[2021-03-18 14:20:31] Created experiment 0:
[2021-03-18 14:20:31]  - Model: vgg-pretrained
[2021-03-18 14:20:31]  - Acquisition function: random
[2021-03-18 14:20:31] Created experiment 1:
[2021-03-18 14:20:31]  - Model: vgg-pretrained
[2021-03-18 14:20:31]  - Acquisition function: least-confidence
[2021-03-18 14:20:31] Created experiment 2:
[2021-03-18 14:20:31]  - Model: vgg-pretrained
[2021-03-18 14:20:31]  - Acquisition function: lc-beam-coreset
[2021-03-18 14:20:31] Created experiment 3:
[2021-03-18 14:20:31]  - Model: vgg-pretrained
[2021-03-18 14:20:31]  - Acquisition function: greedy-coreset
[2021-03-18 14:20:31] Loading cifar10 test set...
[2021-03-18 14:20:32] Experiment repeat 1/1
[2021-03-18 14:20:32] Seeded: 5
[2021-03-18 14:20:32] Using 10.00% labels of the dataset (5000/50000)
[2021-03-18 14:20:32] Randomly labelled 5000/50000
[2021-03-18 14:20:32] Showing labelled: True (5000/50000 visible)
[2021-03-18 14:20:32] Seeded: 5
[2021-03-18 14:20:32] Running: experiment 0
[2021-03-18 14:20:32] Showing labelled: True (5000/50000 visible)
[2021-03-18 14:20:32] Creating pretrained=True VGG16...
[2021-03-18 14:20:34] No parameter reset since we are using a pretrained model.
[2021-03-18 14:20:34] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:20:34] Creating trainer with model on device: cuda
[2021-03-18 14:20:39] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-18 14:21:23] Training accuracy: 0.7989
[2021-03-18 14:21:23] Testing on 10000 data points...
[2021-03-18 14:21:25] Test score for 5000 training labels: 0.7261
[2021-03-18 14:21:25] Showing labelled: False (45000/50000 visible)
[2021-03-18 14:21:25] Found 45000 unlabelled features.
[2021-03-18 14:21:31] Showing labelled: True (10000/50000 visible)
[2021-03-18 14:21:31] Creating pretrained=True VGG16...
[2021-03-18 14:21:32] No parameter reset since we are using a pretrained model.
[2021-03-18 14:21:32] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:21:32] Creating trainer with model on device: cuda
[2021-03-18 14:21:32] Training vgg-pretrained across 10000 data points in cifar10...
[2021-03-18 14:23:02] Training accuracy: 0.8322
[2021-03-18 14:23:02] Testing on 10000 data points...
[2021-03-18 14:23:04] Test score for 10000 training labels: 0.7588
[2021-03-18 14:23:04] Showing labelled: False (40000/50000 visible)
[2021-03-18 14:23:04] Found 40000 unlabelled features.
[2021-03-18 14:23:09] Showing labelled: True (15000/50000 visible)
[2021-03-18 14:23:09] Creating pretrained=True VGG16...
[2021-03-18 14:23:11] No parameter reset since we are using a pretrained model.
[2021-03-18 14:23:11] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:23:11] Creating trainer with model on device: cuda
[2021-03-18 14:23:11] Training vgg-pretrained across 15000 data points in cifar10...
[2021-03-18 14:25:26] Training accuracy: 0.8458
[2021-03-18 14:25:26] Testing on 10000 data points...
[2021-03-18 14:25:28] Test score for 15000 training labels: 0.7842
[2021-03-18 14:25:28] Showing labelled: False (35000/50000 visible)
[2021-03-18 14:25:28] Found 35000 unlabelled features.
[2021-03-18 14:25:33] Showing labelled: True (20000/50000 visible)
[2021-03-18 14:25:33] Creating pretrained=True VGG16...
[2021-03-18 14:25:34] No parameter reset since we are using a pretrained model.
[2021-03-18 14:25:34] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:25:34] Creating trainer with model on device: cuda
[2021-03-18 14:25:34] Training vgg-pretrained across 20000 data points in cifar10...
[2021-03-18 14:28:36] Training accuracy: 0.8671
[2021-03-18 14:28:36] Testing on 10000 data points...
[2021-03-18 14:28:37] Test score for 20000 training labels: 0.7985
[2021-03-18 14:28:37] Seeded: 5
[2021-03-18 14:28:37] Running: experiment 1
[2021-03-18 14:28:37] Showing labelled: True (5000/50000 visible)
[2021-03-18 14:28:37] Creating pretrained=True VGG16...
[2021-03-18 14:28:39] No parameter reset since we are using a pretrained model.
[2021-03-18 14:28:39] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:28:39] Creating trainer with model on device: cuda
[2021-03-18 14:28:39] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-18 14:29:24] Training accuracy: 0.8004
[2021-03-18 14:29:24] Testing on 10000 data points...
[2021-03-18 14:29:26] Test score for 5000 training labels: 0.7264
[2021-03-18 14:29:26] Showing labelled: False (45000/50000 visible)
[2021-03-18 14:29:26] Found 45000 unlabelled features.
[2021-03-18 14:29:35] Showing labelled: True (10000/50000 visible)
[2021-03-18 14:29:35] Creating pretrained=True VGG16...
[2021-03-18 14:29:37] No parameter reset since we are using a pretrained model.
[2021-03-18 14:29:37] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:29:37] Creating trainer with model on device: cuda
[2021-03-18 14:29:37] Training vgg-pretrained across 10000 data points in cifar10...
[2021-03-18 14:31:08] Training accuracy: 0.7507
[2021-03-18 14:31:08] Testing on 10000 data points...
[2021-03-18 14:31:10] Test score for 10000 training labels: 0.7663
[2021-03-18 14:31:10] Showing labelled: False (40000/50000 visible)
[2021-03-18 14:31:10] Found 40000 unlabelled features.
[2021-03-18 14:31:18] Showing labelled: True (15000/50000 visible)
[2021-03-18 14:31:18] Creating pretrained=True VGG16...
[2021-03-18 14:31:19] No parameter reset since we are using a pretrained model.
[2021-03-18 14:31:19] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:31:19] Creating trainer with model on device: cuda
[2021-03-18 14:31:19] Training vgg-pretrained across 15000 data points in cifar10...
[2021-03-18 14:33:36] Training accuracy: 0.7585
[2021-03-18 14:33:36] Testing on 10000 data points...
[2021-03-18 14:33:38] Test score for 15000 training labels: 0.7997
[2021-03-18 14:33:38] Showing labelled: False (35000/50000 visible)
[2021-03-18 14:33:38] Found 35000 unlabelled features.
[2021-03-18 14:33:45] Showing labelled: True (20000/50000 visible)
[2021-03-18 14:33:45] Creating pretrained=True VGG16...
[2021-03-18 14:33:46] No parameter reset since we are using a pretrained model.
[2021-03-18 14:33:46] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:33:46] Creating trainer with model on device: cuda
[2021-03-18 14:33:46] Training vgg-pretrained across 20000 data points in cifar10...
[2021-03-18 14:36:48] Training accuracy: 0.7889
[2021-03-18 14:36:48] Testing on 10000 data points...
[2021-03-18 14:36:50] Test score for 20000 training labels: 0.8158
[2021-03-18 14:36:50] Seeded: 5
[2021-03-18 14:36:50] Running: experiment 2
[2021-03-18 14:36:50] Showing labelled: True (5000/50000 visible)
[2021-03-18 14:36:50] Creating pretrained=True VGG16...
[2021-03-18 14:36:52] No parameter reset since we are using a pretrained model.
[2021-03-18 14:36:52] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:36:52] Creating trainer with model on device: cuda
[2021-03-18 14:36:52] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-18 14:37:37] Training accuracy: 0.8005
[2021-03-18 14:37:37] Testing on 10000 data points...
[2021-03-18 14:37:39] Test score for 5000 training labels: 0.7264
[2021-03-18 14:37:39] Showing labelled: False (45000/50000 visible)
[2021-03-18 14:37:39] Found 45000 unlabelled features.
[2021-03-18 14:37:49] Computing distance between 5000 labelled and 45000 unlabelled vectors of length 512...
[2021-03-18 14:38:01] Searching for coresets with 10 beams...
[2021-03-18 14:40:27] Showing labelled: True (10000/50000 visible)
[2021-03-18 14:40:27] Creating pretrained=True VGG16...
[2021-03-18 14:40:29] No parameter reset since we are using a pretrained model.
[2021-03-18 14:40:29] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:40:29] Creating trainer with model on device: cuda
[2021-03-18 14:40:29] Training vgg-pretrained across 10000 data points in cifar10...
[2021-03-18 14:42:00] Training accuracy: 0.8270
[2021-03-18 14:42:00] Testing on 10000 data points...
[2021-03-18 14:42:02] Test score for 10000 training labels: 0.7583
[2021-03-18 14:42:02] Showing labelled: False (40000/50000 visible)
[2021-03-18 14:42:02] Found 40000 unlabelled features.
[2021-03-18 14:42:12] Computing distance between 10000 labelled and 40000 unlabelled vectors of length 512...
[2021-03-18 14:42:34] Searching for coresets with 10 beams...
[2021-03-18 14:44:44] Showing labelled: True (15000/50000 visible)
[2021-03-18 14:44:44] Creating pretrained=True VGG16...
[2021-03-18 14:44:45] No parameter reset since we are using a pretrained model.
[2021-03-18 14:44:45] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:44:45] Creating trainer with model on device: cuda
[2021-03-18 14:44:45] Training vgg-pretrained across 15000 data points in cifar10...
[2021-03-18 14:47:02] Training accuracy: 0.8295
[2021-03-18 14:47:02] Testing on 10000 data points...
[2021-03-18 14:47:04] Test score for 15000 training labels: 0.7871
[2021-03-18 14:47:04] Showing labelled: False (35000/50000 visible)
[2021-03-18 14:47:04] Found 35000 unlabelled features.
[2021-03-18 14:47:14] Computing distance between 15000 labelled and 35000 unlabelled vectors of length 512...
[2021-03-18 14:47:43] Searching for coresets with 10 beams...
[2021-03-18 14:49:35] Showing labelled: True (20000/50000 visible)
[2021-03-18 14:49:35] Creating pretrained=True VGG16...
[2021-03-18 14:49:37] No parameter reset since we are using a pretrained model.
[2021-03-18 14:49:37] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:49:37] Creating trainer with model on device: cuda
[2021-03-18 14:49:37] Training vgg-pretrained across 20000 data points in cifar10...
[2021-03-18 14:52:38] Training accuracy: 0.8502
[2021-03-18 14:52:38] Testing on 10000 data points...
[2021-03-18 14:52:40] Test score for 20000 training labels: 0.7979
[2021-03-18 14:52:40] Seeded: 5
[2021-03-18 14:52:40] Running: experiment 3
[2021-03-18 14:52:40] Showing labelled: True (5000/50000 visible)
[2021-03-18 14:52:40] Creating pretrained=True VGG16...
[2021-03-18 14:52:42] No parameter reset since we are using a pretrained model.
[2021-03-18 14:52:42] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:52:42] Creating trainer with model on device: cuda
[2021-03-18 14:52:42] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-18 14:53:27] Training accuracy: 0.8011
[2021-03-18 14:53:27] Testing on 10000 data points...
[2021-03-18 14:53:29] Test score for 5000 training labels: 0.7268
[2021-03-18 14:53:29] Showing labelled: False (45000/50000 visible)
[2021-03-18 14:53:29] Found 45000 unlabelled features.
[2021-03-18 14:53:39] Computing distance between 5000 labelled and 45000 unlabelled vectors of length 512...
[2021-03-18 14:53:51] Searching for coresets greedily...
[2021-03-18 14:54:00] Showing labelled: True (10000/50000 visible)
[2021-03-18 14:54:00] Creating pretrained=True VGG16...
[2021-03-18 14:54:02] No parameter reset since we are using a pretrained model.
[2021-03-18 14:54:02] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:54:02] Creating trainer with model on device: cuda
[2021-03-18 14:54:02] Training vgg-pretrained across 10000 data points in cifar10...
[2021-03-18 14:55:33] Training accuracy: 0.8197
[2021-03-18 14:55:33] Testing on 10000 data points...
[2021-03-18 14:55:35] Test score for 10000 training labels: 0.7590
[2021-03-18 14:55:35] Showing labelled: False (40000/50000 visible)
[2021-03-18 14:55:35] Found 40000 unlabelled features.
[2021-03-18 14:55:45] Computing distance between 10000 labelled and 40000 unlabelled vectors of length 512...
[2021-03-18 14:56:07] Searching for coresets greedily...
[2021-03-18 14:56:15] Showing labelled: True (15000/50000 visible)
[2021-03-18 14:56:15] Creating pretrained=True VGG16...
[2021-03-18 14:56:16] No parameter reset since we are using a pretrained model.
[2021-03-18 14:56:16] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:56:16] Creating trainer with model on device: cuda
[2021-03-18 14:56:16] Training vgg-pretrained across 15000 data points in cifar10...
[2021-03-18 14:58:33] Training accuracy: 0.8365
[2021-03-18 14:58:33] Testing on 10000 data points...
[2021-03-18 14:58:35] Test score for 15000 training labels: 0.7857
[2021-03-18 14:58:35] Showing labelled: False (35000/50000 visible)
[2021-03-18 14:58:35] Found 35000 unlabelled features.
[2021-03-18 14:58:45] Computing distance between 15000 labelled and 35000 unlabelled vectors of length 512...
[2021-03-18 14:59:14] Searching for coresets greedily...
[2021-03-18 14:59:20] Showing labelled: True (20000/50000 visible)
[2021-03-18 14:59:20] Creating pretrained=True VGG16...
[2021-03-18 14:59:22] No parameter reset since we are using a pretrained model.
[2021-03-18 14:59:22] vgg-pretrained: initialized 15250250 parameters.
[2021-03-18 14:59:22] Creating trainer with model on device: cuda
[2021-03-18 14:59:22] Training vgg-pretrained across 20000 data points in cifar10...
[2021-03-18 15:02:24] Training accuracy: 0.8261
[2021-03-18 15:02:24] Testing on 10000 data points...
[2021-03-18 15:02:26] Test score for 20000 training labels: 0.7946
[2021-03-18 15:02:26] Updated results: ../results/cifar10/vgg_pretrained_lsmooth/results.json
