[2021-03-17 14:15:16] Seeded: 5
[2021-03-17 14:15:16] Created experiment 0:
[2021-03-17 14:15:16]  - Model: simple-cnn
[2021-03-17 14:15:16]  - Acquisition function: lc-beam-coreset
[2021-03-17 14:15:16] Created experiment 1:
[2021-03-17 14:15:16]  - Model: simple-cnn
[2021-03-17 14:15:16]  - Acquisition function: greedy-coreset
[2021-03-17 14:15:16] Created experiment 2:
[2021-03-17 14:15:16]  - Model: simple-cnn
[2021-03-17 14:15:16]  - Acquisition function: random
[2021-03-17 14:15:16] Created experiment 3:
[2021-03-17 14:15:16]  - Model: simple-cnn
[2021-03-17 14:15:16]  - Acquisition function: least-confidence
[2021-03-17 14:15:16] Loading mnist test set...
[2021-03-17 14:15:16] Experiment repeat 1/1
[2021-03-17 14:15:16] Using 0.33% labels of the dataset (200/60000)
[2021-03-17 14:15:16] Randomly labelled 200/60000
[2021-03-17 14:15:16] Showing labelled: True (200/60000 visible)
[2021-03-17 14:15:16] Running: experiment 0
[2021-03-17 14:15:16] Showing labelled: True (200/60000 visible)
[2021-03-17 14:15:16] simple-cnn: initialized 98314 parameters.
[2021-03-17 14:15:16] Creating trainer with model on device: cuda
[2021-03-17 14:15:20] Training simple-cnn across 200 data points in mnist...
[2021-03-17 14:15:40] Training accuracy: 0.6132
[2021-03-17 14:15:40] Testing on 10000 data points...
[2021-03-17 14:15:41] Test score for 200 training labels: 0.8431
[2021-03-17 14:15:41] Showing labelled: False (59800/60000 visible)
[2021-03-17 14:15:41] Found 59800 unlabelled features.
[2021-03-17 14:15:49] Computing distance between 200 labelled and 59800 unlabelled vectors of length 64...
[2021-03-17 14:15:54] Searching for coresets with 20 beams...
[2021-03-17 14:17:39] Showing labelled: True (400/60000 visible)
[2021-03-17 14:17:39] Training simple-cnn across 400 data points in mnist...
[2021-03-17 14:18:17] Training accuracy: 0.9862
[2021-03-17 14:18:17] Testing on 10000 data points...
[2021-03-17 14:18:18] Test score for 400 training labels: 0.9561
[2021-03-17 14:18:18] Showing labelled: False (59600/60000 visible)
[2021-03-17 14:18:18] Found 59600 unlabelled features.
[2021-03-17 14:18:25] Computing distance between 400 labelled and 59600 unlabelled vectors of length 64...
[2021-03-17 14:18:34] Searching for coresets with 20 beams...
[2021-03-17 14:20:20] Showing labelled: True (600/60000 visible)
[2021-03-17 14:20:20] Training simple-cnn across 600 data points in mnist...
[2021-03-17 14:21:16] Training accuracy: 0.9916
[2021-03-17 14:21:16] Testing on 10000 data points...
[2021-03-17 14:21:17] Test score for 600 training labels: 0.9752
[2021-03-17 14:21:17] Showing labelled: False (59400/60000 visible)
[2021-03-17 14:21:17] Found 59400 unlabelled features.
[2021-03-17 14:21:25] Computing distance between 600 labelled and 59400 unlabelled vectors of length 64...
[2021-03-17 14:21:42] Searching for coresets with 20 beams...
[2021-03-17 14:23:25] Showing labelled: True (800/60000 visible)
[2021-03-17 14:23:25] Training simple-cnn across 800 data points in mnist...
[2021-03-17 14:24:39] Training accuracy: 0.9949
[2021-03-17 14:24:39] Testing on 10000 data points...
[2021-03-17 14:24:40] Test score for 800 training labels: 0.9781
[2021-03-17 14:24:40] Showing labelled: False (59200/60000 visible)
[2021-03-17 14:24:40] Found 59200 unlabelled features.
[2021-03-17 14:24:48] Computing distance between 800 labelled and 59200 unlabelled vectors of length 64...
[2021-03-17 14:25:08] Searching for coresets with 20 beams...
[2021-03-17 14:26:52] Showing labelled: True (1000/60000 visible)
[2021-03-17 14:26:52] Training simple-cnn across 1000 data points in mnist...
[2021-03-17 14:28:24] Training accuracy: 0.9937
[2021-03-17 14:28:24] Testing on 10000 data points...
[2021-03-17 14:28:25] Test score for 1000 training labels: 0.9851
[2021-03-17 14:28:25] Running: experiment 1
[2021-03-17 14:28:25] Showing labelled: True (200/60000 visible)
[2021-03-17 14:28:25] simple-cnn: initialized 98314 parameters.
[2021-03-17 14:28:25] Creating trainer with model on device: cuda
[2021-03-17 14:28:25] Training simple-cnn across 200 data points in mnist...
[2021-03-17 14:28:44] Training accuracy: 0.6613
[2021-03-17 14:28:44] Testing on 10000 data points...
[2021-03-17 14:28:45] Test score for 200 training labels: 0.8906
[2021-03-17 14:28:45] Showing labelled: False (59800/60000 visible)
[2021-03-17 14:28:45] Found 59800 unlabelled features.
[2021-03-17 14:28:52] Computing distance between 200 labelled and 59800 unlabelled vectors of length 64...
[2021-03-17 14:28:57] Searching for coresets greedily...
[2021-03-17 14:28:59] Showing labelled: True (400/60000 visible)
[2021-03-17 14:28:59] Training simple-cnn across 400 data points in mnist...
[2021-03-17 14:29:35] Training accuracy: 0.9848
[2021-03-17 14:29:35] Testing on 10000 data points...
[2021-03-17 14:29:36] Test score for 400 training labels: 0.9373
[2021-03-17 14:29:36] Showing labelled: False (59600/60000 visible)
[2021-03-17 14:29:36] Found 59600 unlabelled features.
[2021-03-17 14:29:43] Computing distance between 400 labelled and 59600 unlabelled vectors of length 64...
[2021-03-17 14:29:52] Searching for coresets greedily...
[2021-03-17 14:29:54] Showing labelled: True (600/60000 visible)
[2021-03-17 14:29:54] Training simple-cnn across 600 data points in mnist...
[2021-03-17 14:30:48] Training accuracy: 0.9925
[2021-03-17 14:30:48] Testing on 10000 data points...
[2021-03-17 14:30:49] Test score for 600 training labels: 0.9788
[2021-03-17 14:30:49] Showing labelled: False (59400/60000 visible)
[2021-03-17 14:30:49] Found 59400 unlabelled features.
[2021-03-17 14:30:56] Computing distance between 600 labelled and 59400 unlabelled vectors of length 64...
[2021-03-17 14:31:12] Searching for coresets greedily...
[2021-03-17 14:31:14] Showing labelled: True (800/60000 visible)
[2021-03-17 14:31:14] Training simple-cnn across 800 data points in mnist...
[2021-03-17 14:32:30] Training accuracy: 0.9954
[2021-03-17 14:32:30] Testing on 10000 data points...
[2021-03-17 14:32:31] Test score for 800 training labels: 0.9841
[2021-03-17 14:32:31] Showing labelled: False (59200/60000 visible)
[2021-03-17 14:32:31] Found 59200 unlabelled features.
[2021-03-17 14:32:39] Computing distance between 800 labelled and 59200 unlabelled vectors of length 64...
[2021-03-17 14:32:59] Searching for coresets greedily...
[2021-03-17 14:33:00] Showing labelled: True (1000/60000 visible)
[2021-03-17 14:33:00] Training simple-cnn across 1000 data points in mnist...
[2021-03-17 14:34:36] Training accuracy: 0.9920
[2021-03-17 14:34:36] Testing on 10000 data points...
[2021-03-17 14:34:37] Test score for 1000 training labels: 0.9876
[2021-03-17 14:34:37] Running: experiment 2
[2021-03-17 14:34:37] Showing labelled: True (200/60000 visible)
[2021-03-17 14:34:37] simple-cnn: initialized 98314 parameters.
[2021-03-17 14:34:37] Creating trainer with model on device: cuda
[2021-03-17 14:34:37] Training simple-cnn across 200 data points in mnist...
[2021-03-17 14:34:56] Training accuracy: 0.6509
[2021-03-17 14:34:56] Testing on 10000 data points...
[2021-03-17 14:34:57] Test score for 200 training labels: 0.6856
[2021-03-17 14:34:57] Showing labelled: False (59800/60000 visible)
[2021-03-17 14:34:57] Found 59800 unlabelled features.
[2021-03-17 14:35:05] Showing labelled: True (400/60000 visible)
[2021-03-17 14:35:05] Training simple-cnn across 400 data points in mnist...
[2021-03-17 14:35:43] Training accuracy: 0.9962
[2021-03-17 14:35:43] Testing on 10000 data points...
[2021-03-17 14:35:44] Test score for 400 training labels: 0.9532
[2021-03-17 14:35:44] Showing labelled: False (59600/60000 visible)
[2021-03-17 14:35:44] Found 59600 unlabelled features.
[2021-03-17 14:35:50] Showing labelled: True (600/60000 visible)
[2021-03-17 14:35:50] Training simple-cnn across 600 data points in mnist...
[2021-03-17 14:36:47] Training accuracy: 0.9977
[2021-03-17 14:36:47] Testing on 10000 data points...
[2021-03-17 14:36:48] Test score for 600 training labels: 0.9643
[2021-03-17 14:36:48] Showing labelled: False (59400/60000 visible)
[2021-03-17 14:36:48] Found 59400 unlabelled features.
[2021-03-17 14:36:55] Showing labelled: True (800/60000 visible)
[2021-03-17 14:36:55] Training simple-cnn across 800 data points in mnist...
[2021-03-17 14:38:11] Training accuracy: 0.9995
[2021-03-17 14:38:11] Testing on 10000 data points...
[2021-03-17 14:38:12] Test score for 800 training labels: 0.9721
[2021-03-17 14:38:12] Showing labelled: False (59200/60000 visible)
[2021-03-17 14:38:12] Found 59200 unlabelled features.
[2021-03-17 14:38:19] Showing labelled: True (1000/60000 visible)
[2021-03-17 14:38:19] Training simple-cnn across 1000 data points in mnist...
[2021-03-17 14:39:52] Training accuracy: 0.9980
[2021-03-17 14:39:52] Testing on 10000 data points...
[2021-03-17 14:39:53] Test score for 1000 training labels: 0.9641
[2021-03-17 14:39:53] Running: experiment 3
[2021-03-17 14:39:53] Showing labelled: True (200/60000 visible)
[2021-03-17 14:39:53] simple-cnn: initialized 98314 parameters.
[2021-03-17 14:39:53] Creating trainer with model on device: cuda
[2021-03-17 14:39:53] Training simple-cnn across 200 data points in mnist...
[2021-03-17 14:40:13] Training accuracy: 0.6478
[2021-03-17 14:40:13] Testing on 10000 data points...
[2021-03-17 14:40:14] Test score for 200 training labels: 0.5438
[2021-03-17 14:40:14] Showing labelled: False (59800/60000 visible)
[2021-03-17 14:40:14] Found 59800 unlabelled features.
[2021-03-17 14:40:22] Showing labelled: True (400/60000 visible)
[2021-03-17 14:40:22] Training simple-cnn across 400 data points in mnist...
[2021-03-17 14:41:01] Training accuracy: 0.9958
[2021-03-17 14:41:01] Testing on 10000 data points...
[2021-03-17 14:41:02] Test score for 400 training labels: 0.9419
[2021-03-17 14:41:02] Showing labelled: False (59600/60000 visible)
[2021-03-17 14:41:02] Found 59600 unlabelled features.
[2021-03-17 14:41:09] Showing labelled: True (600/60000 visible)
[2021-03-17 14:41:09] Training simple-cnn across 600 data points in mnist...
[2021-03-17 14:42:03] Training accuracy: 0.9887
[2021-03-17 14:42:03] Testing on 10000 data points...
[2021-03-17 14:42:04] Test score for 600 training labels: 0.9735
[2021-03-17 14:42:04] Showing labelled: False (59400/60000 visible)
[2021-03-17 14:42:04] Found 59400 unlabelled features.
[2021-03-17 14:42:11] Showing labelled: True (800/60000 visible)
[2021-03-17 14:42:11] Training simple-cnn across 800 data points in mnist...
[2021-03-17 14:43:23] Training accuracy: 0.9921
[2021-03-17 14:43:23] Testing on 10000 data points...
[2021-03-17 14:43:24] Test score for 800 training labels: 0.9801
[2021-03-17 14:43:24] Showing labelled: False (59200/60000 visible)
[2021-03-17 14:43:24] Found 59200 unlabelled features.
[2021-03-17 14:43:32] Showing labelled: True (1000/60000 visible)
[2021-03-17 14:43:32] Training simple-cnn across 1000 data points in mnist...
[2021-03-17 14:45:05] Training accuracy: 0.9910
[2021-03-17 14:45:05] Testing on 10000 data points...
[2021-03-17 14:45:06] Test score for 1000 training labels: 0.9863
[2021-03-17 14:45:06] Updated results: ../results/mnist/lc_beam_coreset_crossentropy_beam20/results.json
