[2021-03-20 21:46:17] Created experiment 0:
[2021-03-20 21:46:17]  - Model: vgg-pretrained
[2021-03-20 21:46:17]  - Acquisition function: lc-beam-pweighted-relconf-coreset
[2021-03-20 21:46:17] Loading cifar10 test set...
[2021-03-20 21:46:18] Experiment repeat 1/1
[2021-03-20 21:46:18] Seeded: 5
[2021-03-20 21:46:18] Using 2.00% labels of the dataset (1000/50000)
[2021-03-20 21:46:18] Randomly labelled 1000/50000
[2021-03-20 21:46:18] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 21:46:18] Seeded: 6
[2021-03-20 21:46:18] Running: experiment 0
[2021-03-20 21:46:18] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-20 21:46:18] Creating pretrained=True VGG16...
[2021-03-20 21:46:20] No parameter reset since we are using a pretrained model.
[2021-03-20 21:46:20] vgg-pretrained: initialized 15245130 parameters.
[2021-03-20 21:46:20] Creating trainer with model on device: cuda
[2021-03-20 21:46:25] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-20 21:47:24] Training accuracy: 0.9151
[2021-03-20 21:47:24] Testing on 10000 data points...
[2021-03-20 21:47:25] Test score for 1000 training labels: 0.7192
[2021-03-20 21:47:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:47:25] Found 50000 unlabelled features.
[2021-03-20 21:47:35] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:47:38] Searching for coresets with 20 beams...
[2021-03-20 21:48:07] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-20 21:48:07] Creating trainer with model on device: cuda
[2021-03-20 21:48:07] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-20 21:48:23] Training accuracy: 0.8968
[2021-03-20 21:48:23] Testing on 10000 data points...
[2021-03-20 21:48:25] Test score for 1400 training labels: 0.7410
[2021-03-20 21:48:25] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:48:25] Found 50000 unlabelled features.
[2021-03-20 21:48:35] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:48:39] Searching for coresets with 20 beams...
[2021-03-20 21:49:07] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-20 21:49:07] Creating trainer with model on device: cuda
[2021-03-20 21:49:07] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-20 21:49:28] Training accuracy: 0.8907
[2021-03-20 21:49:28] Testing on 10000 data points...
[2021-03-20 21:49:30] Test score for 1800 training labels: 0.7555
[2021-03-20 21:49:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:49:30] Found 50000 unlabelled features.
[2021-03-20 21:49:40] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:49:45] Searching for coresets with 20 beams...
[2021-03-20 21:50:14] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-20 21:50:14] Creating trainer with model on device: cuda
[2021-03-20 21:50:14] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-20 21:50:39] Training accuracy: 0.9108
[2021-03-20 21:50:39] Testing on 10000 data points...
[2021-03-20 21:50:41] Test score for 2200 training labels: 0.7806
[2021-03-20 21:50:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:50:41] Found 50000 unlabelled features.
[2021-03-20 21:50:51] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:50:58] Searching for coresets with 20 beams...
[2021-03-20 21:51:27] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-20 21:51:27] Creating trainer with model on device: cuda
[2021-03-20 21:51:27] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-20 21:51:56] Training accuracy: 0.9133
[2021-03-20 21:51:56] Testing on 10000 data points...
[2021-03-20 21:51:58] Test score for 2600 training labels: 0.7906
[2021-03-20 21:51:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:51:58] Found 50000 unlabelled features.
[2021-03-20 21:52:08] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:52:16] Searching for coresets with 20 beams...
[2021-03-20 21:52:45] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-20 21:52:45] Creating trainer with model on device: cuda
[2021-03-20 21:52:45] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-20 21:53:18] Training accuracy: 0.9106
[2021-03-20 21:53:18] Testing on 10000 data points...
[2021-03-20 21:53:20] Test score for 3000 training labels: 0.8004
[2021-03-20 21:53:20] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:53:20] Found 50000 unlabelled features.
[2021-03-20 21:53:31] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:53:39] Searching for coresets with 20 beams...
[2021-03-20 21:54:08] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-20 21:54:08] Creating trainer with model on device: cuda
[2021-03-20 21:54:08] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-20 21:54:47] Training accuracy: 0.9187
[2021-03-20 21:54:47] Testing on 10000 data points...
[2021-03-20 21:54:49] Test score for 3400 training labels: 0.8034
[2021-03-20 21:54:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-20 21:54:49] Found 50000 unlabelled features.
[2021-03-20 21:54:59] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-20 21:55:09] Searching for coresets with 20 beams...
[2021-03-20 21:55:38] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-20 21:55:38] Creating trainer with model on device: cuda
[2021-03-20 21:55:38] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-20 21:56:05] Updated results: ../results/cifar10/vgg_pretrained_relconf_lsmooth/results.json
