[2021-03-21 15:31:23] Created experiment 0:
[2021-03-21 15:31:23]  - Model: vgg-pretrained
[2021-03-21 15:31:23]  - Acquisition function: lc-beam-pweighted-optimcore
[2021-03-21 15:31:23] Loading cifar10 test set...
[2021-03-21 15:31:24] Experiment repeat 1/5
[2021-03-21 15:31:24] Seeded: 5
[2021-03-21 15:31:24] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 15:31:24] Randomly labelled 1000/50000
[2021-03-21 15:31:25] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 15:31:25] Seeded: 6
[2021-03-21 15:31:25] Running: experiment 0
[2021-03-21 15:31:25] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 15:31:25] Creating pretrained=True VGG16...
[2021-03-21 15:31:26] No parameter reset since we are using a pretrained model.
[2021-03-21 15:31:26] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 15:31:26] Creating trainer with model on device: cuda
[2021-03-21 15:31:31] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 15:32:31] Training accuracy: 0.9997
[2021-03-21 15:32:31] Testing on 10000 data points...
[2021-03-21 15:32:33] Test score for 1000 training labels: 0.7239
[2021-03-21 15:32:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:32:33] Found 50000 unlabelled features.
[2021-03-21 15:32:43] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:32:46] Searching for coresets with 20 beams...
[2021-03-21 15:33:14] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 15:33:14] Creating trainer with model on device: cuda
[2021-03-21 15:33:14] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 15:33:35] Training accuracy: 0.9975
[2021-03-21 15:33:35] Testing on 10000 data points...
[2021-03-21 15:33:37] Test score for 1400 training labels: 0.7526
[2021-03-21 15:33:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:33:37] Found 50000 unlabelled features.
[2021-03-21 15:33:47] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:33:51] Searching for coresets with 20 beams...
[2021-03-21 15:34:19] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 15:34:19] Creating trainer with model on device: cuda
[2021-03-21 15:34:19] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 15:34:46] Training accuracy: 0.9962
[2021-03-21 15:34:46] Testing on 10000 data points...
[2021-03-21 15:34:48] Test score for 1800 training labels: 0.7658
[2021-03-21 15:34:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:34:48] Found 50000 unlabelled features.
[2021-03-21 15:34:59] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:35:04] Searching for coresets with 20 beams...
[2021-03-21 15:35:32] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 15:35:32] Creating trainer with model on device: cuda
[2021-03-21 15:35:32] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 15:36:06] Training accuracy: 0.9995
[2021-03-21 15:36:06] Testing on 10000 data points...
[2021-03-21 15:36:08] Test score for 2200 training labels: 0.7863
[2021-03-21 15:36:08] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:36:08] Found 50000 unlabelled features.
[2021-03-21 15:36:18] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:36:25] Searching for coresets with 20 beams...
[2021-03-21 15:36:53] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 15:36:53] Creating trainer with model on device: cuda
[2021-03-21 15:36:53] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 15:37:32] Training accuracy: 0.9993
[2021-03-21 15:37:32] Testing on 10000 data points...
[2021-03-21 15:37:34] Test score for 2600 training labels: 0.8024
[2021-03-21 15:37:34] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:37:34] Found 50000 unlabelled features.
[2021-03-21 15:37:44] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:37:52] Searching for coresets with 20 beams...
[2021-03-21 15:38:20] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 15:38:20] Creating trainer with model on device: cuda
[2021-03-21 15:38:20] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 15:39:05] Training accuracy: 1.0000
[2021-03-21 15:39:05] Testing on 10000 data points...
[2021-03-21 15:39:07] Test score for 3000 training labels: 0.8071
[2021-03-21 15:39:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:39:07] Found 50000 unlabelled features.
[2021-03-21 15:39:18] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:39:27] Searching for coresets with 20 beams...
[2021-03-21 15:39:55] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 15:39:55] Creating trainer with model on device: cuda
[2021-03-21 15:39:55] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 15:40:46] Training accuracy: 1.0000
[2021-03-21 15:40:46] Testing on 10000 data points...
[2021-03-21 15:40:48] Test score for 3400 training labels: 0.8161
[2021-03-21 15:40:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:40:48] Found 50000 unlabelled features.
[2021-03-21 15:40:59] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:41:09] Searching for coresets with 20 beams...
[2021-03-21 15:41:37] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 15:41:37] Creating trainer with model on device: cuda
[2021-03-21 15:41:37] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 15:42:35] Training accuracy: 1.0000
[2021-03-21 15:42:35] Testing on 10000 data points...
[2021-03-21 15:42:37] Test score for 3800 training labels: 0.8241
[2021-03-21 15:42:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:42:37] Found 50000 unlabelled features.
[2021-03-21 15:42:48] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:42:59] Searching for coresets with 20 beams...
[2021-03-21 15:43:27] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 15:43:27] Creating trainer with model on device: cuda
[2021-03-21 15:43:27] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 15:44:31] Training accuracy: 0.9999
[2021-03-21 15:44:31] Testing on 10000 data points...
[2021-03-21 15:44:33] Test score for 4200 training labels: 0.8299
[2021-03-21 15:44:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:44:33] Found 50000 unlabelled features.
[2021-03-21 15:44:44] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:44:56] Searching for coresets with 20 beams...
[2021-03-21 15:45:24] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 15:45:24] Creating trainer with model on device: cuda
[2021-03-21 15:45:24] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 15:46:33] Training accuracy: 1.0000
[2021-03-21 15:46:33] Testing on 10000 data points...
[2021-03-21 15:46:35] Test score for 4600 training labels: 0.8335
[2021-03-21 15:46:35] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:46:35] Found 50000 unlabelled features.
[2021-03-21 15:46:46] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:46:59] Searching for coresets with 20 beams...
[2021-03-21 15:47:28] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 15:47:28] Creating trainer with model on device: cuda
[2021-03-21 15:47:28] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 15:48:43] Training accuracy: 1.0000
[2021-03-21 15:48:43] Testing on 10000 data points...
[2021-03-21 15:48:45] Test score for 5000 training labels: 0.8354
[2021-03-21 15:48:45] Experiment repeat 2/5
[2021-03-21 15:48:45] Seeded: 6
[2021-03-21 15:48:45] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 15:48:45] Randomly labelled 1000/50000
[2021-03-21 15:48:45] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 15:48:45] Seeded: 7
[2021-03-21 15:48:45] Running: experiment 0
[2021-03-21 15:48:45] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 15:48:45] Creating pretrained=True VGG16...
[2021-03-21 15:48:47] No parameter reset since we are using a pretrained model.
[2021-03-21 15:48:47] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 15:48:47] Creating trainer with model on device: cuda
[2021-03-21 15:48:47] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 15:49:47] Training accuracy: 0.9989
[2021-03-21 15:49:47] Testing on 10000 data points...
[2021-03-21 15:49:49] Test score for 1000 training labels: 0.7377
[2021-03-21 15:49:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:49:49] Found 50000 unlabelled features.
[2021-03-21 15:49:59] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:50:02] Searching for coresets with 20 beams...
[2021-03-21 15:50:31] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 15:50:31] Creating trainer with model on device: cuda
[2021-03-21 15:50:31] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 15:50:52] Training accuracy: 0.9989
[2021-03-21 15:50:52] Testing on 10000 data points...
[2021-03-21 15:50:54] Test score for 1400 training labels: 0.7608
[2021-03-21 15:50:54] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:50:54] Found 50000 unlabelled features.
[2021-03-21 15:51:04] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:51:08] Searching for coresets with 20 beams...
[2021-03-21 15:51:37] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 15:51:37] Creating trainer with model on device: cuda
[2021-03-21 15:51:37] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 15:52:04] Training accuracy: 0.9985
[2021-03-21 15:52:04] Testing on 10000 data points...
[2021-03-21 15:52:06] Test score for 1800 training labels: 0.7774
[2021-03-21 15:52:06] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:52:06] Found 50000 unlabelled features.
[2021-03-21 15:52:16] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:52:22] Searching for coresets with 20 beams...
[2021-03-21 15:52:50] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 15:52:50] Creating trainer with model on device: cuda
[2021-03-21 15:52:50] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 15:53:24] Training accuracy: 0.9993
[2021-03-21 15:53:24] Testing on 10000 data points...
[2021-03-21 15:53:26] Test score for 2200 training labels: 0.7936
[2021-03-21 15:53:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:53:26] Found 50000 unlabelled features.
[2021-03-21 15:53:36] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:53:43] Searching for coresets with 20 beams...
[2021-03-21 15:54:11] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 15:54:11] Creating trainer with model on device: cuda
[2021-03-21 15:54:11] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 15:54:50] Training accuracy: 0.9992
[2021-03-21 15:54:50] Testing on 10000 data points...
[2021-03-21 15:54:52] Test score for 2600 training labels: 0.7999
[2021-03-21 15:54:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:54:52] Found 50000 unlabelled features.
[2021-03-21 15:55:03] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:55:11] Searching for coresets with 20 beams...
[2021-03-21 15:55:39] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 15:55:39] Creating trainer with model on device: cuda
[2021-03-21 15:55:39] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 15:56:24] Training accuracy: 0.9996
[2021-03-21 15:56:24] Testing on 10000 data points...
[2021-03-21 15:56:26] Test score for 3000 training labels: 0.8086
[2021-03-21 15:56:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:56:26] Found 50000 unlabelled features.
[2021-03-21 15:56:37] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:56:46] Searching for coresets with 20 beams...
[2021-03-21 15:57:14] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 15:57:14] Creating trainer with model on device: cuda
[2021-03-21 15:57:14] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 15:58:05] Training accuracy: 0.9999
[2021-03-21 15:58:05] Testing on 10000 data points...
[2021-03-21 15:58:07] Test score for 3400 training labels: 0.8168
[2021-03-21 15:58:07] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:58:07] Found 50000 unlabelled features.
[2021-03-21 15:58:18] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 15:58:28] Searching for coresets with 20 beams...
[2021-03-21 15:58:56] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 15:58:56] Creating trainer with model on device: cuda
[2021-03-21 15:58:56] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 15:59:54] Training accuracy: 1.0000
[2021-03-21 15:59:54] Testing on 10000 data points...
[2021-03-21 15:59:56] Test score for 3800 training labels: 0.8181
[2021-03-21 15:59:56] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 15:59:56] Found 50000 unlabelled features.
[2021-03-21 16:00:07] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:00:18] Searching for coresets with 20 beams...
[2021-03-21 16:00:46] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 16:00:46] Creating trainer with model on device: cuda
[2021-03-21 16:00:46] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 16:01:50] Training accuracy: 1.0000
[2021-03-21 16:01:50] Testing on 10000 data points...
[2021-03-21 16:01:52] Test score for 4200 training labels: 0.8260
[2021-03-21 16:01:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:01:52] Found 50000 unlabelled features.
[2021-03-21 16:02:03] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:02:15] Searching for coresets with 20 beams...
[2021-03-21 16:02:43] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 16:02:43] Creating trainer with model on device: cuda
[2021-03-21 16:02:43] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 16:03:53] Training accuracy: 0.9999
[2021-03-21 16:03:53] Testing on 10000 data points...
[2021-03-21 16:03:55] Test score for 4600 training labels: 0.8279
[2021-03-21 16:03:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:03:55] Found 50000 unlabelled features.
[2021-03-21 16:04:06] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:04:19] Searching for coresets with 20 beams...
[2021-03-21 16:04:48] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 16:04:48] Creating trainer with model on device: cuda
[2021-03-21 16:04:48] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 16:06:04] Training accuracy: 0.9995
[2021-03-21 16:06:04] Testing on 10000 data points...
[2021-03-21 16:06:06] Test score for 5000 training labels: 0.8364
[2021-03-21 16:06:06] Experiment repeat 3/5
[2021-03-21 16:06:06] Seeded: 7
[2021-03-21 16:06:06] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 16:06:06] Randomly labelled 1000/50000
[2021-03-21 16:06:06] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 16:06:06] Seeded: 8
[2021-03-21 16:06:06] Running: experiment 0
[2021-03-21 16:06:06] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 16:06:06] Creating pretrained=True VGG16...
[2021-03-21 16:06:07] No parameter reset since we are using a pretrained model.
[2021-03-21 16:06:07] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 16:06:07] Creating trainer with model on device: cuda
[2021-03-21 16:06:07] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 16:07:08] Training accuracy: 1.0000
[2021-03-21 16:07:08] Testing on 10000 data points...
[2021-03-21 16:07:10] Test score for 1000 training labels: 0.7388
[2021-03-21 16:07:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:07:10] Found 50000 unlabelled features.
[2021-03-21 16:07:20] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:07:23] Searching for coresets with 20 beams...
[2021-03-21 16:07:51] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 16:07:51] Creating trainer with model on device: cuda
[2021-03-21 16:07:51] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 16:08:13] Training accuracy: 0.9807
[2021-03-21 16:08:13] Testing on 10000 data points...
[2021-03-21 16:08:15] Test score for 1400 training labels: 0.7281
[2021-03-21 16:08:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:08:15] Found 50000 unlabelled features.
[2021-03-21 16:08:25] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:08:29] Searching for coresets with 20 beams...
[2021-03-21 16:08:57] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 16:08:57] Creating trainer with model on device: cuda
[2021-03-21 16:08:57] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 16:09:25] Training accuracy: 0.9985
[2021-03-21 16:09:25] Testing on 10000 data points...
[2021-03-21 16:09:27] Test score for 1800 training labels: 0.7685
[2021-03-21 16:09:27] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:09:27] Found 50000 unlabelled features.
[2021-03-21 16:09:37] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:09:43] Searching for coresets with 20 beams...
[2021-03-21 16:10:11] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 16:10:11] Creating trainer with model on device: cuda
[2021-03-21 16:10:11] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 16:10:45] Training accuracy: 0.9999
[2021-03-21 16:10:45] Testing on 10000 data points...
[2021-03-21 16:10:47] Test score for 2200 training labels: 0.7823
[2021-03-21 16:10:47] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:10:47] Found 50000 unlabelled features.
[2021-03-21 16:10:57] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:11:04] Searching for coresets with 20 beams...
[2021-03-21 16:11:32] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 16:11:32] Creating trainer with model on device: cuda
[2021-03-21 16:11:32] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 16:12:12] Training accuracy: 0.9993
[2021-03-21 16:12:12] Testing on 10000 data points...
[2021-03-21 16:12:14] Test score for 2600 training labels: 0.7941
[2021-03-21 16:12:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:12:14] Found 50000 unlabelled features.
[2021-03-21 16:12:24] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:12:32] Searching for coresets with 20 beams...
[2021-03-21 16:13:00] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 16:13:00] Creating trainer with model on device: cuda
[2021-03-21 16:13:00] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 16:13:46] Training accuracy: 0.9997
[2021-03-21 16:13:46] Testing on 10000 data points...
[2021-03-21 16:13:48] Test score for 3000 training labels: 0.8024
[2021-03-21 16:13:48] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:13:48] Found 50000 unlabelled features.
[2021-03-21 16:13:58] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:14:07] Searching for coresets with 20 beams...
[2021-03-21 16:14:35] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 16:14:35] Creating trainer with model on device: cuda
[2021-03-21 16:14:35] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 16:15:27] Training accuracy: 0.9996
[2021-03-21 16:15:27] Testing on 10000 data points...
[2021-03-21 16:15:29] Test score for 3400 training labels: 0.8091
[2021-03-21 16:15:29] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:15:29] Found 50000 unlabelled features.
[2021-03-21 16:15:40] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:15:50] Searching for coresets with 20 beams...
[2021-03-21 16:16:18] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 16:16:18] Creating trainer with model on device: cuda
[2021-03-21 16:16:18] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 16:17:16] Training accuracy: 1.0000
[2021-03-21 16:17:16] Testing on 10000 data points...
[2021-03-21 16:17:18] Test score for 3800 training labels: 0.8127
[2021-03-21 16:17:18] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:17:18] Found 50000 unlabelled features.
[2021-03-21 16:17:29] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:17:40] Searching for coresets with 20 beams...
[2021-03-21 16:18:08] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 16:18:08] Creating trainer with model on device: cuda
[2021-03-21 16:18:08] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 16:19:12] Training accuracy: 1.0000
[2021-03-21 16:19:12] Testing on 10000 data points...
[2021-03-21 16:19:14] Test score for 4200 training labels: 0.8229
[2021-03-21 16:19:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:19:14] Found 50000 unlabelled features.
[2021-03-21 16:19:25] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:19:37] Searching for coresets with 20 beams...
[2021-03-21 16:20:06] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 16:20:06] Creating trainer with model on device: cuda
[2021-03-21 16:20:06] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 16:21:15] Training accuracy: 1.0000
[2021-03-21 16:21:15] Testing on 10000 data points...
[2021-03-21 16:21:17] Test score for 4600 training labels: 0.8278
[2021-03-21 16:21:17] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:21:17] Found 50000 unlabelled features.
[2021-03-21 16:21:28] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:21:42] Searching for coresets with 20 beams...
[2021-03-21 16:22:10] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 16:22:10] Creating trainer with model on device: cuda
[2021-03-21 16:22:10] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 16:23:26] Training accuracy: 1.0000
[2021-03-21 16:23:26] Testing on 10000 data points...
[2021-03-21 16:23:28] Test score for 5000 training labels: 0.8347
[2021-03-21 16:23:28] Experiment repeat 4/5
[2021-03-21 16:23:28] Seeded: 8
[2021-03-21 16:23:28] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 16:23:28] Randomly labelled 1000/50000
[2021-03-21 16:23:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 16:23:28] Seeded: 9
[2021-03-21 16:23:28] Running: experiment 0
[2021-03-21 16:23:28] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 16:23:28] Creating pretrained=True VGG16...
[2021-03-21 16:23:29] No parameter reset since we are using a pretrained model.
[2021-03-21 16:23:29] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 16:23:29] Creating trainer with model on device: cuda
[2021-03-21 16:23:29] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 16:24:30] Training accuracy: 0.9986
[2021-03-21 16:24:30] Testing on 10000 data points...
[2021-03-21 16:24:32] Test score for 1000 training labels: 0.7269
[2021-03-21 16:24:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:24:32] Found 50000 unlabelled features.
[2021-03-21 16:24:42] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:24:45] Searching for coresets with 20 beams...
[2021-03-21 16:25:14] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 16:25:14] Creating trainer with model on device: cuda
[2021-03-21 16:25:14] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 16:25:35] Training accuracy: 0.9987
[2021-03-21 16:25:35] Testing on 10000 data points...
[2021-03-21 16:25:37] Test score for 1400 training labels: 0.7618
[2021-03-21 16:25:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:25:37] Found 50000 unlabelled features.
[2021-03-21 16:25:47] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:25:52] Searching for coresets with 20 beams...
[2021-03-21 16:26:20] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 16:26:20] Creating trainer with model on device: cuda
[2021-03-21 16:26:20] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 16:26:47] Training accuracy: 0.9980
[2021-03-21 16:26:47] Testing on 10000 data points...
[2021-03-21 16:26:49] Test score for 1800 training labels: 0.7677
[2021-03-21 16:26:49] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:26:49] Found 50000 unlabelled features.
[2021-03-21 16:27:00] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:27:05] Searching for coresets with 20 beams...
[2021-03-21 16:27:33] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 16:27:33] Creating trainer with model on device: cuda
[2021-03-21 16:27:33] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 16:28:07] Training accuracy: 0.9994
[2021-03-21 16:28:07] Testing on 10000 data points...
[2021-03-21 16:28:09] Test score for 2200 training labels: 0.7771
[2021-03-21 16:28:09] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:28:09] Found 50000 unlabelled features.
[2021-03-21 16:28:20] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:28:26] Searching for coresets with 20 beams...
[2021-03-21 16:28:55] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 16:28:55] Creating trainer with model on device: cuda
[2021-03-21 16:28:55] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 16:29:34] Training accuracy: 1.0000
[2021-03-21 16:29:34] Testing on 10000 data points...
[2021-03-21 16:29:36] Test score for 2600 training labels: 0.7871
[2021-03-21 16:29:36] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:29:36] Found 50000 unlabelled features.
[2021-03-21 16:29:47] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:29:54] Searching for coresets with 20 beams...
[2021-03-21 16:30:23] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 16:30:23] Creating trainer with model on device: cuda
[2021-03-21 16:30:23] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 16:31:08] Training accuracy: 1.0000
[2021-03-21 16:31:08] Testing on 10000 data points...
[2021-03-21 16:31:10] Test score for 3000 training labels: 0.7980
[2021-03-21 16:31:10] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:31:10] Found 50000 unlabelled features.
[2021-03-21 16:31:21] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:31:30] Searching for coresets with 20 beams...
[2021-03-21 16:31:58] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 16:31:58] Creating trainer with model on device: cuda
[2021-03-21 16:31:58] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 16:32:50] Training accuracy: 0.9998
[2021-03-21 16:32:50] Testing on 10000 data points...
[2021-03-21 16:32:52] Test score for 3400 training labels: 0.8101
[2021-03-21 16:32:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:32:52] Found 50000 unlabelled features.
[2021-03-21 16:33:02] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:33:12] Searching for coresets with 20 beams...
[2021-03-21 16:33:41] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 16:33:41] Creating trainer with model on device: cuda
[2021-03-21 16:33:41] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 16:34:39] Training accuracy: 1.0000
[2021-03-21 16:34:39] Testing on 10000 data points...
[2021-03-21 16:34:41] Test score for 3800 training labels: 0.8184
[2021-03-21 16:34:41] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:34:41] Found 50000 unlabelled features.
[2021-03-21 16:34:52] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:35:03] Searching for coresets with 20 beams...
[2021-03-21 16:35:31] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 16:35:31] Creating trainer with model on device: cuda
[2021-03-21 16:35:31] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 16:36:35] Training accuracy: 1.0000
[2021-03-21 16:36:35] Testing on 10000 data points...
[2021-03-21 16:36:37] Test score for 4200 training labels: 0.8223
[2021-03-21 16:36:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:36:37] Found 50000 unlabelled features.
[2021-03-21 16:36:48] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:37:00] Searching for coresets with 20 beams...
[2021-03-21 16:37:29] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 16:37:29] Creating trainer with model on device: cuda
[2021-03-21 16:37:29] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 16:38:38] Training accuracy: 1.0000
[2021-03-21 16:38:38] Testing on 10000 data points...
[2021-03-21 16:38:40] Test score for 4600 training labels: 0.8319
[2021-03-21 16:38:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:38:40] Found 50000 unlabelled features.
[2021-03-21 16:38:51] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:39:04] Searching for coresets with 20 beams...
[2021-03-21 16:39:33] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 16:39:33] Creating trainer with model on device: cuda
[2021-03-21 16:39:33] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 16:40:48] Training accuracy: 1.0000
[2021-03-21 16:40:48] Testing on 10000 data points...
[2021-03-21 16:40:50] Test score for 5000 training labels: 0.8387
[2021-03-21 16:40:50] Experiment repeat 5/5
[2021-03-21 16:40:50] Seeded: 9
[2021-03-21 16:40:50] Using 2.00% labels of the dataset (1000/50000)
[2021-03-21 16:40:50] Randomly labelled 1000/50000
[2021-03-21 16:40:50] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 16:40:50] Seeded: 10
[2021-03-21 16:40:50] Running: experiment 0
[2021-03-21 16:40:51] Showing labelled: True (1000/50000 visible, 0 redundant)
[2021-03-21 16:40:51] Creating pretrained=True VGG16...
[2021-03-21 16:40:52] No parameter reset since we are using a pretrained model.
[2021-03-21 16:40:52] vgg-pretrained: initialized 15245130 parameters.
[2021-03-21 16:40:52] Creating trainer with model on device: cuda
[2021-03-21 16:40:52] Training vgg-pretrained across 1000 data points in cifar10...
[2021-03-21 16:41:53] Training accuracy: 0.9998
[2021-03-21 16:41:53] Testing on 10000 data points...
[2021-03-21 16:41:55] Test score for 1000 training labels: 0.7332
[2021-03-21 16:41:55] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:41:55] Found 50000 unlabelled features.
[2021-03-21 16:42:05] Computing distance between 1000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:42:08] Searching for coresets with 20 beams...
[2021-03-21 16:42:36] Showing labelled: True (1400/50000 visible, 0 redundant)
[2021-03-21 16:42:36] Creating trainer with model on device: cuda
[2021-03-21 16:42:36] Training vgg-pretrained across 1400 data points in cifar10...
[2021-03-21 16:42:58] Training accuracy: 0.9913
[2021-03-21 16:42:58] Testing on 10000 data points...
[2021-03-21 16:43:00] Test score for 1400 training labels: 0.7472
[2021-03-21 16:43:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:43:00] Found 50000 unlabelled features.
[2021-03-21 16:43:10] Computing distance between 1400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:43:14] Searching for coresets with 20 beams...
[2021-03-21 16:43:42] Showing labelled: True (1800/50000 visible, 0 redundant)
[2021-03-21 16:43:42] Creating trainer with model on device: cuda
[2021-03-21 16:43:42] Training vgg-pretrained across 1800 data points in cifar10...
[2021-03-21 16:44:10] Training accuracy: 0.9986
[2021-03-21 16:44:10] Testing on 10000 data points...
[2021-03-21 16:44:12] Test score for 1800 training labels: 0.7700
[2021-03-21 16:44:12] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:44:12] Found 50000 unlabelled features.
[2021-03-21 16:44:22] Computing distance between 1800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:44:28] Searching for coresets with 20 beams...
[2021-03-21 16:44:56] Showing labelled: True (2200/50000 visible, 0 redundant)
[2021-03-21 16:44:56] Creating trainer with model on device: cuda
[2021-03-21 16:44:56] Training vgg-pretrained across 2200 data points in cifar10...
[2021-03-21 16:45:30] Training accuracy: 0.9988
[2021-03-21 16:45:30] Testing on 10000 data points...
[2021-03-21 16:45:32] Test score for 2200 training labels: 0.7842
[2021-03-21 16:45:32] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:45:32] Found 50000 unlabelled features.
[2021-03-21 16:45:43] Computing distance between 2200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:45:49] Searching for coresets with 20 beams...
[2021-03-21 16:46:18] Showing labelled: True (2600/50000 visible, 0 redundant)
[2021-03-21 16:46:18] Creating trainer with model on device: cuda
[2021-03-21 16:46:18] Training vgg-pretrained across 2600 data points in cifar10...
[2021-03-21 16:46:57] Training accuracy: 1.0000
[2021-03-21 16:46:57] Testing on 10000 data points...
[2021-03-21 16:46:59] Test score for 2600 training labels: 0.7975
[2021-03-21 16:46:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:46:59] Found 50000 unlabelled features.
[2021-03-21 16:47:10] Computing distance between 2600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:47:17] Searching for coresets with 20 beams...
[2021-03-21 16:47:46] Showing labelled: True (3000/50000 visible, 0 redundant)
[2021-03-21 16:47:46] Creating trainer with model on device: cuda
[2021-03-21 16:47:46] Training vgg-pretrained across 3000 data points in cifar10...
[2021-03-21 16:48:31] Training accuracy: 0.9993
[2021-03-21 16:48:31] Testing on 10000 data points...
[2021-03-21 16:48:33] Test score for 3000 training labels: 0.8091
[2021-03-21 16:48:33] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:48:33] Found 50000 unlabelled features.
[2021-03-21 16:48:44] Computing distance between 3000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:48:53] Searching for coresets with 20 beams...
[2021-03-21 16:49:21] Showing labelled: True (3400/50000 visible, 0 redundant)
[2021-03-21 16:49:21] Creating trainer with model on device: cuda
[2021-03-21 16:49:21] Training vgg-pretrained across 3400 data points in cifar10...
[2021-03-21 16:50:12] Training accuracy: 0.9999
[2021-03-21 16:50:12] Testing on 10000 data points...
[2021-03-21 16:50:14] Test score for 3400 training labels: 0.8119
[2021-03-21 16:50:14] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:50:14] Found 50000 unlabelled features.
[2021-03-21 16:50:25] Computing distance between 3400 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:50:35] Searching for coresets with 20 beams...
[2021-03-21 16:51:03] Showing labelled: True (3800/50000 visible, 0 redundant)
[2021-03-21 16:51:03] Creating trainer with model on device: cuda
[2021-03-21 16:51:03] Training vgg-pretrained across 3800 data points in cifar10...
[2021-03-21 16:52:02] Training accuracy: 0.9987
[2021-03-21 16:52:02] Testing on 10000 data points...
[2021-03-21 16:52:04] Test score for 3800 training labels: 0.8199
[2021-03-21 16:52:04] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:52:04] Found 50000 unlabelled features.
[2021-03-21 16:52:14] Computing distance between 3800 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:52:26] Searching for coresets with 20 beams...
[2021-03-21 16:52:54] Showing labelled: True (4200/50000 visible, 0 redundant)
[2021-03-21 16:52:54] Creating trainer with model on device: cuda
[2021-03-21 16:52:54] Training vgg-pretrained across 4200 data points in cifar10...
[2021-03-21 16:53:58] Training accuracy: 0.9989
[2021-03-21 16:53:58] Testing on 10000 data points...
[2021-03-21 16:54:00] Test score for 4200 training labels: 0.8232
[2021-03-21 16:54:00] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:54:00] Found 50000 unlabelled features.
[2021-03-21 16:54:10] Computing distance between 4200 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:54:23] Searching for coresets with 20 beams...
[2021-03-21 16:54:51] Showing labelled: True (4600/50000 visible, 0 redundant)
[2021-03-21 16:54:51] Creating trainer with model on device: cuda
[2021-03-21 16:54:51] Training vgg-pretrained across 4600 data points in cifar10...
[2021-03-21 16:56:01] Training accuracy: 0.9999
[2021-03-21 16:56:01] Testing on 10000 data points...
[2021-03-21 16:56:03] Test score for 4600 training labels: 0.8309
[2021-03-21 16:56:03] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-21 16:56:03] Found 50000 unlabelled features.
[2021-03-21 16:56:14] Computing distance between 4600 labelled and 50000 unlabelled vectors of length 512...
[2021-03-21 16:56:27] Searching for coresets with 20 beams...
[2021-03-21 16:56:55] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-21 16:56:55] Creating trainer with model on device: cuda
[2021-03-21 16:56:55] Training vgg-pretrained across 5000 data points in cifar10...
[2021-03-21 16:58:11] Training accuracy: 1.0000
[2021-03-21 16:58:11] Testing on 10000 data points...
[2021-03-21 16:58:13] Test score for 5000 training labels: 0.8360
[2021-03-21 16:58:13] Updated results: ../results/cifar10/main_optimcore/results.json
