[2021-03-18 00:14:59] Created experiment 0:
[2021-03-18 00:14:59]  - Model: vgg
[2021-03-18 00:14:59]  - Acquisition function: random
[2021-03-18 00:14:59] Created experiment 1:
[2021-03-18 00:14:59]  - Model: vgg
[2021-03-18 00:14:59]  - Acquisition function: lc-beam-coreset
[2021-03-18 00:14:59] Created experiment 2:
[2021-03-18 00:14:59]  - Model: vgg
[2021-03-18 00:14:59]  - Acquisition function: greedy-coreset
[2021-03-18 00:14:59] Created experiment 3:
[2021-03-18 00:14:59]  - Model: vgg
[2021-03-18 00:14:59]  - Acquisition function: least-confidence
[2021-03-18 00:14:59] Loading cifar10 test set...
[2021-03-18 00:15:00] Experiment repeat 1/1
[2021-03-18 00:15:00] Seeded: 5
[2021-03-18 00:15:00] Using 10.00% labels of the dataset (5000/50000)
[2021-03-18 00:15:00] Randomly labelled 5000/50000
[2021-03-18 00:15:00] Showing labelled: True (5000/50000 visible)
[2021-03-18 00:15:00] Seeded: 5
[2021-03-18 00:15:00] Running: experiment 0
[2021-03-18 00:15:00] Showing labelled: True (5000/50000 visible)
[2021-03-18 00:15:03] vgg: initialized 16299850 parameters.
[2021-03-18 00:15:03] Creating trainer with model on device: cuda
[2021-03-18 00:15:08] Training vgg across 5000 data points in cifar10...
[2021-03-18 00:15:11] Training accuracy: 0.0987
[2021-03-18 00:15:11] Testing on 10000 data points...
[2021-03-18 00:15:13] Test score for 5000 training labels: 0.1495
[2021-03-18 00:15:13] Showing labelled: False (45000/50000 visible)
[2021-03-18 00:15:13] Found 45000 unlabelled features.
[2021-03-18 00:15:20] Showing labelled: True (10000/50000 visible)
[2021-03-18 00:15:20] Training vgg across 10000 data points in cifar10...
[2021-03-18 00:15:27] Training accuracy: 0.1026
[2021-03-18 00:15:27] Testing on 10000 data points...
[2021-03-18 00:15:29] Test score for 10000 training labels: 0.1291
[2021-03-18 00:15:29] Showing labelled: False (40000/50000 visible)
[2021-03-18 00:15:29] Found 40000 unlabelled features.
[2021-03-18 00:15:34] Showing labelled: True (15000/50000 visible)
[2021-03-18 00:15:34] Training vgg across 15000 data points in cifar10...
[2021-03-18 00:15:45] Training accuracy: 0.1246
[2021-03-18 00:15:45] Testing on 10000 data points...
[2021-03-18 00:15:47] Test score for 15000 training labels: 0.1078
[2021-03-18 00:15:47] Seeded: 5
[2021-03-18 00:15:47] Running: experiment 1
[2021-03-18 00:15:47] Showing labelled: True (5000/50000 visible)
[2021-03-18 00:15:50] vgg: initialized 16299850 parameters.
[2021-03-18 00:15:50] Creating trainer with model on device: cuda
[2021-03-18 00:15:50] Training vgg across 5000 data points in cifar10...
[2021-03-18 00:15:54] Training accuracy: 0.0981
[2021-03-18 00:15:54] Testing on 10000 data points...
[2021-03-18 00:15:56] Test score for 5000 training labels: 0.1484
[2021-03-18 00:15:56] Showing labelled: False (45000/50000 visible)
[2021-03-18 00:15:56] Found 45000 unlabelled features.
[2021-03-18 00:16:07] Computing distance between 5000 labelled and 45000 unlabelled vectors of length 512...
[2021-03-18 00:16:15] Searching for coresets with 3 beams...
[2021-03-18 00:17:02] Showing labelled: True (10000/50000 visible)
[2021-03-18 00:17:02] Training vgg across 10000 data points in cifar10...
[2021-03-18 00:17:09] Training accuracy: 0.1319
[2021-03-18 00:17:09] Testing on 10000 data points...
[2021-03-18 00:17:11] Test score for 10000 training labels: 0.1000
[2021-03-18 00:17:11] Showing labelled: False (40000/50000 visible)
[2021-03-18 00:17:11] Found 40000 unlabelled features.
[2021-03-18 00:17:23] Computing distance between 10000 labelled and 40000 unlabelled vectors of length 512...
[2021-03-18 00:17:38] Searching for coresets with 3 beams...
[2021-03-18 00:18:19] Showing labelled: True (15000/50000 visible)
[2021-03-18 00:18:19] Training vgg across 15000 data points in cifar10...
[2021-03-18 00:18:31] Training accuracy: 0.1392
[2021-03-18 00:18:31] Testing on 10000 data points...
[2021-03-18 00:18:33] Test score for 15000 training labels: 0.1153
[2021-03-18 00:18:33] Seeded: 5
[2021-03-18 00:18:33] Running: experiment 2
[2021-03-18 00:18:33] Showing labelled: True (5000/50000 visible)
[2021-03-18 00:18:35] vgg: initialized 16299850 parameters.
[2021-03-18 00:18:35] Creating trainer with model on device: cuda
[2021-03-18 00:18:35] Training vgg across 5000 data points in cifar10...
[2021-03-18 00:18:39] Training accuracy: 0.0982
[2021-03-18 00:18:39] Testing on 10000 data points...
[2021-03-18 00:18:41] Test score for 5000 training labels: 0.1492
[2021-03-18 00:18:41] Showing labelled: False (45000/50000 visible)
[2021-03-18 00:18:41] Found 45000 unlabelled features.
[2021-03-18 00:18:51] Computing distance between 5000 labelled and 45000 unlabelled vectors of length 512...
[2021-03-18 00:19:00] Searching for coresets greedily...
[2021-03-18 00:19:08] Showing labelled: True (10000/50000 visible)
[2021-03-18 00:19:08] Training vgg across 10000 data points in cifar10...
[2021-03-18 00:19:16] Training accuracy: 0.1396
[2021-03-18 00:19:16] Testing on 10000 data points...
[2021-03-18 00:19:18] Test score for 10000 training labels: 0.1000
[2021-03-18 00:19:18] Showing labelled: False (40000/50000 visible)
[2021-03-18 00:19:18] Found 40000 unlabelled features.
[2021-03-18 00:19:29] Computing distance between 10000 labelled and 40000 unlabelled vectors of length 512...
[2021-03-18 00:19:44] Searching for coresets greedily...
[2021-03-18 00:19:52] Showing labelled: True (15000/50000 visible)
[2021-03-18 00:19:52] Training vgg across 15000 data points in cifar10...
[2021-03-18 00:20:04] Training accuracy: 0.1427
[2021-03-18 00:20:04] Testing on 10000 data points...
[2021-03-18 00:20:06] Test score for 15000 training labels: 0.1000
[2021-03-18 00:20:06] Seeded: 5
[2021-03-18 00:20:06] Running: experiment 3
[2021-03-18 00:20:06] Showing labelled: True (5000/50000 visible)
[2021-03-18 00:20:09] vgg: initialized 16299850 parameters.
[2021-03-18 00:20:09] Creating trainer with model on device: cuda
[2021-03-18 00:20:09] Training vgg across 5000 data points in cifar10...
[2021-03-18 00:20:12] Training accuracy: 0.0980
[2021-03-18 00:20:12] Testing on 10000 data points...
[2021-03-18 00:20:15] Test score for 5000 training labels: 0.1486
[2021-03-18 00:20:15] Showing labelled: False (45000/50000 visible)
[2021-03-18 00:20:15] Found 45000 unlabelled features.
[2021-03-18 00:20:25] Showing labelled: True (10000/50000 visible)
[2021-03-18 00:20:25] Training vgg across 10000 data points in cifar10...
[2021-03-18 00:20:32] Training accuracy: 0.1650
[2021-03-18 00:20:32] Testing on 10000 data points...
[2021-03-18 00:20:34] Test score for 10000 training labels: 0.1033
[2021-03-18 00:20:34] Showing labelled: False (40000/50000 visible)
[2021-03-18 00:20:34] Found 40000 unlabelled features.
[2021-03-18 00:20:43] Showing labelled: True (15000/50000 visible)
[2021-03-18 00:20:43] Training vgg across 15000 data points in cifar10...
[2021-03-18 00:20:55] Training accuracy: 0.1792
[2021-03-18 00:20:55] Testing on 10000 data points...
[2021-03-18 00:20:57] Test score for 15000 training labels: 0.1000
[2021-03-18 00:20:57] Updated results: ../results/cifar10/vgg_test/results.json
