[2021-03-19 22:40:27] Created experiment 0:
[2021-03-19 22:40:27]  - Model: vgg-pretrained
[2021-03-19 22:40:27]  - Acquisition function: greedy-coreset
[2021-03-19 22:40:27] Created experiment 1:
[2021-03-19 22:40:27]  - Model: vgg-pretrained
[2021-03-19 22:40:27]  - Acquisition function: random
[2021-03-19 22:40:27] Created experiment 2:
[2021-03-19 22:40:27]  - Model: vgg-pretrained
[2021-03-19 22:40:27]  - Acquisition function: lc-beam-coreset
[2021-03-19 22:40:27] Loading cifar10 test set...
[2021-03-19 22:40:28] Experiment repeat 1/1
[2021-03-19 22:40:28] Seeded: 5
[2021-03-19 22:40:28] Using 2.00% labels of the dataset (1000/50000)
[2021-03-19 22:40:28] Randomly labelled 1000/50000
[2021-03-19 22:40:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-19 22:40:28] Seeded: 6
[2021-03-19 22:40:28] Running: experiment 0
[2021-03-19 22:40:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-19 22:40:28] Creating pretrained=True VGG16...
[2021-03-19 22:40:30] No parameter reset since we are using a pretrained model.
[2021-03-19 22:40:30] vgg-pretrained: initialized 16299850 parameters.
[2021-03-19 22:40:30] Creating trainer with model on device: cuda
[2021-03-19 22:40:35] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-19 22:40:57] Training accuracy: 0.9967
[2021-03-19 22:40:57] Testing on 10000 data points...
[2021-03-19 22:40:59] Test score for 1000 training labels: 0.7196
[2021-03-19 22:40:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:40:59] Found 50000 unlabelled features.
[2021-03-19 22:41:10] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:41:14] Searching for coresets greedily...
[2021-03-19 22:41:16] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-19 22:41:16] Creating trainer with model on device: cuda
[2021-03-19 22:41:16] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-19 22:41:24] Training accuracy: 0.9808
[2021-03-19 22:41:24] Testing on 10000 data points...
[2021-03-19 22:41:26] Test score for 1400 training labels: 0.7390
[2021-03-19 22:41:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:41:26] Found 50000 unlabelled features.
[2021-03-19 22:41:37] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:41:44] Searching for coresets greedily...
[2021-03-19 22:41:45] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-19 22:41:45] Creating trainer with model on device: cuda
[2021-03-19 22:41:45] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-19 22:41:56] Training accuracy: 0.9900
[2021-03-19 22:41:56] Testing on 10000 data points...
[2021-03-19 22:41:58] Test score for 1800 training labels: 0.7541
[2021-03-19 22:41:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:41:58] Found 50000 unlabelled features.
[2021-03-19 22:42:08] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:42:16] Searching for coresets greedily...
[2021-03-19 22:42:18] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-19 22:42:18] Creating trainer with model on device: cuda
[2021-03-19 22:42:18] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-19 22:42:31] Training accuracy: 0.9954
[2021-03-19 22:42:31] Testing on 10000 data points...
[2021-03-19 22:42:33] Test score for 2200 training labels: 0.7697
[2021-03-19 22:42:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:42:33] Found 50000 unlabelled features.
[2021-03-19 22:42:44] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:42:55] Searching for coresets greedily...
[2021-03-19 22:42:56] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-19 22:42:56] Creating trainer with model on device: cuda
[2021-03-19 22:42:56] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-19 22:43:11] Training accuracy: 0.9976
[2021-03-19 22:43:11] Testing on 10000 data points...
[2021-03-19 22:43:13] Test score for 2600 training labels: 0.7807
[2021-03-19 22:43:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:43:13] Found 50000 unlabelled features.
[2021-03-19 22:43:25] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:43:37] Searching for coresets greedily...
[2021-03-19 22:43:38] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-19 22:43:38] Creating trainer with model on device: cuda
[2021-03-19 22:43:38] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-19 22:43:56] Training accuracy: 0.9988
[2021-03-19 22:43:56] Testing on 10000 data points...
[2021-03-19 22:43:58] Test score for 3000 training labels: 0.7900
[2021-03-19 22:43:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:43:58] Found 50000 unlabelled features.
[2021-03-19 22:44:08] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:44:22] Searching for coresets greedily...
[2021-03-19 22:44:24] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-19 22:44:24] Creating trainer with model on device: cuda
[2021-03-19 22:44:24] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-19 22:44:44] Training accuracy: 0.9965
[2021-03-19 22:44:44] Testing on 10000 data points...
[2021-03-19 22:44:46] Test score for 3400 training labels: 0.7992
[2021-03-19 22:44:46] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:44:46] Found 50000 unlabelled features.
[2021-03-19 22:44:57] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:45:13] Searching for coresets greedily...
[2021-03-19 22:45:14] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-19 22:45:14] Creating trainer with model on device: cuda
[2021-03-19 22:45:14] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-19 22:45:38] Training accuracy: 0.9990
[2021-03-19 22:45:38] Testing on 10000 data points...
[2021-03-19 22:45:40] Test score for 3800 training labels: 0.8106
[2021-03-19 22:45:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:45:40] Found 50000 unlabelled features.
[2021-03-19 22:45:51] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:46:09] Searching for coresets greedily...
[2021-03-19 22:46:11] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-19 22:46:11] Creating trainer with model on device: cuda
[2021-03-19 22:46:11] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-19 22:46:36] Training accuracy: 0.9975
[2021-03-19 22:46:36] Testing on 10000 data points...
[2021-03-19 22:46:38] Test score for 4200 training labels: 0.8153
[2021-03-19 22:46:38] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:46:38] Found 50000 unlabelled features.
[2021-03-19 22:46:49] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:47:09] Searching for coresets greedily...
[2021-03-19 22:47:10] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-19 22:47:10] Creating trainer with model on device: cuda
[2021-03-19 22:47:10] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-19 22:47:38] Training accuracy: 0.9998
[2021-03-19 22:47:38] Testing on 10000 data points...
[2021-03-19 22:47:40] Test score for 4600 training labels: 0.8189
[2021-03-19 22:47:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:47:40] Found 50000 unlabelled features.
[2021-03-19 22:47:52] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:48:13] Searching for coresets greedily...
[2021-03-19 22:48:15] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-19 22:48:15] Creating trainer with model on device: cuda
[2021-03-19 22:48:15] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-19 22:48:45] Training accuracy: 0.9993
[2021-03-19 22:48:45] Testing on 10000 data points...
[2021-03-19 22:48:47] Test score for 5000 training labels: 0.8256
[2021-03-19 22:48:47] Seeded: 6
[2021-03-19 22:48:47] Running: experiment 1
[2021-03-19 22:48:47] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-19 22:48:47] Creating pretrained=True VGG16...
[2021-03-19 22:48:49] No parameter reset since we are using a pretrained model.
[2021-03-19 22:48:49] vgg-pretrained: initialized 16299850 parameters.
[2021-03-19 22:48:49] Creating trainer with model on device: cuda
[2021-03-19 22:48:49] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-19 22:49:11] Training accuracy: 0.9950
[2021-03-19 22:49:11] Testing on 10000 data points...
[2021-03-19 22:49:13] Test score for 1000 training labels: 0.7218
[2021-03-19 22:49:13] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:49:13] Found 50000 unlabelled features.
[2021-03-19 22:49:19] Showing labelled: True (1382/50000 visible, 18 redundant)
[2021-03-19 22:49:19] Creating trainer with model on device: cuda
[2021-03-19 22:49:19] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-19 22:49:27] Training accuracy: 0.9790
[2021-03-19 22:49:27] Testing on 10000 data points...
[2021-03-19 22:49:29] Test score for 1400 training labels: 0.7412
[2021-03-19 22:49:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:49:29] Found 50000 unlabelled features.
[2021-03-19 22:49:36] Showing labelled: True (1769/50000 visible, 31 redundant)
[2021-03-19 22:49:36] Creating trainer with model on device: cuda
[2021-03-19 22:49:36] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-19 22:49:47] Training accuracy: 0.9452
[2021-03-19 22:49:47] Testing on 10000 data points...
[2021-03-19 22:49:49] Test score for 1800 training labels: 0.7363
[2021-03-19 22:49:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:49:49] Found 50000 unlabelled features.
[2021-03-19 22:49:54] Showing labelled: True (2150/50000 visible, 50 redundant)
[2021-03-19 22:49:54] Creating trainer with model on device: cuda
[2021-03-19 22:49:54] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-19 22:50:07] Training accuracy: 0.9972
[2021-03-19 22:50:07] Testing on 10000 data points...
[2021-03-19 22:50:09] Test score for 2200 training labels: 0.7653
[2021-03-19 22:50:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:50:10] Found 50000 unlabelled features.
[2021-03-19 22:50:16] Showing labelled: True (2531/50000 visible, 69 redundant)
[2021-03-19 22:50:16] Creating trainer with model on device: cuda
[2021-03-19 22:50:16] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-19 22:50:32] Training accuracy: 0.9969
[2021-03-19 22:50:32] Testing on 10000 data points...
[2021-03-19 22:50:34] Test score for 2600 training labels: 0.7692
[2021-03-19 22:50:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:50:34] Found 50000 unlabelled features.
[2021-03-19 22:50:40] Showing labelled: True (2910/50000 visible, 90 redundant)
[2021-03-19 22:50:40] Creating trainer with model on device: cuda
[2021-03-19 22:50:40] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-19 22:50:58] Training accuracy: 0.9986
[2021-03-19 22:50:58] Testing on 10000 data points...
[2021-03-19 22:51:00] Test score for 3000 training labels: 0.7781
[2021-03-19 22:51:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:51:00] Found 50000 unlabelled features.
[2021-03-19 22:51:06] Showing labelled: True (3286/50000 visible, 114 redundant)
[2021-03-19 22:51:06] Creating trainer with model on device: cuda
[2021-03-19 22:51:06] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-19 22:51:26] Training accuracy: 0.9992
[2021-03-19 22:51:26] Testing on 10000 data points...
[2021-03-19 22:51:28] Test score for 3400 training labels: 0.7832
[2021-03-19 22:51:28] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:51:28] Found 50000 unlabelled features.
[2021-03-19 22:51:33] Showing labelled: True (3662/50000 visible, 138 redundant)
[2021-03-19 22:51:33] Creating trainer with model on device: cuda
[2021-03-19 22:51:33] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-19 22:51:56] Training accuracy: 0.9992
[2021-03-19 22:51:56] Testing on 10000 data points...
[2021-03-19 22:51:58] Test score for 3800 training labels: 0.7867
[2021-03-19 22:51:58] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:51:58] Found 50000 unlabelled features.
[2021-03-19 22:52:04] Showing labelled: True (4033/50000 visible, 167 redundant)
[2021-03-19 22:52:04] Creating trainer with model on device: cuda
[2021-03-19 22:52:04] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-19 22:52:28] Training accuracy: 0.9992
[2021-03-19 22:52:28] Testing on 10000 data points...
[2021-03-19 22:52:30] Test score for 4200 training labels: 0.7918
[2021-03-19 22:52:30] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:52:30] Found 50000 unlabelled features.
[2021-03-19 22:52:36] Showing labelled: True (4398/50000 visible, 202 redundant)
[2021-03-19 22:52:36] Creating trainer with model on device: cuda
[2021-03-19 22:52:36] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-19 22:53:03] Training accuracy: 0.9998
[2021-03-19 22:53:03] Testing on 10000 data points...
[2021-03-19 22:53:05] Test score for 4600 training labels: 0.7955
[2021-03-19 22:53:05] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:53:05] Found 50000 unlabelled features.
[2021-03-19 22:53:12] Showing labelled: True (4760/50000 visible, 240 redundant)
[2021-03-19 22:53:12] Creating trainer with model on device: cuda
[2021-03-19 22:53:12] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-19 22:53:41] Training accuracy: 0.9997
[2021-03-19 22:53:41] Testing on 10000 data points...
[2021-03-19 22:53:43] Test score for 5000 training labels: 0.8023
[2021-03-19 22:53:43] Seeded: 6
[2021-03-19 22:53:43] Running: experiment 2
[2021-03-19 22:53:43] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-19 22:53:43] Creating pretrained=True VGG16...
[2021-03-19 22:53:45] No parameter reset since we are using a pretrained model.
[2021-03-19 22:53:45] vgg-pretrained: initialized 16299850 parameters.
[2021-03-19 22:53:45] Creating trainer with model on device: cuda
[2021-03-19 22:53:45] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-19 22:54:08] Training accuracy: 0.9955
[2021-03-19 22:54:08] Testing on 10000 data points...
[2021-03-19 22:54:10] Test score for 1000 training labels: 0.7234
[2021-03-19 22:54:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-19 22:54:10] Found 50000 unlabelled features.
[2021-03-19 22:54:21] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 1024...
[2021-03-19 22:54:26] Searching for coresets with 20 beams...
[2021-03-19 22:54:26] Updated results: ../results/cifar10/vgg_pretrained_gcoreset/results.json
