[2021-03-20 00:53:04] Created experiment 0:
[2021-03-20 00:53:04]  - Model: mlp
[2021-03-20 00:53:04]  - Acquisition function: lc-beam-coreset
[2021-03-20 00:53:04] Loading quadrant test set...
[2021-03-20 00:53:04] Saved figure to: ../data/quadrant_data.png
[2021-03-20 00:53:04] Saved figure to: ../data/quadrant_test.png
[2021-03-20 00:53:04] Experiment repeat 1/3
[2021-03-20 00:53:04] Seeded: 5
[2021-03-20 00:53:04] Using 2.00% labels of the dataset (20/1000)
[2021-03-20 00:53:04] Randomly labelled 20/1000
[2021-03-20 00:53:04] Showing labelled: True (20/1000 visible, 0 redundant)
[2021-03-20 00:53:04] Seeded: 6
[2021-03-20 00:53:04] Running: experiment 0
[2021-03-20 00:53:04] Showing labelled: True (20/1000 visible, 0 redundant)
[2021-03-20 00:53:04] Reset parameters according to He et al.
[2021-03-20 00:53:04] mlp: initialized 17412 parameters.
[2021-03-20 00:53:04] Creating trainer with model on device: cuda
[2021-03-20 00:53:09] Training mlp across 20 data points in quadrant...
[2021-03-20 00:53:10] Training accuracy: 0.8426
[2021-03-20 00:53:10] Testing on 1000 data points...
[2021-03-20 00:53:10] Test score for 20 training labels: 0.6950
[2021-03-20 00:53:10] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:10] Found 1000 unlabelled features.
[2021-03-20 00:53:10] Computing distance between 20 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:10] Searching for coresets with 30 beams...
[2021-03-20 00:53:10] Showing labelled: True (40/1000 visible, 0 redundant)
[2021-03-20 00:53:10] Reset parameters according to He et al.
[2021-03-20 00:53:10] mlp: initialized 17412 parameters.
[2021-03-20 00:53:10] Creating trainer with model on device: cuda
[2021-03-20 00:53:10] Training mlp across 40 data points in quadrant...
[2021-03-20 00:53:12] Training accuracy: 0.9443
[2021-03-20 00:53:12] Testing on 1000 data points...
[2021-03-20 00:53:12] Test score for 40 training labels: 0.9260
[2021-03-20 00:53:12] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:12] Found 1000 unlabelled features.
[2021-03-20 00:53:12] Computing distance between 40 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:12] Searching for coresets with 30 beams...
[2021-03-20 00:53:12] Showing labelled: True (60/1000 visible, 0 redundant)
[2021-03-20 00:53:12] Reset parameters according to He et al.
[2021-03-20 00:53:12] mlp: initialized 17412 parameters.
[2021-03-20 00:53:12] Creating trainer with model on device: cuda
[2021-03-20 00:53:12] Training mlp across 60 data points in quadrant...
[2021-03-20 00:53:14] Training accuracy: 0.9016
[2021-03-20 00:53:14] Testing on 1000 data points...
[2021-03-20 00:53:14] Test score for 60 training labels: 0.9640
[2021-03-20 00:53:14] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:14] Found 1000 unlabelled features.
[2021-03-20 00:53:14] Computing distance between 60 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:14] Searching for coresets with 30 beams...
[2021-03-20 00:53:14] Showing labelled: True (80/1000 visible, 0 redundant)
[2021-03-20 00:53:14] Reset parameters according to He et al.
[2021-03-20 00:53:14] mlp: initialized 17412 parameters.
[2021-03-20 00:53:14] Creating trainer with model on device: cuda
[2021-03-20 00:53:14] Training mlp across 80 data points in quadrant...
[2021-03-20 00:53:18] Training accuracy: 0.8921
[2021-03-20 00:53:18] Testing on 1000 data points...
[2021-03-20 00:53:18] Test score for 80 training labels: 0.9900
[2021-03-20 00:53:18] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:18] Found 1000 unlabelled features.
[2021-03-20 00:53:18] Computing distance between 80 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:18] Searching for coresets with 30 beams...
[2021-03-20 00:53:18] Showing labelled: True (100/1000 visible, 0 redundant)
[2021-03-20 00:53:18] Reset parameters according to He et al.
[2021-03-20 00:53:18] mlp: initialized 17412 parameters.
[2021-03-20 00:53:18] Creating trainer with model on device: cuda
[2021-03-20 00:53:18] Training mlp across 100 data points in quadrant...
[2021-03-20 00:53:22] Training accuracy: 0.9022
[2021-03-20 00:53:22] Testing on 1000 data points...
[2021-03-20 00:53:22] Test score for 100 training labels: 0.9840
[2021-03-20 00:53:22] Experiment repeat 2/3
[2021-03-20 00:53:22] Seeded: 6
[2021-03-20 00:53:22] Using 2.00% labels of the dataset (20/1000)
[2021-03-20 00:53:22] Randomly labelled 20/1000
[2021-03-20 00:53:22] Showing labelled: True (20/1000 visible, 0 redundant)
[2021-03-20 00:53:22] Seeded: 7
[2021-03-20 00:53:22] Running: experiment 0
[2021-03-20 00:53:22] Showing labelled: True (20/1000 visible, 0 redundant)
[2021-03-20 00:53:22] Reset parameters according to He et al.
[2021-03-20 00:53:22] mlp: initialized 17412 parameters.
[2021-03-20 00:53:22] Creating trainer with model on device: cuda
[2021-03-20 00:53:22] Training mlp across 20 data points in quadrant...
[2021-03-20 00:53:23] Training accuracy: 0.8700
[2021-03-20 00:53:23] Testing on 1000 data points...
[2021-03-20 00:53:23] Test score for 20 training labels: 0.8260
[2021-03-20 00:53:23] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:23] Found 1000 unlabelled features.
[2021-03-20 00:53:23] Computing distance between 20 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:23] Searching for coresets with 30 beams...
[2021-03-20 00:53:23] Showing labelled: True (40/1000 visible, 0 redundant)
[2021-03-20 00:53:23] Reset parameters according to He et al.
[2021-03-20 00:53:23] mlp: initialized 17412 parameters.
[2021-03-20 00:53:23] Creating trainer with model on device: cuda
[2021-03-20 00:53:23] Training mlp across 40 data points in quadrant...
[2021-03-20 00:53:25] Training accuracy: 0.9145
[2021-03-20 00:53:25] Testing on 1000 data points...
[2021-03-20 00:53:25] Test score for 40 training labels: 0.9560
[2021-03-20 00:53:25] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:25] Found 1000 unlabelled features.
[2021-03-20 00:53:25] Computing distance between 40 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:25] Searching for coresets with 30 beams...
[2021-03-20 00:53:25] Showing labelled: True (60/1000 visible, 0 redundant)
[2021-03-20 00:53:25] Reset parameters according to He et al.
[2021-03-20 00:53:25] mlp: initialized 17412 parameters.
[2021-03-20 00:53:25] Creating trainer with model on device: cuda
[2021-03-20 00:53:25] Training mlp across 60 data points in quadrant...
[2021-03-20 00:53:27] Training accuracy: 0.9343
[2021-03-20 00:53:27] Testing on 1000 data points...
[2021-03-20 00:53:27] Test score for 60 training labels: 0.9480
[2021-03-20 00:53:27] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:27] Found 1000 unlabelled features.
[2021-03-20 00:53:27] Computing distance between 60 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:27] Searching for coresets with 30 beams...
[2021-03-20 00:53:27] Showing labelled: True (80/1000 visible, 0 redundant)
[2021-03-20 00:53:27] Reset parameters according to He et al.
[2021-03-20 00:53:27] mlp: initialized 17412 parameters.
[2021-03-20 00:53:27] Creating trainer with model on device: cuda
[2021-03-20 00:53:27] Training mlp across 80 data points in quadrant...
[2021-03-20 00:53:30] Training accuracy: 0.9452
[2021-03-20 00:53:30] Testing on 1000 data points...
[2021-03-20 00:53:30] Test score for 80 training labels: 0.9530
[2021-03-20 00:53:30] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:30] Found 1000 unlabelled features.
[2021-03-20 00:53:30] Computing distance between 80 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:30] Searching for coresets with 30 beams...
[2021-03-20 00:53:30] Showing labelled: True (100/1000 visible, 0 redundant)
[2021-03-20 00:53:30] Reset parameters according to He et al.
[2021-03-20 00:53:30] mlp: initialized 17412 parameters.
[2021-03-20 00:53:30] Creating trainer with model on device: cuda
[2021-03-20 00:53:30] Training mlp across 100 data points in quadrant...
[2021-03-20 00:53:34] Training accuracy: 0.9456
[2021-03-20 00:53:34] Testing on 1000 data points...
[2021-03-20 00:53:34] Test score for 100 training labels: 0.9720
[2021-03-20 00:53:34] Experiment repeat 3/3
[2021-03-20 00:53:34] Seeded: 7
[2021-03-20 00:53:34] Using 2.00% labels of the dataset (20/1000)
[2021-03-20 00:53:34] Randomly labelled 20/1000
[2021-03-20 00:53:34] Showing labelled: True (20/1000 visible, 0 redundant)
[2021-03-20 00:53:34] Seeded: 8
[2021-03-20 00:53:34] Running: experiment 0
[2021-03-20 00:53:34] Showing labelled: True (20/1000 visible, 0 redundant)
[2021-03-20 00:53:34] Reset parameters according to He et al.
[2021-03-20 00:53:34] mlp: initialized 17412 parameters.
[2021-03-20 00:53:34] Creating trainer with model on device: cuda
[2021-03-20 00:53:34] Training mlp across 20 data points in quadrant...
[2021-03-20 00:53:35] Training accuracy: 0.8209
[2021-03-20 00:53:35] Testing on 1000 data points...
[2021-03-20 00:53:35] Test score for 20 training labels: 0.7920
[2021-03-20 00:53:35] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:35] Found 1000 unlabelled features.
[2021-03-20 00:53:35] Computing distance between 20 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:35] Searching for coresets with 30 beams...
[2021-03-20 00:53:35] Showing labelled: True (40/1000 visible, 0 redundant)
[2021-03-20 00:53:35] Reset parameters according to He et al.
[2021-03-20 00:53:35] mlp: initialized 17412 parameters.
[2021-03-20 00:53:35] Creating trainer with model on device: cuda
[2021-03-20 00:53:35] Training mlp across 40 data points in quadrant...
[2021-03-20 00:53:37] Training accuracy: 0.9032
[2021-03-20 00:53:37] Testing on 1000 data points...
[2021-03-20 00:53:37] Test score for 40 training labels: 0.9540
[2021-03-20 00:53:37] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:37] Found 1000 unlabelled features.
[2021-03-20 00:53:37] Computing distance between 40 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:37] Searching for coresets with 30 beams...
[2021-03-20 00:53:37] Showing labelled: True (60/1000 visible, 0 redundant)
[2021-03-20 00:53:37] Reset parameters according to He et al.
[2021-03-20 00:53:37] mlp: initialized 17412 parameters.
[2021-03-20 00:53:37] Creating trainer with model on device: cuda
[2021-03-20 00:53:37] Training mlp across 60 data points in quadrant...
[2021-03-20 00:53:39] Training accuracy: 0.8633
[2021-03-20 00:53:39] Testing on 1000 data points...
[2021-03-20 00:53:39] Test score for 60 training labels: 0.9460
[2021-03-20 00:53:39] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:39] Found 1000 unlabelled features.
[2021-03-20 00:53:39] Computing distance between 60 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:39] Searching for coresets with 30 beams...
[2021-03-20 00:53:39] Showing labelled: True (80/1000 visible, 0 redundant)
[2021-03-20 00:53:39] Reset parameters according to He et al.
[2021-03-20 00:53:39] mlp: initialized 17412 parameters.
[2021-03-20 00:53:39] Creating trainer with model on device: cuda
[2021-03-20 00:53:39] Training mlp across 80 data points in quadrant...
[2021-03-20 00:53:43] Training accuracy: 0.9264
[2021-03-20 00:53:43] Testing on 1000 data points...
[2021-03-20 00:53:43] Test score for 80 training labels: 0.9650
[2021-03-20 00:53:43] Showing labelled: False (1000/1000 visible, 0 redundant)
[2021-03-20 00:53:43] Found 1000 unlabelled features.
[2021-03-20 00:53:43] Computing distance between 80 labelled and 1000 unlabelled vectors of length 128...
[2021-03-20 00:53:43] Searching for coresets with 30 beams...
[2021-03-20 00:53:43] Showing labelled: True (100/1000 visible, 0 redundant)
[2021-03-20 00:53:43] Reset parameters according to He et al.
[2021-03-20 00:53:43] mlp: initialized 17412 parameters.
[2021-03-20 00:53:43] Creating trainer with model on device: cuda
[2021-03-20 00:53:43] Training mlp across 100 data points in quadrant...
[2021-03-20 00:53:46] Training accuracy: 0.9100
[2021-03-20 00:53:46] Testing on 1000 data points...
[2021-03-20 00:53:46] Test score for 100 training labels: 0.9780
[2021-03-20 00:53:46] Updated results: ../results/quadrant/mine_lcbeam/results.json
