:10] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 07:06:17] Training accuracy: 0.9001
[2021-03-20 07:06:17] Testing on 10000 data points...
[2021-03-20 07:06:19] Test score for 4200 training labels: 0.8294
[2021-03-20 07:06:19] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:06:19] Found 50000 unlabelled features.
[2021-03-20 07:06:32] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:06:44] Searching for coresets with 20 beams...
[2021-03-20 07:07:12] Showing labelled: True (4413/50000 visible, 187 redundant)
[2021-03-20 07:07:12] Creating trainer with model on device: cuda
[2021-03-20 07:07:12] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 07:08:26] Training accuracy: 0.9030
[2021-03-20 07:08:26] Testing on 10000 data points...
[2021-03-20 07:08:28] Test score for 4600 training labels: 0.8371
[2021-03-20 07:08:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:08:28] Found 50000 unlabelled features.
[2021-03-20 07:08:41] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:08:54] Searching for coresets with 20 beams...
[2021-03-20 07:09:23] Showing labelled: True (4798/50000 visible, 202 redundant)
[2021-03-20 07:09:23] Creating trainer with model on device: cuda
[2021-03-20 07:09:23] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 07:10:42] Training accuracy: 0.8880
[2021-03-20 07:10:42] Testing on 10000 data points...
[2021-03-20 07:10:44] Test score for 5000 training labels: 0.8452
[2021-03-20 07:10:44] Seeded: 9
[2021-03-20 07:10:44] Running: experiment 2
[2021-03-20 07:10:44] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 07:10:44] Creating pretrained=True VGG16...
[2021-03-20 07:10:46] No parameter reset since we are using a pretrained model.
[2021-03-20 07:10:46] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 07:10:46] Creating trainer with model on device: cuda
[2021-03-20 07:10:46] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 07:11:50] Training accuracy: 0.8960
[2021-03-20 07:11:50] Testing on 10000 data points...
[2021-03-20 07:11:52] Test score for 1000 training labels: 0.7401
[2021-03-20 07:11:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:11:52] Found 50000 unlabelled features.
[2021-03-20 07:12:04] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:12:07] Searching for coresets with 20 beams...
[2021-03-20 07:12:36] Showing labelled: True (1399/50000 visible, 1 redundant)
[2021-03-20 07:12:36] Creating trainer with model on device: cuda
[2021-03-20 07:12:36] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 07:12:58] Training accuracy: 0.8857
[2021-03-20 07:12:58] Testing on 10000 data points...
[2021-03-20 07:13:00] Test score for 1400 training labels: 0.7648
[2021-03-20 07:13:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:13:00] Found 50000 unlabelled features.
[2021-03-20 07:13:13] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:13:17] Searching for coresets with 20 beams...
[2021-03-20 07:13:45] Showing labelled: True (1792/50000 visible, 8 redundant)
[2021-03-20 07:13:45] Creating trainer with model on device: cuda
[2021-03-20 07:13:45] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 07:14:14] Training accuracy: 0.8713
[2021-03-20 07:14:14] Testing on 10000 data points...
[2021-03-20 07:14:16] Test score for 1800 training labels: 0.7758
[2021-03-20 07:14:16] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:14:16] Found 50000 unlabelled features.
[2021-03-20 07:14:29] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:14:34] Searching for coresets with 20 beams...
[2021-03-20 07:15:03] Showing labelled: True (2185/50000 visible, 15 redundant)
[2021-03-20 07:15:03] Creating trainer with model on device: cuda
[2021-03-20 07:15:03] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 07:15:39] Training accuracy: 0.8876
[2021-03-20 07:15:39] Testing on 10000 data points...
[2021-03-20 07:15:41] Test score for 2200 training labels: 0.7887
[2021-03-20 07:15:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:15:41] Found 50000 unlabelled features.
[2021-03-20 07:15:53] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:15:59] Searching for coresets with 20 beams...
[2021-03-20 07:16:28] Showing labelled: True (2564/50000 visible, 36 redundant)
[2021-03-20 07:16:28] Creating trainer with model on device: cuda
[2021-03-20 07:16:28] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 07:17:11] Training accuracy: 0.8909
[2021-03-20 07:17:11] Testing on 10000 data points...
[2021-03-20 07:17:13] Test score for 2600 training labels: 0.7998
[2021-03-20 07:17:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:17:13] Found 50000 unlabelled features.
[2021-03-20 07:17:26] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:17:34] Searching for coresets with 20 beams...
[2021-03-20 07:18:02] Showing labelled: True (2935/50000 visible, 65 redundant)
[2021-03-20 07:18:02] Creating trainer with model on device: cuda
[2021-03-20 07:18:02] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 07:18:50] Training accuracy: 0.8935
[2021-03-20 07:18:50] Testing on 10000 data points...
[2021-03-20 07:18:52] Test score for 3000 training labels: 0.8098
[2021-03-20 07:18:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:18:52] Found 50000 unlabelled features.
[2021-03-20 07:19:05] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:19:14] Searching for coresets with 20 beams...
[2021-03-20 07:19:42] Showing labelled: True (3320/50000 visible, 80 redundant)
[2021-03-20 07:19:42] Creating trainer with model on device: cuda
[2021-03-20 07:19:42] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 07:20:37] Training accuracy: 0.8961
[2021-03-20 07:20:37] Testing on 10000 data points...
[2021-03-20 07:20:39] Test score for 3400 training labels: 0.8180
[2021-03-20 07:20:39] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:20:39] Found 50000 unlabelled features.
[2021-03-20 07:20:51] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:21:01] Searching for coresets with 20 beams...
[2021-03-20 07:21:30] Showing labelled: True (3699/50000 visible, 101 redundant)
[2021-03-20 07:21:30] Creating trainer with model on device: cuda
[2021-03-20 07:21:30] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 07:22:31] Training accuracy: 0.8924
[2021-03-20 07:22:31] Testing on 10000 data points...
[2021-03-20 07:22:33] Test score for 3800 training labels: 0.8254
[2021-03-20 07:22:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:22:33] Found 50000 unlabelled features.
[2021-03-20 07:22:46] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:22:56] Searching for coresets with 20 beams...
[2021-03-20 07:23:25] Showing labelled: True (4079/50000 visible, 121 redundant)
[2021-03-20 07:23:25] Creating trainer with model on device: cuda
[2021-03-20 07:23:25] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 07:24:32] Training accuracy: 0.8997
[2021-03-20 07:24:32] Testing on 10000 data points...
[2021-03-20 07:24:34] Test score for 4200 training labels: 0.8352
[2021-03-20 07:24:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:24:34] Found 50000 unlabelled features.
[2021-03-20 07:24:48] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:24:59] Searching for coresets with 20 beams...
[2021-03-20 07:25:28] Showing labelled: True (4453/50000 visible, 147 redundant)
[2021-03-20 07:25:28] Creating trainer with model on device: cuda
[2021-03-20 07:25:28] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 07:26:41] Training accuracy: 0.9033
[2021-03-20 07:26:41] Testing on 10000 data points...
[2021-03-20 07:26:43] Test score for 4600 training labels: 0.8414
[2021-03-20 07:26:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:26:43] Found 50000 unlabelled features.
[2021-03-20 07:26:56] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:27:09] Searching for coresets with 20 beams...
[2021-03-20 07:27:38] Showing labelled: True (4826/50000 visible, 174 redundant)
[2021-03-20 07:27:38] Creating trainer with model on device: cuda
[2021-03-20 07:27:38] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 07:28:57] Training accuracy: 0.8904
[2021-03-20 07:28:57] Testing on 10000 data points...
[2021-03-20 07:28:59] Test score for 5000 training labels: 0.8403
[2021-03-20 07:28:59] Seeded: 9
[2021-03-20 07:28:59] Running: experiment 3
[2021-03-20 07:28:59] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 07:28:59] Creating pretrained=True VGG16...
[2021-03-20 07:29:01] No parameter reset since we are using a pretrained model.
[2021-03-20 07:29:01] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 07:29:01] Creating trainer with model on device: cuda
[2021-03-20 07:29:01] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 07:30:05] Training accuracy: 0.8954
[2021-03-20 07:30:05] Testing on 10000 data points...
[2021-03-20 07:30:06] Test score for 1000 training labels: 0.7435
[2021-03-20 07:30:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:30:07] Found 50000 unlabelled features.
[2021-03-20 07:30:19] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:30:22] Searching for coresets greedily...
[2021-03-20 07:30:23] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-20 07:30:23] Creating trainer with model on device: cuda
[2021-03-20 07:30:23] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 07:30:45] Training accuracy: 0.8827
[2021-03-20 07:30:45] Testing on 10000 data points...
[2021-03-20 07:30:47] Test score for 1400 training labels: 0.7651
[2021-03-20 07:30:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:30:47] Found 50000 unlabelled features.
[2021-03-20 07:30:59] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:31:03] Searching for coresets greedily...
[2021-03-20 07:31:04] Showing labelled: True (1787/50000 visible, 13 redundant)
[2021-03-20 07:31:04] Creating trainer with model on device: cuda
[2021-03-20 07:31:04] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 07:31:32] Training accuracy: 0.8757
[2021-03-20 07:31:32] Testing on 10000 data points...
[2021-03-20 07:31:34] Test score for 1800 training labels: 0.7791
[2021-03-20 07:31:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:31:34] Found 50000 unlabelled features.
[2021-03-20 07:31:47] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:31:52] Searching for coresets greedily...
[2021-03-20 07:31:53] Showing labelled: True (2170/50000 visible, 30 redundant)
[2021-03-20 07:31:53] Creating trainer with model on device: cuda
[2021-03-20 07:31:53] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 07:32:28] Training accuracy: 0.8803
[2021-03-20 07:32:28] Testing on 10000 data points...
[2021-03-20 07:32:30] Test score for 2200 training labels: 0.7867
[2021-03-20 07:32:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:32:30] Found 50000 unlabelled features.
[2021-03-20 07:32:43] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:32:49] Searching for coresets greedily...
[2021-03-20 07:32:50] Showing labelled: True (2553/50000 visible, 47 redundant)
[2021-03-20 07:32:50] Creating trainer with model on device: cuda
[2021-03-20 07:32:50] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 07:33:31] Training accuracy: 0.8851
[2021-03-20 07:33:31] Testing on 10000 data points...
[2021-03-20 07:33:33] Test score for 2600 training labels: 0.8016
[2021-03-20 07:33:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:33:33] Found 50000 unlabelled features.
[2021-03-20 07:33:46] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:33:53] Searching for coresets greedily...
[2021-03-20 07:33:54] Showing labelled: True (2923/50000 visible, 77 redundant)
[2021-03-20 07:33:54] Creating trainer with model on device: cuda
[2021-03-20 07:33:54] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 07:34:41] Training accuracy: 0.8995
[2021-03-20 07:34:41] Testing on 10000 data points...
[2021-03-20 07:34:43] Test score for 3000 training labels: 0.8084
[2021-03-20 07:34:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:34:43] Found 50000 unlabelled features.
[2021-03-20 07:34:56] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:35:04] Searching for coresets greedily...
[2021-03-20 07:35:05] Showing labelled: True (3298/50000 visible, 102 redundant)
[2021-03-20 07:35:05] Creating trainer with model on device: cuda
[2021-03-20 07:35:05] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 07:35:59] Training accuracy: 0.9089
[2021-03-20 07:35:59] Testing on 10000 data points...
[2021-03-20 07:36:01] Test score for 3400 training labels: 0.8173
[2021-03-20 07:36:01] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:36:01] Found 50000 unlabelled features.
[2021-03-20 07:36:14] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:36:23] Searching for coresets greedily...
[2021-03-20 07:36:24] Showing labelled: True (3669/50000 visible, 131 redundant)
[2021-03-20 07:36:24] Creating trainer with model on device: cuda
[2021-03-20 07:36:24] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 07:37:25] Training accuracy: 0.8962
[2021-03-20 07:37:25] Testing on 10000 data points...
[2021-03-20 07:37:27] Test score for 3800 training labels: 0.8267
[2021-03-20 07:37:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:37:27] Found 50000 unlabelled features.
[2021-03-20 07:37:40] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:37:50] Searching for coresets greedily...
[2021-03-20 07:37:51] Showing labelled: True (4038/50000 visible, 162 redundant)
[2021-03-20 07:37:51] Creating trainer with model on device: cuda
[2021-03-20 07:37:51] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 07:38:58] Training accuracy: 0.8974
[2021-03-20 07:38:58] Testing on 10000 data points...
[2021-03-20 07:39:00] Test score for 4200 training labels: 0.8280
[2021-03-20 07:39:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:39:00] Found 50000 unlabelled features.
[2021-03-20 07:39:13] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:39:24] Searching for coresets greedily...
[2021-03-20 07:39:25] Showing labelled: True (4407/50000 visible, 193 redundant)
[2021-03-20 07:39:25] Creating trainer with model on device: cuda
[2021-03-20 07:39:25] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 07:40:38] Training accuracy: 0.9040
[2021-03-20 07:40:38] Testing on 10000 data points...
[2021-03-20 07:40:40] Test score for 4600 training labels: 0.8407
[2021-03-20 07:40:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:40:40] Found 50000 unlabelled features.
[2021-03-20 07:40:54] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 07:41:06] Searching for coresets greedily...
[2021-03-20 07:41:07] Showing labelled: True (4780/50000 visible, 220 redundant)
[2021-03-20 07:41:07] Creating trainer with model on device: cuda
[2021-03-20 07:41:07] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 07:42:27] Training accuracy: 0.8947
[2021-03-20 07:42:27] Testing on 10000 data points...
[2021-03-20 07:42:28] Test score for 5000 training labels: 0.8454
[2021-03-20 07:42:28] Seeded: 9
[2021-03-20 07:42:28] Running: experiment 4
[2021-03-20 07:42:29] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 07:42:29] Creating pretrained=True VGG16...
[2021-03-20 07:42:30] No parameter reset since we are using a pretrained model.
[2021-03-20 07:42:30] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 07:42:30] Creating trainer with model on device: cuda
[2021-03-20 07:42:30] Training vgg-pretrained across 1000 data points in cifar10-augmented...
[2021-03-20 07:43:34] Training accuracy: 0.8945
[2021-03-20 07:43:34] Testing on 10000 data points...
[2021-03-20 07:43:36] Test score for 1000 training labels: 0.7425
[2021-03-20 07:43:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:43:36] Found 50000 unlabelled features.
[2021-03-20 07:43:48] Showing labelled: True (1398/50000 visible, 2 redundant)
[2021-03-20 07:43:48] Creating trainer with model on device: cuda
[2021-03-20 07:43:48] Training vgg-pretrained across 1400 data points in cifar10-augmented...
[2021-03-20 07:44:10] Training accuracy: 0.8546
[2021-03-20 07:44:10] Testing on 10000 data points...
[2021-03-20 07:44:12] Test score for 1400 training labels: 0.7649
[2021-03-20 07:44:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:44:12] Found 50000 unlabelled features.
[2021-03-20 07:44:24] Showing labelled: True (1777/50000 visible, 23 redundant)
[2021-03-20 07:44:24] Creating trainer with model on device: cuda
[2021-03-20 07:44:24] Training vgg-pretrained across 1800 data points in cifar10-augmented...
[2021-03-20 07:44:52] Training accuracy: 0.8562
[2021-03-20 07:44:52] Testing on 10000 data points...
[2021-03-20 07:44:54] Test score for 1800 training labels: 0.7813
[2021-03-20 07:44:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:44:54] Found 50000 unlabelled features.
[2021-03-20 07:45:06] Showing labelled: True (2160/50000 visible, 40 redundant)
[2021-03-20 07:45:06] Creating trainer with model on device: cuda
[2021-03-20 07:45:06] Training vgg-pretrained across 2200 data points in cifar10-augmented...
[2021-03-20 07:45:42] Training accuracy: 0.8755
[2021-03-20 07:45:42] Testing on 10000 data points...
[2021-03-20 07:45:43] Test score for 2200 training labels: 0.8006
[2021-03-20 07:45:43] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:45:43] Found 50000 unlabelled features.
[2021-03-20 07:45:55] Showing labelled: True (2538/50000 visible, 62 redundant)
[2021-03-20 07:45:55] Creating trainer with model on device: cuda
[2021-03-20 07:45:55] Training vgg-pretrained across 2600 data points in cifar10-augmented...
[2021-03-20 07:46:36] Training accuracy: 0.8820
[2021-03-20 07:46:36] Testing on 10000 data points...
[2021-03-20 07:46:38] Test score for 2600 training labels: 0.8119
[2021-03-20 07:46:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:46:38] Found 50000 unlabelled features.
[2021-03-20 07:46:50] Showing labelled: True (2921/50000 visible, 79 redundant)
[2021-03-20 07:46:50] Creating trainer with model on device: cuda
[2021-03-20 07:46:50] Training vgg-pretrained across 3000 data points in cifar10-augmented...
[2021-03-20 07:47:38] Training accuracy: 0.8928
[2021-03-20 07:47:38] Testing on 10000 data points...
[2021-03-20 07:47:40] Test score for 3000 training labels: 0.8208
[2021-03-20 07:47:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:47:40] Found 50000 unlabelled features.
[2021-03-20 07:47:52] Showing labelled: True (3305/50000 visible, 95 redundant)
[2021-03-20 07:47:52] Creating trainer with model on device: cuda
[2021-03-20 07:47:52] Training vgg-pretrained across 3400 data points in cifar10-augmented...
[2021-03-20 07:48:46] Training accuracy: 0.8982
[2021-03-20 07:48:46] Testing on 10000 data points...
[2021-03-20 07:48:48] Test score for 3400 training labels: 0.8250
[2021-03-20 07:48:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:48:48] Found 50000 unlabelled features.
[2021-03-20 07:49:00] Showing labelled: True (3677/50000 visible, 123 redundant)
[2021-03-20 07:49:00] Creating trainer with model on device: cuda
[2021-03-20 07:49:00] Training vgg-pretrained across 3800 data points in cifar10-augmented...
[2021-03-20 07:50:01] Training accuracy: 0.8897
[2021-03-20 07:50:01] Testing on 10000 data points...
[2021-03-20 07:50:03] Test score for 3800 training labels: 0.8399
[2021-03-20 07:50:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:50:03] Found 50000 unlabelled features.
[2021-03-20 07:50:15] Showing labelled: True (4057/50000 visible, 143 redundant)
[2021-03-20 07:50:15] Creating trainer with model on device: cuda
[2021-03-20 07:50:15] Training vgg-pretrained across 4200 data points in cifar10-augmented...
[2021-03-20 07:51:21] Training accuracy: 0.9046
[2021-03-20 07:51:21] Testing on 10000 data points...
[2021-03-20 07:51:23] Test score for 4200 training labels: 0.8446
[2021-03-20 07:51:23] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:51:23] Found 50000 unlabelled features.
[2021-03-20 07:51:35] Showing labelled: True (4445/50000 visible, 155 redundant)
[2021-03-20 07:51:35] Creating trainer with model on device: cuda
[2021-03-20 07:51:35] Training vgg-pretrained across 4600 data points in cifar10-augmented...
[2021-03-20 07:52:48] Training accuracy: 0.8968
[2021-03-20 07:52:48] Testing on 10000 data points...
[2021-03-20 07:52:50] Test score for 4600 training labels: 0.8437
[2021-03-20 07:52:50] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 07:52:50] Found 50000 unlabelled features.
[2021-03-20 07:53:02] Showing labelled: True (4827/50000 visible, 173 redundant)
[2021-03-20 07:53:02] Creating trainer with model on device: cuda
[2021-03-20 07:53:02] Training vgg-pretrained across 5000 data points in cifar10-augmented...
[2021-03-20 07:54:21] Training accuracy: 0.8985
[2021-03-20 07:54:21] Testing on 10000 data points...
[2021-03-20 07:54:23] Test score for 5000 training labels: 0.8518
[2021-03-20 07:54:23] Updated results: ../results/cifar10-aug/vgg_pretrained_top2entropy/results.json
