[2021-03-20 01:11:23] Created experiment 0:
[2021-03-20 01:11:23]  - Model: mlp
[2021-03-20 01:11:23]  - Acquisition function: random
[2021-03-20 01:11:23] Created experiment 1:
[2021-03-20 01:11:23]  - Model: mlp
[2021-03-20 01:11:23]  - Acquisition function: least-confidence
[2021-03-20 01:11:23] Created experiment 2:
[2021-03-20 01:11:23]  - Model: mlp
[2021-03-20 01:11:23]  - Acquisition function: lc-beam-coreset
[2021-03-20 01:11:23] Created experiment 3:
[2021-03-20 01:11:23]  - Model: mlp
[2021-03-20 01:11:23]  - Acquisition function: greedy-coreset
[2021-03-20 01:11:23] Loading blobsthin-balanced test set...
[2021-03-20 01:11:23] Generated 5000 blobs:
[2021-03-20 01:11:23] Class 0: 500
[2021-03-20 01:11:23] Class 1: 500
[2021-03-20 01:11:23] Class 2: 500
[2021-03-20 01:11:23] Class 3: 500
[2021-03-20 01:11:23] Class 4: 500
[2021-03-20 01:11:23] Class 5: 500
[2021-03-20 01:11:23] Class 6: 500
[2021-03-20 01:11:23] Class 7: 500
[2021-03-20 01:11:23] Class 8: 500
[2021-03-20 01:11:23] Class 9: 500
[2021-03-20 01:11:24] Saved figure to: ../data/blobsthin-balanced_data.png
[2021-03-20 01:11:24] Generated 5000 test points.
[2021-03-20 01:11:24] Saved figure to: ../data/blobsthin-balanced_test.png
[2021-03-20 01:11:24] Experiment repeat 1/3
[2021-03-20 01:11:24] Seeded: 5
[2021-03-20 01:11:24] Using 1.00% labels of the dataset (50/5000)
[2021-03-20 01:11:24] Randomly labelled 50/5000
[2021-03-20 01:11:24] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:11:24] Seeded: 6
[2021-03-20 01:11:24] Running: experiment 0
[2021-03-20 01:11:24] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:11:24] Reset parameters according to He et al.
[2021-03-20 01:11:24] mlp: initialized 18186 parameters.
[2021-03-20 01:11:24] Creating trainer with model on device: cuda
[2021-03-20 01:11:29] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:11:29] Training accuracy: 0.7241
[2021-03-20 01:11:29] Testing on 5000 data points...
[2021-03-20 01:11:29] Test score for 50 training labels: 0.7154
[2021-03-20 01:11:29] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:29] Found 5000 unlabelled features.
[2021-03-20 01:11:29] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:11:29] Reset parameters according to He et al.
[2021-03-20 01:11:29] mlp: initialized 18186 parameters.
[2021-03-20 01:11:29] Creating trainer with model on device: cuda
[2021-03-20 01:11:29] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:11:30] Training accuracy: 0.7577
[2021-03-20 01:11:30] Testing on 5000 data points...
[2021-03-20 01:11:31] Test score for 100 training labels: 0.8918
[2021-03-20 01:11:31] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:31] Found 5000 unlabelled features.
[2021-03-20 01:11:31] Showing labelled: True (149/5000 visible, 1 redundant)
[2021-03-20 01:11:31] Reset parameters according to He et al.
[2021-03-20 01:11:31] mlp: initialized 18186 parameters.
[2021-03-20 01:11:31] Creating trainer with model on device: cuda
[2021-03-20 01:11:31] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:11:32] Training accuracy: 0.8089
[2021-03-20 01:11:32] Testing on 5000 data points...
[2021-03-20 01:11:33] Test score for 150 training labels: 0.9302
[2021-03-20 01:11:33] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:33] Found 5000 unlabelled features.
[2021-03-20 01:11:33] Showing labelled: True (197/5000 visible, 3 redundant)
[2021-03-20 01:11:33] Reset parameters according to He et al.
[2021-03-20 01:11:33] mlp: initialized 18186 parameters.
[2021-03-20 01:11:33] Creating trainer with model on device: cuda
[2021-03-20 01:11:33] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:11:35] Training accuracy: 0.8415
[2021-03-20 01:11:35] Testing on 5000 data points...
[2021-03-20 01:11:35] Test score for 200 training labels: 0.9094
[2021-03-20 01:11:35] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:35] Found 5000 unlabelled features.
[2021-03-20 01:11:35] Showing labelled: True (245/5000 visible, 5 redundant)
[2021-03-20 01:11:35] Reset parameters according to He et al.
[2021-03-20 01:11:35] mlp: initialized 18186 parameters.
[2021-03-20 01:11:35] Creating trainer with model on device: cuda
[2021-03-20 01:11:35] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:11:38] Training accuracy: 0.8608
[2021-03-20 01:11:38] Testing on 5000 data points...
[2021-03-20 01:11:38] Test score for 250 training labels: 0.9170
[2021-03-20 01:11:38] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:38] Found 5000 unlabelled features.
[2021-03-20 01:11:38] Showing labelled: True (289/5000 visible, 11 redundant)
[2021-03-20 01:11:38] Reset parameters according to He et al.
[2021-03-20 01:11:38] mlp: initialized 18186 parameters.
[2021-03-20 01:11:38] Creating trainer with model on device: cuda
[2021-03-20 01:11:38] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:11:42] Training accuracy: 0.8942
[2021-03-20 01:11:42] Testing on 5000 data points...
[2021-03-20 01:11:42] Test score for 300 training labels: 0.9034
[2021-03-20 01:11:42] Seeded: 6
[2021-03-20 01:11:42] Running: experiment 1
[2021-03-20 01:11:42] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:11:42] Reset parameters according to He et al.
[2021-03-20 01:11:42] mlp: initialized 18186 parameters.
[2021-03-20 01:11:42] Creating trainer with model on device: cuda
[2021-03-20 01:11:42] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:11:42] Training accuracy: 0.7241
[2021-03-20 01:11:42] Testing on 5000 data points...
[2021-03-20 01:11:42] Test score for 50 training labels: 0.7154
[2021-03-20 01:11:42] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:42] Found 5000 unlabelled features.
[2021-03-20 01:11:42] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:11:42] Reset parameters according to He et al.
[2021-03-20 01:11:42] mlp: initialized 18186 parameters.
[2021-03-20 01:11:42] Creating trainer with model on device: cuda
[2021-03-20 01:11:42] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:11:43] Training accuracy: 0.8621
[2021-03-20 01:11:43] Testing on 5000 data points...
[2021-03-20 01:11:44] Test score for 100 training labels: 0.6242
[2021-03-20 01:11:44] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:44] Found 5000 unlabelled features.
[2021-03-20 01:11:44] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:11:44] Reset parameters according to He et al.
[2021-03-20 01:11:44] mlp: initialized 18186 parameters.
[2021-03-20 01:11:44] Creating trainer with model on device: cuda
[2021-03-20 01:11:44] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:11:45] Training accuracy: 0.9133
[2021-03-20 01:11:45] Testing on 5000 data points...
[2021-03-20 01:11:46] Test score for 150 training labels: 0.6536
[2021-03-20 01:11:46] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:46] Found 5000 unlabelled features.
[2021-03-20 01:11:46] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:11:46] Reset parameters according to He et al.
[2021-03-20 01:11:46] mlp: initialized 18186 parameters.
[2021-03-20 01:11:46] Creating trainer with model on device: cuda
[2021-03-20 01:11:46] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:11:48] Training accuracy: 0.9279
[2021-03-20 01:11:48] Testing on 5000 data points...
[2021-03-20 01:11:48] Test score for 200 training labels: 0.7840
[2021-03-20 01:11:48] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:48] Found 5000 unlabelled features.
[2021-03-20 01:11:48] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:11:48] Reset parameters according to He et al.
[2021-03-20 01:11:48] mlp: initialized 18186 parameters.
[2021-03-20 01:11:48] Creating trainer with model on device: cuda
[2021-03-20 01:11:48] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:11:51] Training accuracy: 0.9547
[2021-03-20 01:11:51] Testing on 5000 data points...
[2021-03-20 01:11:51] Test score for 250 training labels: 0.7934
[2021-03-20 01:11:51] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:51] Found 5000 unlabelled features.
[2021-03-20 01:11:51] Showing labelled: True (298/5000 visible, 2 redundant)
[2021-03-20 01:11:51] Reset parameters according to He et al.
[2021-03-20 01:11:51] mlp: initialized 18186 parameters.
[2021-03-20 01:11:51] Creating trainer with model on device: cuda
[2021-03-20 01:11:51] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:11:54] Training accuracy: 0.9412
[2021-03-20 01:11:54] Testing on 5000 data points...
[2021-03-20 01:11:54] Test score for 300 training labels: 0.7998
[2021-03-20 01:11:54] Seeded: 6
[2021-03-20 01:11:54] Running: experiment 2
[2021-03-20 01:11:54] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:11:54] Reset parameters according to He et al.
[2021-03-20 01:11:54] mlp: initialized 18186 parameters.
[2021-03-20 01:11:54] Creating trainer with model on device: cuda
[2021-03-20 01:11:54] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:11:55] Training accuracy: 0.7241
[2021-03-20 01:11:55] Testing on 5000 data points...
[2021-03-20 01:11:55] Test score for 50 training labels: 0.7154
[2021-03-20 01:11:55] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:55] Found 5000 unlabelled features.
[2021-03-20 01:11:55] Computing distance between 50 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:11:55] Searching for coresets with 20 beams...
[2021-03-20 01:11:55] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:11:55] Reset parameters according to He et al.
[2021-03-20 01:11:55] mlp: initialized 18186 parameters.
[2021-03-20 01:11:55] Creating trainer with model on device: cuda
[2021-03-20 01:11:55] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:11:56] Training accuracy: 0.7253
[2021-03-20 01:11:56] Testing on 5000 data points...
[2021-03-20 01:11:56] Test score for 100 training labels: 0.8788
[2021-03-20 01:11:56] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:56] Found 5000 unlabelled features.
[2021-03-20 01:11:57] Computing distance between 100 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:11:57] Searching for coresets with 20 beams...
[2021-03-20 01:11:57] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:11:57] Reset parameters according to He et al.
[2021-03-20 01:11:57] mlp: initialized 18186 parameters.
[2021-03-20 01:11:57] Creating trainer with model on device: cuda
[2021-03-20 01:11:57] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:11:59] Training accuracy: 0.7813
[2021-03-20 01:11:59] Testing on 5000 data points...
[2021-03-20 01:11:59] Test score for 150 training labels: 0.8442
[2021-03-20 01:11:59] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:11:59] Found 5000 unlabelled features.
[2021-03-20 01:11:59] Computing distance between 150 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:11:59] Searching for coresets with 20 beams...
[2021-03-20 01:11:59] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:11:59] Reset parameters according to He et al.
[2021-03-20 01:11:59] mlp: initialized 18186 parameters.
[2021-03-20 01:11:59] Creating trainer with model on device: cuda
[2021-03-20 01:11:59] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:12:01] Training accuracy: 0.8164
[2021-03-20 01:12:01] Testing on 5000 data points...
[2021-03-20 01:12:01] Test score for 200 training labels: 0.9590
[2021-03-20 01:12:01] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:01] Found 5000 unlabelled features.
[2021-03-20 01:12:01] Computing distance between 200 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:01] Searching for coresets with 20 beams...
[2021-03-20 01:12:02] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:12:02] Reset parameters according to He et al.
[2021-03-20 01:12:02] mlp: initialized 18186 parameters.
[2021-03-20 01:12:02] Creating trainer with model on device: cuda
[2021-03-20 01:12:02] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:12:05] Training accuracy: 0.8393
[2021-03-20 01:12:05] Testing on 5000 data points...
[2021-03-20 01:12:05] Test score for 250 training labels: 0.9786
[2021-03-20 01:12:05] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:05] Found 5000 unlabelled features.
[2021-03-20 01:12:05] Computing distance between 250 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:05] Searching for coresets with 20 beams...
[2021-03-20 01:12:05] Showing labelled: True (300/5000 visible, 0 redundant)
[2021-03-20 01:12:05] Reset parameters according to He et al.
[2021-03-20 01:12:05] mlp: initialized 18186 parameters.
[2021-03-20 01:12:05] Creating trainer with model on device: cuda
[2021-03-20 01:12:05] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:12:09] Training accuracy: 0.8616
[2021-03-20 01:12:09] Testing on 5000 data points...
[2021-03-20 01:12:09] Test score for 300 training labels: 0.9924
[2021-03-20 01:12:09] Seeded: 6
[2021-03-20 01:12:09] Running: experiment 3
[2021-03-20 01:12:09] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:12:09] Reset parameters according to He et al.
[2021-03-20 01:12:09] mlp: initialized 18186 parameters.
[2021-03-20 01:12:09] Creating trainer with model on device: cuda
[2021-03-20 01:12:09] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:12:10] Training accuracy: 0.7241
[2021-03-20 01:12:10] Testing on 5000 data points...
[2021-03-20 01:12:10] Test score for 50 training labels: 0.7154
[2021-03-20 01:12:10] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:10] Found 5000 unlabelled features.
[2021-03-20 01:12:10] Computing distance between 50 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:10] Searching for coresets greedily...
[2021-03-20 01:12:10] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:12:10] Reset parameters according to He et al.
[2021-03-20 01:12:10] mlp: initialized 18186 parameters.
[2021-03-20 01:12:10] Creating trainer with model on device: cuda
[2021-03-20 01:12:10] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:12:11] Training accuracy: 0.7438
[2021-03-20 01:12:11] Testing on 5000 data points...
[2021-03-20 01:12:11] Test score for 100 training labels: 0.8980
[2021-03-20 01:12:11] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:11] Found 5000 unlabelled features.
[2021-03-20 01:12:11] Computing distance between 100 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:11] Searching for coresets greedily...
[2021-03-20 01:12:11] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:12:11] Reset parameters according to He et al.
[2021-03-20 01:12:11] mlp: initialized 18186 parameters.
[2021-03-20 01:12:11] Creating trainer with model on device: cuda
[2021-03-20 01:12:11] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:12:13] Training accuracy: 0.7722
[2021-03-20 01:12:13] Testing on 5000 data points...
[2021-03-20 01:12:13] Test score for 150 training labels: 0.7980
[2021-03-20 01:12:13] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:13] Found 5000 unlabelled features.
[2021-03-20 01:12:13] Computing distance between 150 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:13] Searching for coresets greedily...
[2021-03-20 01:12:13] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:12:13] Reset parameters according to He et al.
[2021-03-20 01:12:13] mlp: initialized 18186 parameters.
[2021-03-20 01:12:13] Creating trainer with model on device: cuda
[2021-03-20 01:12:13] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:12:15] Training accuracy: 0.8209
[2021-03-20 01:12:15] Testing on 5000 data points...
[2021-03-20 01:12:15] Test score for 200 training labels: 0.9526
[2021-03-20 01:12:15] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:15] Found 5000 unlabelled features.
[2021-03-20 01:12:15] Computing distance between 200 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:15] Searching for coresets greedily...
[2021-03-20 01:12:15] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:12:15] Reset parameters according to He et al.
[2021-03-20 01:12:15] mlp: initialized 18186 parameters.
[2021-03-20 01:12:15] Creating trainer with model on device: cuda
[2021-03-20 01:12:15] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:12:18] Training accuracy: 0.8203
[2021-03-20 01:12:18] Testing on 5000 data points...
[2021-03-20 01:12:19] Test score for 250 training labels: 0.9696
[2021-03-20 01:12:19] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:19] Found 5000 unlabelled features.
[2021-03-20 01:12:19] Computing distance between 250 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:19] Searching for coresets greedily...
[2021-03-20 01:12:19] Showing labelled: True (300/5000 visible, 0 redundant)
[2021-03-20 01:12:19] Reset parameters according to He et al.
[2021-03-20 01:12:19] mlp: initialized 18186 parameters.
[2021-03-20 01:12:19] Creating trainer with model on device: cuda
[2021-03-20 01:12:19] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:12:22] Training accuracy: 0.8547
[2021-03-20 01:12:22] Testing on 5000 data points...
[2021-03-20 01:12:22] Test score for 300 training labels: 0.9892
[2021-03-20 01:12:22] Experiment repeat 2/3
[2021-03-20 01:12:22] Seeded: 6
[2021-03-20 01:12:22] Using 1.00% labels of the dataset (50/5000)
[2021-03-20 01:12:22] Randomly labelled 50/5000
[2021-03-20 01:12:22] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:12:22] Seeded: 7
[2021-03-20 01:12:22] Running: experiment 0
[2021-03-20 01:12:22] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:12:22] Reset parameters according to He et al.
[2021-03-20 01:12:22] mlp: initialized 18186 parameters.
[2021-03-20 01:12:22] Creating trainer with model on device: cuda
[2021-03-20 01:12:22] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:12:23] Training accuracy: 0.6273
[2021-03-20 01:12:23] Testing on 5000 data points...
[2021-03-20 01:12:23] Test score for 50 training labels: 0.8020
[2021-03-20 01:12:23] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:23] Found 5000 unlabelled features.
[2021-03-20 01:12:23] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:12:23] Reset parameters according to He et al.
[2021-03-20 01:12:23] mlp: initialized 18186 parameters.
[2021-03-20 01:12:23] Creating trainer with model on device: cuda
[2021-03-20 01:12:23] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:12:24] Training accuracy: 0.7135
[2021-03-20 01:12:24] Testing on 5000 data points...
[2021-03-20 01:12:24] Test score for 100 training labels: 0.8922
[2021-03-20 01:12:24] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:24] Found 5000 unlabelled features.
[2021-03-20 01:12:24] Showing labelled: True (149/5000 visible, 1 redundant)
[2021-03-20 01:12:24] Reset parameters according to He et al.
[2021-03-20 01:12:24] mlp: initialized 18186 parameters.
[2021-03-20 01:12:24] Creating trainer with model on device: cuda
[2021-03-20 01:12:24] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:12:26] Training accuracy: 0.7499
[2021-03-20 01:12:26] Testing on 5000 data points...
[2021-03-20 01:12:26] Test score for 150 training labels: 0.8992
[2021-03-20 01:12:26] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:26] Found 5000 unlabelled features.
[2021-03-20 01:12:26] Showing labelled: True (199/5000 visible, 1 redundant)
[2021-03-20 01:12:26] Reset parameters according to He et al.
[2021-03-20 01:12:26] mlp: initialized 18186 parameters.
[2021-03-20 01:12:26] Creating trainer with model on device: cuda
[2021-03-20 01:12:26] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:12:28] Training accuracy: 0.8412
[2021-03-20 01:12:28] Testing on 5000 data points...
[2021-03-20 01:12:28] Test score for 200 training labels: 0.9672
[2021-03-20 01:12:28] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:28] Found 5000 unlabelled features.
[2021-03-20 01:12:28] Showing labelled: True (246/5000 visible, 4 redundant)
[2021-03-20 01:12:28] Reset parameters according to He et al.
[2021-03-20 01:12:28] mlp: initialized 18186 parameters.
[2021-03-20 01:12:28] Creating trainer with model on device: cuda
[2021-03-20 01:12:28] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:12:31] Training accuracy: 0.8475
[2021-03-20 01:12:31] Testing on 5000 data points...
[2021-03-20 01:12:31] Test score for 250 training labels: 0.9734
[2021-03-20 01:12:31] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:31] Found 5000 unlabelled features.
[2021-03-20 01:12:31] Showing labelled: True (293/5000 visible, 7 redundant)
[2021-03-20 01:12:31] Reset parameters according to He et al.
[2021-03-20 01:12:31] mlp: initialized 18186 parameters.
[2021-03-20 01:12:31] Creating trainer with model on device: cuda
[2021-03-20 01:12:31] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:12:34] Training accuracy: 0.8811
[2021-03-20 01:12:34] Testing on 5000 data points...
[2021-03-20 01:12:35] Test score for 300 training labels: 0.9684
[2021-03-20 01:12:35] Seeded: 7
[2021-03-20 01:12:35] Running: experiment 1
[2021-03-20 01:12:35] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:12:35] Reset parameters according to He et al.
[2021-03-20 01:12:35] mlp: initialized 18186 parameters.
[2021-03-20 01:12:35] Creating trainer with model on device: cuda
[2021-03-20 01:12:35] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:12:35] Training accuracy: 0.6273
[2021-03-20 01:12:35] Testing on 5000 data points...
[2021-03-20 01:12:35] Test score for 50 training labels: 0.8020
[2021-03-20 01:12:35] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:35] Found 5000 unlabelled features.
[2021-03-20 01:12:35] Showing labelled: True (99/5000 visible, 1 redundant)
[2021-03-20 01:12:35] Reset parameters according to He et al.
[2021-03-20 01:12:35] mlp: initialized 18186 parameters.
[2021-03-20 01:12:35] Creating trainer with model on device: cuda
[2021-03-20 01:12:35] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:12:36] Training accuracy: 0.8190
[2021-03-20 01:12:36] Testing on 5000 data points...
[2021-03-20 01:12:36] Test score for 100 training labels: 0.8974
[2021-03-20 01:12:36] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:36] Found 5000 unlabelled features.
[2021-03-20 01:12:36] Showing labelled: True (147/5000 visible, 3 redundant)
[2021-03-20 01:12:36] Reset parameters according to He et al.
[2021-03-20 01:12:36] mlp: initialized 18186 parameters.
[2021-03-20 01:12:36] Creating trainer with model on device: cuda
[2021-03-20 01:12:36] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:12:38] Training accuracy: 0.8751
[2021-03-20 01:12:38] Testing on 5000 data points...
[2021-03-20 01:12:38] Test score for 150 training labels: 0.8000
[2021-03-20 01:12:38] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:38] Found 5000 unlabelled features.
[2021-03-20 01:12:38] Showing labelled: True (194/5000 visible, 6 redundant)
[2021-03-20 01:12:38] Reset parameters according to He et al.
[2021-03-20 01:12:38] mlp: initialized 18186 parameters.
[2021-03-20 01:12:38] Creating trainer with model on device: cuda
[2021-03-20 01:12:38] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:12:41] Training accuracy: 0.8402
[2021-03-20 01:12:41] Testing on 5000 data points...
[2021-03-20 01:12:41] Test score for 200 training labels: 0.8070
[2021-03-20 01:12:41] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:41] Found 5000 unlabelled features.
[2021-03-20 01:12:41] Showing labelled: True (219/5000 visible, 31 redundant)
[2021-03-20 01:12:41] Reset parameters according to He et al.
[2021-03-20 01:12:41] mlp: initialized 18186 parameters.
[2021-03-20 01:12:41] Creating trainer with model on device: cuda
[2021-03-20 01:12:41] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:12:44] Training accuracy: 0.8197
[2021-03-20 01:12:44] Testing on 5000 data points...
[2021-03-20 01:12:44] Test score for 250 training labels: 0.8620
[2021-03-20 01:12:44] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:44] Found 5000 unlabelled features.
[2021-03-20 01:12:44] Showing labelled: True (269/5000 visible, 31 redundant)
[2021-03-20 01:12:44] Reset parameters according to He et al.
[2021-03-20 01:12:44] mlp: initialized 18186 parameters.
[2021-03-20 01:12:44] Creating trainer with model on device: cuda
[2021-03-20 01:12:44] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:12:47] Training accuracy: 0.7823
[2021-03-20 01:12:47] Testing on 5000 data points...
[2021-03-20 01:12:48] Test score for 300 training labels: 0.8832
[2021-03-20 01:12:48] Seeded: 7
[2021-03-20 01:12:48] Running: experiment 2
[2021-03-20 01:12:48] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:12:48] Reset parameters according to He et al.
[2021-03-20 01:12:48] mlp: initialized 18186 parameters.
[2021-03-20 01:12:48] Creating trainer with model on device: cuda
[2021-03-20 01:12:48] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:12:48] Training accuracy: 0.6273
[2021-03-20 01:12:48] Testing on 5000 data points...
[2021-03-20 01:12:48] Test score for 50 training labels: 0.8020
[2021-03-20 01:12:48] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:48] Found 5000 unlabelled features.
[2021-03-20 01:12:48] Computing distance between 50 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:48] Searching for coresets with 20 beams...
[2021-03-20 01:12:48] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:12:48] Reset parameters according to He et al.
[2021-03-20 01:12:48] mlp: initialized 18186 parameters.
[2021-03-20 01:12:48] Creating trainer with model on device: cuda
[2021-03-20 01:12:48] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:12:50] Training accuracy: 0.7528
[2021-03-20 01:12:50] Testing on 5000 data points...
[2021-03-20 01:12:50] Test score for 100 training labels: 0.8900
[2021-03-20 01:12:50] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:50] Found 5000 unlabelled features.
[2021-03-20 01:12:50] Computing distance between 100 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:50] Searching for coresets with 20 beams...
[2021-03-20 01:12:50] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:12:50] Reset parameters according to He et al.
[2021-03-20 01:12:50] mlp: initialized 18186 parameters.
[2021-03-20 01:12:50] Creating trainer with model on device: cuda
[2021-03-20 01:12:50] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:12:52] Training accuracy: 0.7473
[2021-03-20 01:12:52] Testing on 5000 data points...
[2021-03-20 01:12:52] Test score for 150 training labels: 0.8858
[2021-03-20 01:12:52] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:52] Found 5000 unlabelled features.
[2021-03-20 01:12:52] Computing distance between 150 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:52] Searching for coresets with 20 beams...
[2021-03-20 01:12:52] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:12:52] Reset parameters according to He et al.
[2021-03-20 01:12:52] mlp: initialized 18186 parameters.
[2021-03-20 01:12:52] Creating trainer with model on device: cuda
[2021-03-20 01:12:52] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:12:54] Training accuracy: 0.8051
[2021-03-20 01:12:54] Testing on 5000 data points...
[2021-03-20 01:12:54] Test score for 200 training labels: 0.9732
[2021-03-20 01:12:54] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:54] Found 5000 unlabelled features.
[2021-03-20 01:12:54] Computing distance between 200 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:54] Searching for coresets with 20 beams...
[2021-03-20 01:12:55] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:12:55] Reset parameters according to He et al.
[2021-03-20 01:12:55] mlp: initialized 18186 parameters.
[2021-03-20 01:12:55] Creating trainer with model on device: cuda
[2021-03-20 01:12:55] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:12:57] Training accuracy: 0.8349
[2021-03-20 01:12:57] Testing on 5000 data points...
[2021-03-20 01:12:57] Test score for 250 training labels: 0.9868
[2021-03-20 01:12:57] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:12:57] Found 5000 unlabelled features.
[2021-03-20 01:12:57] Computing distance between 250 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:12:58] Searching for coresets with 20 beams...
[2021-03-20 01:12:58] Showing labelled: True (300/5000 visible, 0 redundant)
[2021-03-20 01:12:58] Reset parameters according to He et al.
[2021-03-20 01:12:58] mlp: initialized 18186 parameters.
[2021-03-20 01:12:58] Creating trainer with model on device: cuda
[2021-03-20 01:12:58] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:13:01] Training accuracy: 0.8643
[2021-03-20 01:13:01] Testing on 5000 data points...
[2021-03-20 01:13:01] Test score for 300 training labels: 0.9702
[2021-03-20 01:13:01] Seeded: 7
[2021-03-20 01:13:01] Running: experiment 3
[2021-03-20 01:13:01] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:13:01] Reset parameters according to He et al.
[2021-03-20 01:13:01] mlp: initialized 18186 parameters.
[2021-03-20 01:13:01] Creating trainer with model on device: cuda
[2021-03-20 01:13:01] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:13:02] Training accuracy: 0.6273
[2021-03-20 01:13:02] Testing on 5000 data points...
[2021-03-20 01:13:02] Test score for 50 training labels: 0.8020
[2021-03-20 01:13:02] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:02] Found 5000 unlabelled features.
[2021-03-20 01:13:02] Computing distance between 50 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:02] Searching for coresets greedily...
[2021-03-20 01:13:02] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:13:02] Reset parameters according to He et al.
[2021-03-20 01:13:02] mlp: initialized 18186 parameters.
[2021-03-20 01:13:02] Creating trainer with model on device: cuda
[2021-03-20 01:13:02] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:13:03] Training accuracy: 0.7544
[2021-03-20 01:13:03] Testing on 5000 data points...
[2021-03-20 01:13:03] Test score for 100 training labels: 0.8316
[2021-03-20 01:13:03] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:03] Found 5000 unlabelled features.
[2021-03-20 01:13:03] Computing distance between 100 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:03] Searching for coresets greedily...
[2021-03-20 01:13:03] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:13:03] Reset parameters according to He et al.
[2021-03-20 01:13:03] mlp: initialized 18186 parameters.
[2021-03-20 01:13:03] Creating trainer with model on device: cuda
[2021-03-20 01:13:03] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:13:05] Training accuracy: 0.7656
[2021-03-20 01:13:05] Testing on 5000 data points...
[2021-03-20 01:13:05] Test score for 150 training labels: 0.8846
[2021-03-20 01:13:05] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:05] Found 5000 unlabelled features.
[2021-03-20 01:13:05] Computing distance between 150 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:05] Searching for coresets greedily...
[2021-03-20 01:13:05] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:13:05] Reset parameters according to He et al.
[2021-03-20 01:13:05] mlp: initialized 18186 parameters.
[2021-03-20 01:13:05] Creating trainer with model on device: cuda
[2021-03-20 01:13:05] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:13:08] Training accuracy: 0.8102
[2021-03-20 01:13:08] Testing on 5000 data points...
[2021-03-20 01:13:08] Test score for 200 training labels: 0.9206
[2021-03-20 01:13:08] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:08] Found 5000 unlabelled features.
[2021-03-20 01:13:08] Computing distance between 200 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:08] Searching for coresets greedily...
[2021-03-20 01:13:08] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:13:08] Reset parameters according to He et al.
[2021-03-20 01:13:08] mlp: initialized 18186 parameters.
[2021-03-20 01:13:08] Creating trainer with model on device: cuda
[2021-03-20 01:13:08] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:13:11] Training accuracy: 0.8641
[2021-03-20 01:13:11] Testing on 5000 data points...
[2021-03-20 01:13:11] Test score for 250 training labels: 0.9610
[2021-03-20 01:13:11] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:11] Found 5000 unlabelled features.
[2021-03-20 01:13:11] Computing distance between 250 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:11] Searching for coresets greedily...
[2021-03-20 01:13:11] Showing labelled: True (300/5000 visible, 0 redundant)
[2021-03-20 01:13:11] Reset parameters according to He et al.
[2021-03-20 01:13:11] mlp: initialized 18186 parameters.
[2021-03-20 01:13:11] Creating trainer with model on device: cuda
[2021-03-20 01:13:11] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:13:14] Training accuracy: 0.8559
[2021-03-20 01:13:14] Testing on 5000 data points...
[2021-03-20 01:13:14] Test score for 300 training labels: 0.9892
[2021-03-20 01:13:14] Experiment repeat 3/3
[2021-03-20 01:13:14] Seeded: 7
[2021-03-20 01:13:14] Using 1.00% labels of the dataset (50/5000)
[2021-03-20 01:13:14] Randomly labelled 50/5000
[2021-03-20 01:13:14] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:13:14] Seeded: 8
[2021-03-20 01:13:14] Running: experiment 0
[2021-03-20 01:13:14] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:13:14] Reset parameters according to He et al.
[2021-03-20 01:13:14] mlp: initialized 18186 parameters.
[2021-03-20 01:13:14] Creating trainer with model on device: cuda
[2021-03-20 01:13:14] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:13:15] Training accuracy: 0.6309
[2021-03-20 01:13:15] Testing on 5000 data points...
[2021-03-20 01:13:15] Test score for 50 training labels: 0.7536
[2021-03-20 01:13:15] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:15] Found 5000 unlabelled features.
[2021-03-20 01:13:15] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:13:15] Reset parameters according to He et al.
[2021-03-20 01:13:15] mlp: initialized 18186 parameters.
[2021-03-20 01:13:15] Creating trainer with model on device: cuda
[2021-03-20 01:13:15] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:13:16] Training accuracy: 0.6914
[2021-03-20 01:13:16] Testing on 5000 data points...
[2021-03-20 01:13:16] Test score for 100 training labels: 0.9086
[2021-03-20 01:13:16] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:16] Found 5000 unlabelled features.
[2021-03-20 01:13:16] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:13:16] Reset parameters according to He et al.
[2021-03-20 01:13:16] mlp: initialized 18186 parameters.
[2021-03-20 01:13:16] Creating trainer with model on device: cuda
[2021-03-20 01:13:16] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:13:18] Training accuracy: 0.7869
[2021-03-20 01:13:18] Testing on 5000 data points...
[2021-03-20 01:13:18] Test score for 150 training labels: 0.8988
[2021-03-20 01:13:18] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:18] Found 5000 unlabelled features.
[2021-03-20 01:13:18] Showing labelled: True (199/5000 visible, 1 redundant)
[2021-03-20 01:13:18] Reset parameters according to He et al.
[2021-03-20 01:13:18] mlp: initialized 18186 parameters.
[2021-03-20 01:13:18] Creating trainer with model on device: cuda
[2021-03-20 01:13:18] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:13:21] Training accuracy: 0.8304
[2021-03-20 01:13:21] Testing on 5000 data points...
[2021-03-20 01:13:21] Test score for 200 training labels: 0.9496
[2021-03-20 01:13:21] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:21] Found 5000 unlabelled features.
[2021-03-20 01:13:21] Showing labelled: True (246/5000 visible, 4 redundant)
[2021-03-20 01:13:21] Reset parameters according to He et al.
[2021-03-20 01:13:21] mlp: initialized 18186 parameters.
[2021-03-20 01:13:21] Creating trainer with model on device: cuda
[2021-03-20 01:13:21] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:13:23] Training accuracy: 0.8492
[2021-03-20 01:13:23] Testing on 5000 data points...
[2021-03-20 01:13:23] Test score for 250 training labels: 0.9256
[2021-03-20 01:13:24] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:24] Found 5000 unlabelled features.
[2021-03-20 01:13:24] Showing labelled: True (292/5000 visible, 8 redundant)
[2021-03-20 01:13:24] Reset parameters according to He et al.
[2021-03-20 01:13:24] mlp: initialized 18186 parameters.
[2021-03-20 01:13:24] Creating trainer with model on device: cuda
[2021-03-20 01:13:24] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:13:27] Training accuracy: 0.8811
[2021-03-20 01:13:27] Testing on 5000 data points...
[2021-03-20 01:13:27] Test score for 300 training labels: 0.9102
[2021-03-20 01:13:27] Seeded: 8
[2021-03-20 01:13:27] Running: experiment 1
[2021-03-20 01:13:27] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:13:27] Reset parameters according to He et al.
[2021-03-20 01:13:27] mlp: initialized 18186 parameters.
[2021-03-20 01:13:27] Creating trainer with model on device: cuda
[2021-03-20 01:13:27] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:13:28] Training accuracy: 0.6309
[2021-03-20 01:13:28] Testing on 5000 data points...
[2021-03-20 01:13:28] Test score for 50 training labels: 0.7536
[2021-03-20 01:13:28] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:28] Found 5000 unlabelled features.
[2021-03-20 01:13:28] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:13:28] Reset parameters according to He et al.
[2021-03-20 01:13:28] mlp: initialized 18186 parameters.
[2021-03-20 01:13:28] Creating trainer with model on device: cuda
[2021-03-20 01:13:28] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:13:29] Training accuracy: 0.7918
[2021-03-20 01:13:29] Testing on 5000 data points...
[2021-03-20 01:13:29] Test score for 100 training labels: 0.7248
[2021-03-20 01:13:29] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:29] Found 5000 unlabelled features.
[2021-03-20 01:13:29] Showing labelled: True (148/5000 visible, 2 redundant)
[2021-03-20 01:13:29] Reset parameters according to He et al.
[2021-03-20 01:13:29] mlp: initialized 18186 parameters.
[2021-03-20 01:13:29] Creating trainer with model on device: cuda
[2021-03-20 01:13:29] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:13:31] Training accuracy: 0.8550
[2021-03-20 01:13:31] Testing on 5000 data points...
[2021-03-20 01:13:31] Test score for 150 training labels: 0.7728
[2021-03-20 01:13:31] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:31] Found 5000 unlabelled features.
[2021-03-20 01:13:31] Showing labelled: True (198/5000 visible, 2 redundant)
[2021-03-20 01:13:31] Reset parameters according to He et al.
[2021-03-20 01:13:31] mlp: initialized 18186 parameters.
[2021-03-20 01:13:31] Creating trainer with model on device: cuda
[2021-03-20 01:13:31] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:13:34] Training accuracy: 0.9010
[2021-03-20 01:13:34] Testing on 5000 data points...
[2021-03-20 01:13:34] Test score for 200 training labels: 0.8000
[2021-03-20 01:13:34] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:34] Found 5000 unlabelled features.
[2021-03-20 01:13:34] Showing labelled: True (237/5000 visible, 13 redundant)
[2021-03-20 01:13:34] Reset parameters according to He et al.
[2021-03-20 01:13:34] mlp: initialized 18186 parameters.
[2021-03-20 01:13:34] Creating trainer with model on device: cuda
[2021-03-20 01:13:34] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:13:37] Training accuracy: 0.8076
[2021-03-20 01:13:37] Testing on 5000 data points...
[2021-03-20 01:13:37] Test score for 250 training labels: 0.9132
[2021-03-20 01:13:37] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:37] Found 5000 unlabelled features.
[2021-03-20 01:13:37] Showing labelled: True (281/5000 visible, 19 redundant)
[2021-03-20 01:13:37] Reset parameters according to He et al.
[2021-03-20 01:13:37] mlp: initialized 18186 parameters.
[2021-03-20 01:13:37] Creating trainer with model on device: cuda
[2021-03-20 01:13:37] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:13:40] Training accuracy: 0.7786
[2021-03-20 01:13:40] Testing on 5000 data points...
[2021-03-20 01:13:40] Test score for 300 training labels: 0.7730
[2021-03-20 01:13:40] Seeded: 8
[2021-03-20 01:13:40] Running: experiment 2
[2021-03-20 01:13:40] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:13:40] Reset parameters according to He et al.
[2021-03-20 01:13:40] mlp: initialized 18186 parameters.
[2021-03-20 01:13:40] Creating trainer with model on device: cuda
[2021-03-20 01:13:40] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:13:41] Training accuracy: 0.6309
[2021-03-20 01:13:41] Testing on 5000 data points...
[2021-03-20 01:13:41] Test score for 50 training labels: 0.7536
[2021-03-20 01:13:41] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:41] Found 5000 unlabelled features.
[2021-03-20 01:13:41] Computing distance between 50 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:41] Searching for coresets with 20 beams...
[2021-03-20 01:13:41] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:13:41] Reset parameters according to He et al.
[2021-03-20 01:13:41] mlp: initialized 18186 parameters.
[2021-03-20 01:13:41] Creating trainer with model on device: cuda
[2021-03-20 01:13:41] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:13:42] Training accuracy: 0.6920
[2021-03-20 01:13:42] Testing on 5000 data points...
[2021-03-20 01:13:42] Test score for 100 training labels: 0.8108
[2021-03-20 01:13:42] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:42] Found 5000 unlabelled features.
[2021-03-20 01:13:42] Computing distance between 100 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:42] Searching for coresets with 20 beams...
[2021-03-20 01:13:43] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:13:43] Reset parameters according to He et al.
[2021-03-20 01:13:43] mlp: initialized 18186 parameters.
[2021-03-20 01:13:43] Creating trainer with model on device: cuda
[2021-03-20 01:13:43] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:13:45] Training accuracy: 0.7586
[2021-03-20 01:13:45] Testing on 5000 data points...
[2021-03-20 01:13:45] Test score for 150 training labels: 0.9110
[2021-03-20 01:13:45] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:45] Found 5000 unlabelled features.
[2021-03-20 01:13:45] Computing distance between 150 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:45] Searching for coresets with 20 beams...
[2021-03-20 01:13:45] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:13:45] Reset parameters according to He et al.
[2021-03-20 01:13:45] mlp: initialized 18186 parameters.
[2021-03-20 01:13:45] Creating trainer with model on device: cuda
[2021-03-20 01:13:45] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:13:47] Training accuracy: 0.8212
[2021-03-20 01:13:47] Testing on 5000 data points...
[2021-03-20 01:13:47] Test score for 200 training labels: 0.8984
[2021-03-20 01:13:47] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:47] Found 5000 unlabelled features.
[2021-03-20 01:13:47] Computing distance between 200 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:47] Searching for coresets with 20 beams...
[2021-03-20 01:13:47] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:13:47] Reset parameters according to He et al.
[2021-03-20 01:13:47] mlp: initialized 18186 parameters.
[2021-03-20 01:13:47] Creating trainer with model on device: cuda
[2021-03-20 01:13:47] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:13:50] Training accuracy: 0.8399
[2021-03-20 01:13:50] Testing on 5000 data points...
[2021-03-20 01:13:50] Test score for 250 training labels: 0.9356
[2021-03-20 01:13:50] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:50] Found 5000 unlabelled features.
[2021-03-20 01:13:50] Computing distance between 250 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:50] Searching for coresets with 20 beams...
[2021-03-20 01:13:51] Showing labelled: True (300/5000 visible, 0 redundant)
[2021-03-20 01:13:51] Reset parameters according to He et al.
[2021-03-20 01:13:51] mlp: initialized 18186 parameters.
[2021-03-20 01:13:51] Creating trainer with model on device: cuda
[2021-03-20 01:13:51] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:13:54] Training accuracy: 0.8576
[2021-03-20 01:13:54] Testing on 5000 data points...
[2021-03-20 01:13:54] Test score for 300 training labels: 0.9606
[2021-03-20 01:13:54] Seeded: 8
[2021-03-20 01:13:54] Running: experiment 3
[2021-03-20 01:13:54] Showing labelled: True (50/5000 visible, 0 redundant)
[2021-03-20 01:13:54] Reset parameters according to He et al.
[2021-03-20 01:13:54] mlp: initialized 18186 parameters.
[2021-03-20 01:13:54] Creating trainer with model on device: cuda
[2021-03-20 01:13:54] Training mlp across 50 data points in blobsthin-balanced...
[2021-03-20 01:13:55] Training accuracy: 0.6309
[2021-03-20 01:13:55] Testing on 5000 data points...
[2021-03-20 01:13:55] Test score for 50 training labels: 0.7536
[2021-03-20 01:13:55] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:55] Found 5000 unlabelled features.
[2021-03-20 01:13:55] Computing distance between 50 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:55] Searching for coresets greedily...
[2021-03-20 01:13:55] Showing labelled: True (100/5000 visible, 0 redundant)
[2021-03-20 01:13:55] Reset parameters according to He et al.
[2021-03-20 01:13:55] mlp: initialized 18186 parameters.
[2021-03-20 01:13:55] Creating trainer with model on device: cuda
[2021-03-20 01:13:55] Training mlp across 100 data points in blobsthin-balanced...
[2021-03-20 01:13:56] Training accuracy: 0.6924
[2021-03-20 01:13:56] Testing on 5000 data points...
[2021-03-20 01:13:56] Test score for 100 training labels: 0.7984
[2021-03-20 01:13:56] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:56] Found 5000 unlabelled features.
[2021-03-20 01:13:56] Computing distance between 100 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:56] Searching for coresets greedily...
[2021-03-20 01:13:56] Showing labelled: True (150/5000 visible, 0 redundant)
[2021-03-20 01:13:56] Reset parameters according to He et al.
[2021-03-20 01:13:56] mlp: initialized 18186 parameters.
[2021-03-20 01:13:56] Creating trainer with model on device: cuda
[2021-03-20 01:13:56] Training mlp across 150 data points in blobsthin-balanced...
[2021-03-20 01:13:58] Training accuracy: 0.7413
[2021-03-20 01:13:58] Testing on 5000 data points...
[2021-03-20 01:13:58] Test score for 150 training labels: 0.8974
[2021-03-20 01:13:58] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:13:58] Found 5000 unlabelled features.
[2021-03-20 01:13:58] Computing distance between 150 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:13:58] Searching for coresets greedily...
[2021-03-20 01:13:58] Showing labelled: True (200/5000 visible, 0 redundant)
[2021-03-20 01:13:58] Reset parameters according to He et al.
[2021-03-20 01:13:58] mlp: initialized 18186 parameters.
[2021-03-20 01:13:58] Creating trainer with model on device: cuda
[2021-03-20 01:13:58] Training mlp across 200 data points in blobsthin-balanced...
[2021-03-20 01:14:00] Training accuracy: 0.8179
[2021-03-20 01:14:00] Testing on 5000 data points...
[2021-03-20 01:14:00] Test score for 200 training labels: 0.8828
[2021-03-20 01:14:00] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:14:00] Found 5000 unlabelled features.
[2021-03-20 01:14:01] Computing distance between 200 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:14:01] Searching for coresets greedily...
[2021-03-20 01:14:01] Showing labelled: True (250/5000 visible, 0 redundant)
[2021-03-20 01:14:01] Reset parameters according to He et al.
[2021-03-20 01:14:01] mlp: initialized 18186 parameters.
[2021-03-20 01:14:01] Creating trainer with model on device: cuda
[2021-03-20 01:14:01] Training mlp across 250 data points in blobsthin-balanced...
[2021-03-20 01:14:03] Training accuracy: 0.8544
[2021-03-20 01:14:03] Testing on 5000 data points...
[2021-03-20 01:14:03] Test score for 250 training labels: 0.9006
[2021-03-20 01:14:03] Showing labelled: False (5000/5000 visible, 0 redundant)
[2021-03-20 01:14:03] Found 5000 unlabelled features.
[2021-03-20 01:14:03] Computing distance between 250 labelled and 5000 unlabelled vectors of length 128...
[2021-03-20 01:14:04] Searching for coresets greedily...
[2021-03-20 01:14:04] Showing labelled: True (300/5000 visible, 0 redundant)
[2021-03-20 01:14:04] Reset parameters according to He et al.
[2021-03-20 01:14:04] mlp: initialized 18186 parameters.
[2021-03-20 01:14:04] Creating trainer with model on device: cuda
[2021-03-20 01:14:04] Training mlp across 300 data points in blobsthin-balanced...
[2021-03-20 01:14:07] Training accuracy: 0.8670
[2021-03-20 01:14:07] Testing on 5000 data points...
[2021-03-20 01:14:07] Test score for 300 training labels: 0.9654
[2021-03-20 01:14:07] Updated results: ../results/blobsthin-balanced/mlp/results.json
