[2021-03-17 15:31:21] Seeded: 5
[2021-03-17 15:31:21] Created experiment 0:
[2021-03-17 15:31:21]  - Model: simple-cnn
[2021-03-17 15:31:21]  - Acquisition function: lc-beam-coreset
[2021-03-17 15:31:21] Created experiment 1:
[2021-03-17 15:31:21]  - Model: simple-cnn
[2021-03-17 15:31:21]  - Acquisition function: greedy-coreset
[2021-03-17 15:31:21] Created experiment 2:
[2021-03-17 15:31:21]  - Model: simple-cnn
[2021-03-17 15:31:21]  - Acquisition function: random
[2021-03-17 15:31:21] Created experiment 3:
[2021-03-17 15:31:21]  - Model: simple-cnn
[2021-03-17 15:31:21]  - Acquisition function: least-confidence
[2021-03-17 15:31:21] Loading cifar10 test set...
[2021-03-17 15:31:28] Experiment repeat 1/1
[2021-03-17 15:31:28] Using 10.00% labels of the dataset (5000/50000)
[2021-03-17 15:31:28] Randomly labelled 5000/50000
[2021-03-17 15:31:28] Showing labelled: True (5000/50000 visible)
[2021-03-17 15:31:28] Running: experiment 0
[2021-03-17 15:31:28] Showing labelled: True (5000/50000 visible)
[2021-03-17 15:31:28] simple-cnn: initialized 98602 parameters.
[2021-03-17 15:31:28] Creating trainer with model on device: cuda
[2021-03-17 15:31:33] Training simple-cnn across 5000 data points in cifar10...
[2021-03-17 15:39:05] Training accuracy: 0.9169
[2021-03-17 15:39:05] Testing on 10000 data points...
[2021-03-17 15:39:06] Test score for 5000 training labels: 0.2778
[2021-03-17 15:39:06] Showing labelled: False (45000/50000 visible)
[2021-03-17 15:39:06] Found 45000 unlabelled features.
[2021-03-17 15:39:12] Computing distance between 5000 labelled and 45000 unlabelled vectors of length 64...
[2021-03-17 15:41:18] Searching for coresets with 20 beams...
[2021-03-17 16:12:10] Showing labelled: True (10000/50000 visible)
[2021-03-17 16:12:10] Training simple-cnn across 10000 data points in cifar10...
[2021-03-17 16:27:06] Training accuracy: 0.9904
[2021-03-17 16:27:06] Testing on 10000 data points...
[2021-03-17 16:27:07] Test score for 10000 training labels: 0.5695
[2021-03-17 16:27:07] Showing labelled: False (40000/50000 visible)
[2021-03-17 16:27:07] Found 40000 unlabelled features.
[2021-03-17 16:27:12] Computing distance between 10000 labelled and 40000 unlabelled vectors of length 64...
[2021-03-17 16:30:01] Searching for coresets with 20 beams...
[2021-03-17 16:57:19] Showing labelled: True (15000/50000 visible)
[2021-03-17 16:57:19] Training simple-cnn across 15000 data points in cifar10...
[2021-03-17 17:20:06] Training accuracy: 0.9898
[2021-03-17 17:20:06] Testing on 10000 data points...
[2021-03-17 17:20:07] Test score for 15000 training labels: 0.5809
[2021-03-17 17:20:07] Showing labelled: False (35000/50000 visible)
[2021-03-17 17:20:07] Found 35000 unlabelled features.
[2021-03-17 17:20:14] Computing distance between 15000 labelled and 35000 unlabelled vectors of length 64...
[2021-03-17 17:23:56] Searching for coresets with 20 beams...
[2021-03-17 17:47:48] Showing labelled: True (20000/50000 visible)
[2021-03-17 17:47:48] Training simple-cnn across 20000 data points in cifar10...
[2021-03-17 18:17:35] Training accuracy: 0.9899
[2021-03-17 18:17:35] Testing on 10000 data points...
[2021-03-17 18:17:36] Test score for 20000 training labels: 0.5995
[2021-03-17 18:17:36] Running: experiment 1
[2021-03-17 18:17:36] Showing labelled: True (5000/50000 visible)
[2021-03-17 18:17:36] simple-cnn: initialized 98602 parameters.
[2021-03-17 18:17:36] Creating trainer with model on device: cuda
[2021-03-17 18:17:36] Training simple-cnn across 5000 data points in cifar10...
[2021-03-17 18:24:55] Training accuracy: 0.9220
[2021-03-17 18:24:55] Testing on 10000 data points...
[2021-03-17 18:24:56] Test score for 5000 training labels: 0.5284
[2021-03-17 18:24:56] Showing labelled: False (45000/50000 visible)
[2021-03-17 18:24:56] Found 45000 unlabelled features.
[2021-03-17 18:25:02] Computing distance between 5000 labelled and 45000 unlabelled vectors of length 64...
[2021-03-17 18:26:36] Searching for coresets greedily...
[2021-03-17 18:27:07] Showing labelled: True (10000/50000 visible)
[2021-03-17 18:27:07] Training simple-cnn across 10000 data points in cifar10...
[2021-03-17 18:42:02] Training accuracy: 0.9902
[2021-03-17 18:42:02] Testing on 10000 data points...
[2021-03-17 18:42:03] Test score for 10000 training labels: 0.5571
[2021-03-17 18:42:03] Showing labelled: False (40000/50000 visible)
[2021-03-17 18:42:03] Found 40000 unlabelled features.
[2021-03-17 18:42:09] Computing distance between 10000 labelled and 40000 unlabelled vectors of length 64...
[2021-03-17 18:44:57] Searching for coresets greedily...
[2021-03-17 18:45:25] Showing labelled: True (15000/50000 visible)
[2021-03-17 18:45:25] Training simple-cnn across 15000 data points in cifar10...
[2021-03-17 19:07:25] Training accuracy: 0.9899
[2021-03-17 19:07:25] Testing on 10000 data points...
[2021-03-17 19:07:26] Test score for 15000 training labels: 0.5697
[2021-03-17 19:07:26] Showing labelled: False (35000/50000 visible)
[2021-03-17 19:07:26] Found 35000 unlabelled features.
[2021-03-17 19:07:33] Computing distance between 15000 labelled and 35000 unlabelled vectors of length 64...
[2021-03-17 19:11:14] Searching for coresets greedily...
[2021-03-17 19:11:38] Showing labelled: True (20000/50000 visible)
[2021-03-17 19:11:38] Training simple-cnn across 20000 data points in cifar10...
[2021-03-17 19:42:27] Training accuracy: 0.9896
[2021-03-17 19:42:27] Testing on 10000 data points...
[2021-03-17 19:42:28] Test score for 20000 training labels: 0.5816
[2021-03-17 19:42:28] Running: experiment 2
[2021-03-17 19:42:28] Showing labelled: True (5000/50000 visible)
[2021-03-17 19:42:28] simple-cnn: initialized 98602 parameters.
[2021-03-17 19:42:28] Creating trainer with model on device: cuda
[2021-03-17 19:42:28] Training simple-cnn across 5000 data points in cifar10...
[2021-03-17 19:50:11] Training accuracy: 0.9199
[2021-03-17 19:50:11] Testing on 10000 data points...
[2021-03-17 19:50:12] Test score for 5000 training labels: 0.5346
[2021-03-17 19:50:12] Showing labelled: False (45000/50000 visible)
[2021-03-17 19:50:12] Found 45000 unlabelled features.
[2021-03-17 19:50:17] Showing labelled: True (10000/50000 visible)
[2021-03-17 19:50:17] Training simple-cnn across 10000 data points in cifar10...
[2021-03-17 20:05:00] Training accuracy: 0.9904
[2021-03-17 20:05:00] Testing on 10000 data points...
[2021-03-17 20:05:02] Test score for 10000 training labels: 0.5704
[2021-03-17 20:05:02] Showing labelled: False (40000/50000 visible)
[2021-03-17 20:05:02] Found 40000 unlabelled features.
[2021-03-17 20:05:06] Showing labelled: True (15000/50000 visible)
[2021-03-17 20:05:06] Training simple-cnn across 15000 data points in cifar10...
[2021-03-17 20:27:43] Training accuracy: 0.9897
[2021-03-17 20:27:43] Testing on 10000 data points...
[2021-03-17 20:27:45] Test score for 15000 training labels: 0.5816
[2021-03-17 20:27:45] Showing labelled: False (35000/50000 visible)
[2021-03-17 20:27:45] Found 35000 unlabelled features.
[2021-03-17 20:27:48] Showing labelled: True (20000/50000 visible)
[2021-03-17 20:27:48] Training simple-cnn across 20000 data points in cifar10...
[2021-03-17 20:59:06] Training accuracy: 0.9894
[2021-03-17 20:59:06] Testing on 10000 data points...
[2021-03-17 20:59:08] Test score for 20000 training labels: 0.5954
[2021-03-17 20:59:08] Running: experiment 3
[2021-03-17 20:59:08] Showing labelled: True (5000/50000 visible)
[2021-03-17 20:59:08] simple-cnn: initialized 98602 parameters.
[2021-03-17 20:59:08] Creating trainer with model on device: cuda
[2021-03-17 20:59:08] Training simple-cnn across 5000 data points in cifar10...
[2021-03-17 21:05:38] Updated results: ../results/cifar10/lc_beam_coreset_crossentropy_beam20/results.json
