[2021-03-21 14:19:18] Created experiment 0:
[2021-03-21 14:19:18]  - Model: vgg-pretrained
[2021-03-21 14:19:18]  - Acquisition function: lc-beam-pweighted-relconf-coreset
[2021-03-21 14:19:18] Created experiment 1:
[2021-03-21 14:19:18]  - Model: vgg-pretrained
[2021-03-21 14:19:18]  - Acquisition function: greedy-coreset
[2021-03-21 14:19:18] Loading cifar10 test set...
[2021-03-21 14:19:19] Experiment repeat 1/2
[2021-03-21 14:19:19] Seeded: 5
[2021-03-21 14:19:19] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 14:19:19] Randomly labelled 1000/50000
[2021-03-21 14:19:19] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 14:19:19] Seeded: 6
[2021-03-21 14:19:19] Running: experiment 0
[2021-03-21 14:19:19] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 14:19:19] Creating pretrained=True VGG16...
[2021-03-21 14:19:21] No parameter reset since we are using a pretrained model.
[2021-03-21 14:19:21] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 14:19:21] Creating trainer with model on device: cuda
[2021-03-21 14:19:26] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 14:20:02] Training accuracy: 0.9995
[2021-03-21 14:20:02] Testing on 10000 data points...
[2021-03-21 14:20:04] Test score for 1000 training labels: 0.7209
[2021-03-21 14:20:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:20:04] Found 50000 unlabelled features.
[2021-03-21 14:20:14] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:20:17] Searching for coresets with 20 beams...
[2021-03-21 14:20:46] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 14:20:46] Creating trainer with model on device: cuda
[2021-03-21 14:20:46] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 14:20:54] Training accuracy: 0.9535
[2021-03-21 14:20:54] Testing on 10000 data points...
[2021-03-21 14:20:56] Test score for 1400 training labels: 0.7443
[2021-03-21 14:20:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:20:56] Found 50000 unlabelled features.
[2021-03-21 14:21:06] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:21:10] Searching for coresets with 20 beams...
[2021-03-21 14:21:39] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 14:21:39] Creating trainer with model on device: cuda
[2021-03-21 14:21:39] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 14:21:50] Training accuracy: 0.9858
[2021-03-21 14:21:50] Testing on 10000 data points...
[2021-03-21 14:21:51] Test score for 1800 training labels: 0.7675
[2021-03-21 14:21:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:21:51] Found 50000 unlabelled features.
[2021-03-21 14:22:01] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:22:07] Searching for coresets with 20 beams...
[2021-03-21 14:22:36] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 14:22:36] Creating trainer with model on device: cuda
[2021-03-21 14:22:36] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 14:22:49] Training accuracy: 0.9891
[2021-03-21 14:22:49] Testing on 10000 data points...
[2021-03-21 14:22:51] Test score for 2200 training labels: 0.7821
[2021-03-21 14:22:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:22:51] Found 50000 unlabelled features.
[2021-03-21 14:23:01] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:23:07] Searching for coresets with 20 beams...
[2021-03-21 14:23:36] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 14:23:36] Creating trainer with model on device: cuda
[2021-03-21 14:23:36] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 14:23:52] Training accuracy: 0.9912
[2021-03-21 14:23:52] Testing on 10000 data points...
[2021-03-21 14:23:54] Test score for 2600 training labels: 0.7932
[2021-03-21 14:23:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:23:54] Found 50000 unlabelled features.
[2021-03-21 14:24:04] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:24:11] Searching for coresets with 20 beams...
[2021-03-21 14:24:40] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 14:24:40] Creating trainer with model on device: cuda
[2021-03-21 14:24:40] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 14:24:58] Training accuracy: 0.9941
[2021-03-21 14:24:58] Testing on 10000 data points...
[2021-03-21 14:25:00] Test score for 3000 training labels: 0.8022
[2021-03-21 14:25:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:25:00] Found 50000 unlabelled features.
[2021-03-21 14:25:10] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:25:19] Searching for coresets with 20 beams...
[2021-03-21 14:25:48] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 14:25:48] Creating trainer with model on device: cuda
[2021-03-21 14:25:48] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 14:26:08] Training accuracy: 0.9862
[2021-03-21 14:26:08] Testing on 10000 data points...
[2021-03-21 14:26:10] Test score for 3400 training labels: 0.8150
[2021-03-21 14:26:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:26:10] Found 50000 unlabelled features.
[2021-03-21 14:26:21] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:26:30] Searching for coresets with 20 beams...
[2021-03-21 14:26:59] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 14:26:59] Creating trainer with model on device: cuda
[2021-03-21 14:26:59] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 14:27:22] Training accuracy: 0.9981
[2021-03-21 14:27:22] Testing on 10000 data points...
[2021-03-21 14:27:24] Test score for 3800 training labels: 0.8191
[2021-03-21 14:27:24] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:27:24] Found 50000 unlabelled features.
[2021-03-21 14:27:35] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:27:46] Searching for coresets with 20 beams...
[2021-03-21 14:28:15] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 14:28:15] Creating trainer with model on device: cuda
[2021-03-21 14:28:15] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 14:28:40] Training accuracy: 0.9982
[2021-03-21 14:28:40] Testing on 10000 data points...
[2021-03-21 14:28:42] Test score for 4200 training labels: 0.8230
[2021-03-21 14:28:42] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:28:42] Found 50000 unlabelled features.
[2021-03-21 14:28:52] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:29:04] Searching for coresets with 20 beams...
[2021-03-21 14:29:33] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 14:29:33] Creating trainer with model on device: cuda
[2021-03-21 14:29:33] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 14:30:01] Training accuracy: 0.9978
[2021-03-21 14:30:01] Testing on 10000 data points...
[2021-03-21 14:30:03] Test score for 4600 training labels: 0.8300
[2021-03-21 14:30:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 14:30:03] Found 50000 unlabelled features.
[2021-03-21 14:30:13] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 14:30:26] Searching for coresets with 20 beams...
[2021-03-21 14:30:55] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 14:30:55] Creating trainer with model on device: cuda
[2021-03-21 14:30:55] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 14:31:26] Training accuracy: 0.9992
[2021-03-21 14:31:26] Testing on 10000 data points...
[2021-03-21 14:31:27] Test score for 5000 training labels: 0.8355
[2021-03-21 14:31:27] Seeded: 6
[2021-03-21 14:31:27] Running: experiment 1
[2021-03-21 14:31:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 14:31:28] Creating pretrained=True VGG16...
[2021-03-21 14:31:29] No parameter reset since we are using a pretrained model.
[2021-03-21 14:31:29] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 14:31:29] Creating trainer with model on device: cuda
[2021-03-21 14:31:29] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 14:31:44] Updated results: ../results/cifar10/main_notovertrained/results.json
