[2021-03-18 20:22:21] Created experiment 0:
[2021-03-18 20:22:21]  - Model: vgg
[2021-03-18 20:22:21]  - Acquisition function: random
[2021-03-18 20:22:21] Created experiment 1:
[2021-03-18 20:22:21]  - Model: vgg
[2021-03-18 20:22:21]  - Acquisition function: greedy-coreset
[2021-03-18 20:22:21] Created experiment 2:
[2021-03-18 20:22:21]  - Model: vgg
[2021-03-18 20:22:21]  - Acquisition function: least-confidence
[2021-03-18 20:22:21] Created experiment 3:
[2021-03-18 20:22:21]  - Model: vgg
[2021-03-18 20:22:21]  - Acquisition function: lc-beam-coreset
[2021-03-18 20:22:21] Loading cifar10-augmented test set...
[2021-03-18 20:22:22] Experiment repeat 1/1
[2021-03-18 20:22:22] Seeded: 5
[2021-03-18 20:22:22] Using 10.00% labels of the dataset (5000/50000)
[2021-03-18 20:22:22] Randomly labelled 5000/50000
[2021-03-18 20:22:22] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-18 20:22:22] Seeded: 6
[2021-03-18 20:22:22] Running: experiment 0
[2021-03-18 20:22:22] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-18 20:22:22] Creating pretrained=False VGG16...
[2021-03-18 20:22:25] Reset parameters according to He et al.
[2021-03-18 20:22:25] vgg: initialized 15250250 parameters.
[2021-03-18 20:22:25] Creating trainer with model on device: cuda
[2021-03-18 20:22:29] Training vgg across 5000 data points in cifar10-augmented...
[2021-03-18 20:29:13] Training accuracy: 0.7325
[2021-03-18 20:29:13] Testing on 10000 data points...
[2021-03-18 20:29:15] Test score for 5000 training labels: 0.6305
[2021-03-18 20:29:15] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 20:29:15] Found 50000 unlabelled features.
[2021-03-18 20:29:24] Showing labelled: True (9512/50000 visible, 488 redundant)
[2021-03-18 20:29:24] Creating pretrained=False VGG16...
[2021-03-18 20:29:27] Reset parameters according to He et al.
[2021-03-18 20:29:27] vgg: initialized 15250250 parameters.
[2021-03-18 20:29:27] Creating trainer with model on device: cuda
[2021-03-18 20:29:27] Training vgg across 10000 data points in cifar10-augmented...
[2021-03-18 20:42:57] Training accuracy: 0.9478
[2021-03-18 20:42:57] Testing on 10000 data points...
[2021-03-18 20:42:59] Test score for 10000 training labels: 0.7348
[2021-03-18 20:42:59] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 20:42:59] Found 50000 unlabelled features.
[2021-03-18 20:43:08] Showing labelled: True (13552/50000 visible, 1448 redundant)
[2021-03-18 20:43:08] Creating pretrained=False VGG16...
[2021-03-18 20:43:10] Reset parameters according to He et al.
[2021-03-18 20:43:10] vgg: initialized 15250250 parameters.
[2021-03-18 20:43:10] Creating trainer with model on device: cuda
[2021-03-18 20:43:10] Training vgg across 15000 data points in cifar10-augmented...
[2021-03-18 21:03:38] Training accuracy: 0.9749
[2021-03-18 21:03:38] Testing on 10000 data points...
[2021-03-18 21:03:40] Test score for 15000 training labels: 0.7793
[2021-03-18 21:03:40] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 21:03:40] Found 50000 unlabelled features.
[2021-03-18 21:03:49] Showing labelled: True (17219/50000 visible, 2781 redundant)
[2021-03-18 21:03:49] Creating pretrained=False VGG16...
[2021-03-18 21:03:52] Reset parameters according to He et al.
[2021-03-18 21:03:52] vgg: initialized 15250250 parameters.
[2021-03-18 21:03:52] Creating trainer with model on device: cuda
[2021-03-18 21:03:52] Training vgg across 20000 data points in cifar10-augmented...
[2021-03-18 21:32:39] Training accuracy: 0.9923
[2021-03-18 21:32:39] Testing on 10000 data points...
[2021-03-18 21:32:41] Test score for 20000 training labels: 0.8007
[2021-03-18 21:32:41] Seeded: 6
[2021-03-18 21:32:41] Running: experiment 1
[2021-03-18 21:32:41] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-18 21:32:41] Creating pretrained=False VGG16...
[2021-03-18 21:32:44] Reset parameters according to He et al.
[2021-03-18 21:32:44] vgg: initialized 15250250 parameters.
[2021-03-18 21:32:44] Creating trainer with model on device: cuda
[2021-03-18 21:32:44] Training vgg across 5000 data points in cifar10-augmented...
[2021-03-18 21:39:24] Training accuracy: 0.7339
[2021-03-18 21:39:24] Testing on 10000 data points...
[2021-03-18 21:39:26] Test score for 5000 training labels: 0.6283
[2021-03-18 21:39:26] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 21:39:26] Found 50000 unlabelled features.
[2021-03-18 21:39:40] Computing distance between 5000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-18 21:39:54] Searching for coresets greedily...
[2021-03-18 21:40:03] Showing labelled: True (9720/50000 visible, 280 redundant)
[2021-03-18 21:40:03] Creating pretrained=False VGG16...
[2021-03-18 21:40:06] Reset parameters according to He et al.
[2021-03-18 21:40:06] vgg: initialized 15250250 parameters.
[2021-03-18 21:40:06] Creating trainer with model on device: cuda
[2021-03-18 21:40:06] Training vgg across 10000 data points in cifar10-augmented...
[2021-03-18 21:53:35] Training accuracy: 0.9518
[2021-03-18 21:53:35] Testing on 10000 data points...
[2021-03-18 21:53:37] Test score for 10000 training labels: 0.6984
[2021-03-18 21:53:37] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 21:53:37] Found 50000 unlabelled features.
[2021-03-18 21:53:53] Computing distance between 10000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-18 21:54:23] Searching for coresets greedily...
[2021-03-18 21:54:32] Showing labelled: True (13429/50000 visible, 1571 redundant)
[2021-03-18 21:54:32] Creating pretrained=False VGG16...
[2021-03-18 21:54:34] Reset parameters according to He et al.
[2021-03-18 21:54:35] vgg: initialized 15250250 parameters.
[2021-03-18 21:54:35] Creating trainer with model on device: cuda
[2021-03-18 21:54:35] Training vgg across 15000 data points in cifar10-augmented...
[2021-03-18 22:19:49] Training accuracy: 0.9840
[2021-03-18 22:19:49] Testing on 10000 data points...
[2021-03-18 22:19:51] Test score for 15000 training labels: 0.7605
[2021-03-18 22:19:51] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 22:19:51] Found 50000 unlabelled features.
[2021-03-18 22:20:09] Computing distance between 15000 labelled and 50000 unlabelled vectors of length 512...
[2021-03-18 22:20:53] Searching for coresets greedily...
[2021-03-18 22:21:02] Showing labelled: True (16367/50000 visible, 3633 redundant)
[2021-03-18 22:21:02] Creating pretrained=False VGG16...
[2021-03-18 22:21:04] Reset parameters according to He et al.
[2021-03-18 22:21:05] vgg: initialized 15250250 parameters.
[2021-03-18 22:21:05] Creating trainer with model on device: cuda
[2021-03-18 22:21:05] Training vgg across 20000 data points in cifar10-augmented...
[2021-03-18 22:48:07] Training accuracy: 0.9945
[2021-03-18 22:48:07] Testing on 10000 data points...
[2021-03-18 22:48:09] Test score for 20000 training labels: 0.7877
[2021-03-18 22:48:09] Seeded: 6
[2021-03-18 22:48:09] Running: experiment 2
[2021-03-18 22:48:09] Showing labelled: True (5000/50000 visible, 0 redundant)
[2021-03-18 22:48:09] Creating pretrained=False VGG16...
[2021-03-18 22:48:11] Reset parameters according to He et al.
[2021-03-18 22:48:11] vgg: initialized 15250250 parameters.
[2021-03-18 22:48:11] Creating trainer with model on device: cuda
[2021-03-18 22:48:11] Training vgg across 5000 data points in cifar10-augmented...
[2021-03-18 22:54:50] Training accuracy: 0.7307
[2021-03-18 22:54:50] Testing on 10000 data points...
[2021-03-18 22:54:52] Test score for 5000 training labels: 0.6261
[2021-03-18 22:54:52] Showing labelled: False (50000/50000 visible, 0 redundant)
[2021-03-18 22:54:52] Found 50000 unlabelled features.
[2021-03-18 22:55:04] Showing labelled: True (9561/50000 visible, 439 redundant)
[2021-03-18 22:55:04] Creating pretrained=False VGG16...
[2021-03-18 22:55:08] Reset parameters according to He et al.
[2021-03-18 22:55:08] vgg: initialized 15250250 parameters.
[2021-03-18 22:55:08] Creating trainer with model on device: cuda
[2021-03-18 22:55:08] Training vgg across 10000 data points in cifar10-augmented...
[2021-03-18 23:06:40] Updated results: ../results/cifar10-aug/vgg/results.json
