[2021-03-16 00:37:38] Seeded: 5
[2021-03-16 00:37:38] Created experiment 0:
[2021-03-16 00:37:38]  - Model: mlp
[2021-03-16 00:37:38]  - Acquisition function: sp-coreset
[2021-03-16 00:37:38] Created experiment 1:
[2021-03-16 00:37:38]  - Model: mlp
[2021-03-16 00:37:38]  - Acquisition function: greedy-coreset
[2021-03-16 00:37:38] Created experiment 2:
[2021-03-16 00:37:38]  - Model: mlp
[2021-03-16 00:37:38]  - Acquisition function: random
[2021-03-16 00:37:38] Loading blobsthin test set...
[2021-03-16 00:37:38] Generated 5000 blobs:
[2021-03-16 00:37:38] Class 0: 200
[2021-03-16 00:37:38] Class 1: 300
[2021-03-16 00:37:38] Class 2: 1000
[2021-03-16 00:37:38] Class 3: 100
[2021-03-16 00:37:38] Class 4: 50
[2021-03-16 00:37:38] Class 5: 200
[2021-03-16 00:37:38] Class 6: 400
[2021-03-16 00:37:38] Class 7: 1800
[2021-03-16 00:37:38] Class 8: 50
[2021-03-16 00:37:38] Class 9: 900
[2021-03-16 00:37:39] Saved figure to: ../data/blobsthin_data.png
[2021-03-16 00:37:39] Generated 5000 test points.
[2021-03-16 00:37:39] Saved figure to: ../data/blobsthin_test.png
[2021-03-16 00:37:39] Experiment repeat 1/2
[2021-03-16 00:37:39] Randomly labelled 50/5000
[2021-03-16 00:37:39] Showing labelled: True (50/5000 visible)
[2021-03-16 00:37:39] Running: experiment 0
[2021-03-16 00:37:39] Showing labelled: True (50/5000 visible)
[2021-03-16 00:37:39] mlp: initialized 1482 parameters.
[2021-03-16 00:37:39] Creating trainer with model on device: cuda
[2021-03-16 00:37:44] Training mlp across 50 data points in blobsthin...
[2021-03-16 00:37:44] Training loss: 1.5627
[2021-03-16 00:37:44] Testing on 5000 data points...
[2021-03-16 00:37:44] Test accuracy: 0.5000
[2021-03-16 00:37:44] Score for 50 labels: 0.5000
[2021-03-16 00:37:44] Showing labelled: False (4950/5000 visible)
[2021-03-16 00:37:44] Found 4950 unlabelled features.
[2021-03-16 00:37:44] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:44] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:45] Showing labelled: True (75/5000 visible)
[2021-03-16 00:37:45] Training mlp across 75 data points in blobsthin...
[2021-03-16 00:37:45] Training loss: 0.6830
[2021-03-16 00:37:45] Testing on 5000 data points...
[2021-03-16 00:37:45] Test accuracy: 0.7000
[2021-03-16 00:37:45] Score for 75 labels: 0.7000
[2021-03-16 00:37:45] Showing labelled: False (4925/5000 visible)
[2021-03-16 00:37:45] Found 4925 unlabelled features.
[2021-03-16 00:37:45] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:46] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:46] Showing labelled: True (100/5000 visible)
[2021-03-16 00:37:46] Training mlp across 100 data points in blobsthin...
[2021-03-16 00:37:47] Training loss: 0.3053
[2021-03-16 00:37:47] Testing on 5000 data points...
[2021-03-16 00:37:47] Test accuracy: 0.7980
[2021-03-16 00:37:47] Score for 100 labels: 0.7980
[2021-03-16 00:37:47] Showing labelled: False (4900/5000 visible)
[2021-03-16 00:37:47] Found 4900 unlabelled features.
[2021-03-16 00:37:47] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:47] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:48] Showing labelled: True (125/5000 visible)
[2021-03-16 00:37:48] Training mlp across 125 data points in blobsthin...
[2021-03-16 00:37:48] Training loss: 0.1768
[2021-03-16 00:37:48] Testing on 5000 data points...
[2021-03-16 00:37:49] Test accuracy: 0.8038
[2021-03-16 00:37:49] Score for 125 labels: 0.8038
[2021-03-16 00:37:49] Showing labelled: False (4875/5000 visible)
[2021-03-16 00:37:49] Found 4875 unlabelled features.
[2021-03-16 00:37:49] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:49] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:49] Showing labelled: True (150/5000 visible)
[2021-03-16 00:37:49] Training mlp across 150 data points in blobsthin...
[2021-03-16 00:37:50] Training loss: 0.1154
[2021-03-16 00:37:50] Testing on 5000 data points...
[2021-03-16 00:37:50] Test accuracy: 0.8388
[2021-03-16 00:37:50] Score for 150 labels: 0.8388
[2021-03-16 00:37:50] Showing labelled: False (4850/5000 visible)
[2021-03-16 00:37:50] Found 4850 unlabelled features.
[2021-03-16 00:37:50] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:51] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:51] Showing labelled: True (175/5000 visible)
[2021-03-16 00:37:51] Training mlp across 175 data points in blobsthin...
[2021-03-16 00:37:52] Training loss: 0.0902
[2021-03-16 00:37:52] Testing on 5000 data points...
[2021-03-16 00:37:52] Test accuracy: 0.8786
[2021-03-16 00:37:52] Score for 175 labels: 0.8786
[2021-03-16 00:37:52] Showing labelled: False (4825/5000 visible)
[2021-03-16 00:37:52] Found 4825 unlabelled features.
[2021-03-16 00:37:52] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:53] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:53] Showing labelled: True (200/5000 visible)
[2021-03-16 00:37:53] Training mlp across 200 data points in blobsthin...
[2021-03-16 00:37:54] Training loss: 0.0594
[2021-03-16 00:37:54] Testing on 5000 data points...
[2021-03-16 00:37:54] Test accuracy: 0.8912
[2021-03-16 00:37:54] Score for 200 labels: 0.8912
[2021-03-16 00:37:54] Showing labelled: False (4800/5000 visible)
[2021-03-16 00:37:54] Found 4800 unlabelled features.
[2021-03-16 00:37:54] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:54] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:55] Showing labelled: True (225/5000 visible)
[2021-03-16 00:37:55] Training mlp across 225 data points in blobsthin...
[2021-03-16 00:37:56] Training loss: 0.0357
[2021-03-16 00:37:56] Testing on 5000 data points...
[2021-03-16 00:37:56] Test accuracy: 0.9940
[2021-03-16 00:37:56] Score for 225 labels: 0.9940
[2021-03-16 00:37:56] Showing labelled: False (4775/5000 visible)
[2021-03-16 00:37:56] Found 4775 unlabelled features.
[2021-03-16 00:37:56] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:57] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:57] Showing labelled: True (250/5000 visible)
[2021-03-16 00:37:57] Training mlp across 250 data points in blobsthin...
[2021-03-16 00:37:59] Training loss: 0.0228
[2021-03-16 00:37:59] Testing on 5000 data points...
[2021-03-16 00:37:59] Test accuracy: 0.9934
[2021-03-16 00:37:59] Score for 250 labels: 0.9934
[2021-03-16 00:37:59] Showing labelled: False (4750/5000 visible)
[2021-03-16 00:37:59] Found 4750 unlabelled features.
[2021-03-16 00:37:59] Training GMM (k=50) on labelled features...
[2021-03-16 00:37:59] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:37:59] Showing labelled: True (275/5000 visible)
[2021-03-16 00:37:59] Training mlp across 275 data points in blobsthin...
[2021-03-16 00:38:01] Training loss: 0.0162
[2021-03-16 00:38:01] Testing on 5000 data points...
[2021-03-16 00:38:01] Test accuracy: 0.9948
[2021-03-16 00:38:01] Score for 275 labels: 0.9948
[2021-03-16 00:38:01] Showing labelled: False (4725/5000 visible)
[2021-03-16 00:38:01] Found 4725 unlabelled features.
[2021-03-16 00:38:01] Training GMM (k=50) on labelled features...
[2021-03-16 00:38:01] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:38:02] Showing labelled: True (300/5000 visible)
[2021-03-16 00:38:02] Training mlp across 300 data points in blobsthin...
[2021-03-16 00:38:03] Training loss: 0.0114
[2021-03-16 00:38:03] Testing on 5000 data points...
[2021-03-16 00:38:03] Test accuracy: 0.9954
[2021-03-16 00:38:03] Score for 300 labels: 0.9954
[2021-03-16 00:38:03] Running: experiment 1
[2021-03-16 00:38:03] Showing labelled: True (50/5000 visible)
[2021-03-16 00:38:03] mlp: initialized 1482 parameters.
[2021-03-16 00:38:03] Creating trainer with model on device: cuda
[2021-03-16 00:38:03] Training mlp across 50 data points in blobsthin...
[2021-03-16 00:38:03] Training loss: 1.5609
[2021-03-16 00:38:03] Testing on 5000 data points...
[2021-03-16 00:38:03] Test accuracy: 0.5000
[2021-03-16 00:38:03] Score for 50 labels: 0.5000
[2021-03-16 00:38:03] Showing labelled: False (4950/5000 visible)
[2021-03-16 00:38:03] Found 4950 unlabelled features.
[2021-03-16 00:38:03] Searching core-set greedily...
[2021-03-16 00:38:06] Showing labelled: True (75/5000 visible)
[2021-03-16 00:38:06] Training mlp across 75 data points in blobsthin...
[2021-03-16 00:38:06] Training loss: 0.6880
[2021-03-16 00:38:06] Testing on 5000 data points...
[2021-03-16 00:38:06] Test accuracy: 0.8700
[2021-03-16 00:38:06] Score for 75 labels: 0.8700
[2021-03-16 00:38:06] Showing labelled: False (4925/5000 visible)
[2021-03-16 00:38:06] Found 4925 unlabelled features.
[2021-03-16 00:38:06] Searching core-set greedily...
[2021-03-16 00:38:09] Showing labelled: True (100/5000 visible)
[2021-03-16 00:38:09] Training mlp across 100 data points in blobsthin...
[2021-03-16 00:38:10] Training loss: 0.3457
[2021-03-16 00:38:10] Testing on 5000 data points...
[2021-03-16 00:38:10] Test accuracy: 0.9006
[2021-03-16 00:38:10] Score for 100 labels: 0.9006
[2021-03-16 00:38:10] Showing labelled: False (4900/5000 visible)
[2021-03-16 00:38:10] Found 4900 unlabelled features.
[2021-03-16 00:38:10] Searching core-set greedily...
[2021-03-16 00:38:14] Showing labelled: True (125/5000 visible)
[2021-03-16 00:38:14] Training mlp across 125 data points in blobsthin...
[2021-03-16 00:38:14] Training loss: 0.1958
[2021-03-16 00:38:14] Testing on 5000 data points...
[2021-03-16 00:38:14] Test accuracy: 0.9790
[2021-03-16 00:38:14] Score for 125 labels: 0.9790
[2021-03-16 00:38:14] Showing labelled: False (4875/5000 visible)
[2021-03-16 00:38:14] Found 4875 unlabelled features.
[2021-03-16 00:38:14] Searching core-set greedily...
[2021-03-16 00:38:18] Showing labelled: True (150/5000 visible)
[2021-03-16 00:38:18] Training mlp across 150 data points in blobsthin...
[2021-03-16 00:38:19] Training loss: 0.1169
[2021-03-16 00:38:19] Testing on 5000 data points...
[2021-03-16 00:38:19] Test accuracy: 0.9952
[2021-03-16 00:38:19] Score for 150 labels: 0.9952
[2021-03-16 00:38:19] Showing labelled: False (4850/5000 visible)
[2021-03-16 00:38:19] Found 4850 unlabelled features.
[2021-03-16 00:38:19] Searching core-set greedily...
[2021-03-16 00:38:24] Showing labelled: True (175/5000 visible)
[2021-03-16 00:38:24] Training mlp across 175 data points in blobsthin...
[2021-03-16 00:38:25] Training loss: 0.0810
[2021-03-16 00:38:25] Testing on 5000 data points...
[2021-03-16 00:38:25] Test accuracy: 0.9968
[2021-03-16 00:38:25] Score for 175 labels: 0.9968
[2021-03-16 00:38:25] Showing labelled: False (4825/5000 visible)
[2021-03-16 00:38:25] Found 4825 unlabelled features.
[2021-03-16 00:38:25] Searching core-set greedily...
[2021-03-16 00:38:30] Showing labelled: True (200/5000 visible)
[2021-03-16 00:38:30] Training mlp across 200 data points in blobsthin...
[2021-03-16 00:38:31] Training loss: 0.0590
[2021-03-16 00:38:31] Testing on 5000 data points...
[2021-03-16 00:38:31] Test accuracy: 0.9978
[2021-03-16 00:38:31] Score for 200 labels: 0.9978
[2021-03-16 00:38:31] Showing labelled: False (4800/5000 visible)
[2021-03-16 00:38:31] Found 4800 unlabelled features.
[2021-03-16 00:38:31] Searching core-set greedily...
[2021-03-16 00:38:36] Showing labelled: True (225/5000 visible)
[2021-03-16 00:38:36] Training mlp across 225 data points in blobsthin...
[2021-03-16 00:38:37] Training loss: 0.0452
[2021-03-16 00:38:37] Testing on 5000 data points...
[2021-03-16 00:38:37] Test accuracy: 0.9972
[2021-03-16 00:38:37] Score for 225 labels: 0.9972
[2021-03-16 00:38:37] Showing labelled: False (4775/5000 visible)
[2021-03-16 00:38:37] Found 4775 unlabelled features.
[2021-03-16 00:38:37] Searching core-set greedily...
[2021-03-16 00:38:43] Showing labelled: True (250/5000 visible)
[2021-03-16 00:38:43] Training mlp across 250 data points in blobsthin...
[2021-03-16 00:38:44] Training loss: 0.0328
[2021-03-16 00:38:44] Testing on 5000 data points...
[2021-03-16 00:38:44] Test accuracy: 0.9972
[2021-03-16 00:38:44] Score for 250 labels: 0.9972
[2021-03-16 00:38:44] Showing labelled: False (4750/5000 visible)
[2021-03-16 00:38:44] Found 4750 unlabelled features.
[2021-03-16 00:38:44] Searching core-set greedily...
[2021-03-16 00:38:48] Showing labelled: True (275/5000 visible)
[2021-03-16 00:38:48] Training mlp across 275 data points in blobsthin...
[2021-03-16 00:38:50] Training loss: 0.0285
[2021-03-16 00:38:50] Testing on 5000 data points...
[2021-03-16 00:38:50] Test accuracy: 0.9980
[2021-03-16 00:38:50] Score for 275 labels: 0.9980
[2021-03-16 00:38:50] Showing labelled: False (4725/5000 visible)
[2021-03-16 00:38:50] Found 4725 unlabelled features.
[2021-03-16 00:38:50] Searching core-set greedily...
[2021-03-16 00:38:55] Showing labelled: True (300/5000 visible)
[2021-03-16 00:38:55] Training mlp across 300 data points in blobsthin...
[2021-03-16 00:38:56] Training loss: 0.0230
[2021-03-16 00:38:56] Testing on 5000 data points...
[2021-03-16 00:38:56] Test accuracy: 0.9980
[2021-03-16 00:38:56] Score for 300 labels: 0.9980
[2021-03-16 00:38:56] Running: experiment 2
[2021-03-16 00:38:56] Showing labelled: True (50/5000 visible)
[2021-03-16 00:38:56] mlp: initialized 1482 parameters.
[2021-03-16 00:38:56] Creating trainer with model on device: cuda
[2021-03-16 00:38:56] Training mlp across 50 data points in blobsthin...
[2021-03-16 00:38:57] Training loss: 1.5098
[2021-03-16 00:38:57] Testing on 5000 data points...
[2021-03-16 00:38:57] Test accuracy: 0.5000
[2021-03-16 00:38:57] Score for 50 labels: 0.5000
[2021-03-16 00:38:57] Showing labelled: False (4950/5000 visible)
[2021-03-16 00:38:57] Found 4950 unlabelled features.
[2021-03-16 00:38:57] Showing labelled: True (75/5000 visible)
[2021-03-16 00:38:57] Training mlp across 75 data points in blobsthin...
[2021-03-16 00:38:57] Training loss: 0.6984
[2021-03-16 00:38:57] Testing on 5000 data points...
[2021-03-16 00:38:57] Test accuracy: 0.6358
[2021-03-16 00:38:57] Score for 75 labels: 0.6358
[2021-03-16 00:38:57] Showing labelled: False (4925/5000 visible)
[2021-03-16 00:38:57] Found 4925 unlabelled features.
[2021-03-16 00:38:57] Showing labelled: True (100/5000 visible)
[2021-03-16 00:38:57] Training mlp across 100 data points in blobsthin...
[2021-03-16 00:38:58] Training loss: 0.3333
[2021-03-16 00:38:58] Testing on 5000 data points...
[2021-03-16 00:38:58] Test accuracy: 0.7904
[2021-03-16 00:38:58] Score for 100 labels: 0.7904
[2021-03-16 00:38:58] Showing labelled: False (4900/5000 visible)
[2021-03-16 00:38:58] Found 4900 unlabelled features.
[2021-03-16 00:38:58] Showing labelled: True (125/5000 visible)
[2021-03-16 00:38:58] Training mlp across 125 data points in blobsthin...
[2021-03-16 00:38:58] Training loss: 0.1797
[2021-03-16 00:38:58] Testing on 5000 data points...
[2021-03-16 00:38:58] Test accuracy: 0.8452
[2021-03-16 00:38:58] Score for 125 labels: 0.8452
[2021-03-16 00:38:58] Showing labelled: False (4875/5000 visible)
[2021-03-16 00:38:58] Found 4875 unlabelled features.
[2021-03-16 00:38:59] Showing labelled: True (150/5000 visible)
[2021-03-16 00:38:59] Training mlp across 150 data points in blobsthin...
[2021-03-16 00:38:59] Training loss: 0.1098
[2021-03-16 00:38:59] Testing on 5000 data points...
[2021-03-16 00:38:59] Test accuracy: 0.9482
[2021-03-16 00:38:59] Score for 150 labels: 0.9482
[2021-03-16 00:38:59] Showing labelled: False (4850/5000 visible)
[2021-03-16 00:38:59] Found 4850 unlabelled features.
[2021-03-16 00:38:59] Showing labelled: True (175/5000 visible)
[2021-03-16 00:38:59] Training mlp across 175 data points in blobsthin...
[2021-03-16 00:39:00] Training loss: 0.0656
[2021-03-16 00:39:00] Testing on 5000 data points...
[2021-03-16 00:39:00] Test accuracy: 0.9626
[2021-03-16 00:39:00] Score for 175 labels: 0.9626
[2021-03-16 00:39:00] Showing labelled: False (4825/5000 visible)
[2021-03-16 00:39:00] Found 4825 unlabelled features.
[2021-03-16 00:39:00] Showing labelled: True (200/5000 visible)
[2021-03-16 00:39:00] Training mlp across 200 data points in blobsthin...
[2021-03-16 00:39:01] Training loss: 0.0421
[2021-03-16 00:39:01] Testing on 5000 data points...
[2021-03-16 00:39:01] Test accuracy: 0.9752
[2021-03-16 00:39:01] Score for 200 labels: 0.9752
[2021-03-16 00:39:01] Showing labelled: False (4800/5000 visible)
[2021-03-16 00:39:01] Found 4800 unlabelled features.
[2021-03-16 00:39:01] Showing labelled: True (225/5000 visible)
[2021-03-16 00:39:01] Training mlp across 225 data points in blobsthin...
[2021-03-16 00:39:02] Training loss: 0.0287
[2021-03-16 00:39:02] Testing on 5000 data points...
[2021-03-16 00:39:02] Test accuracy: 0.9786
[2021-03-16 00:39:02] Score for 225 labels: 0.9786
[2021-03-16 00:39:02] Showing labelled: False (4775/5000 visible)
[2021-03-16 00:39:02] Found 4775 unlabelled features.
[2021-03-16 00:39:02] Showing labelled: True (250/5000 visible)
[2021-03-16 00:39:02] Training mlp across 250 data points in blobsthin...
[2021-03-16 00:39:04] Training loss: 0.0212
[2021-03-16 00:39:04] Testing on 5000 data points...
[2021-03-16 00:39:04] Test accuracy: 0.9842
[2021-03-16 00:39:04] Score for 250 labels: 0.9842
[2021-03-16 00:39:04] Showing labelled: False (4750/5000 visible)
[2021-03-16 00:39:04] Found 4750 unlabelled features.
[2021-03-16 00:39:04] Showing labelled: True (275/5000 visible)
[2021-03-16 00:39:04] Training mlp across 275 data points in blobsthin...
[2021-03-16 00:39:05] Training loss: 0.0167
[2021-03-16 00:39:05] Testing on 5000 data points...
[2021-03-16 00:39:05] Test accuracy: 0.9848
[2021-03-16 00:39:05] Score for 275 labels: 0.9848
[2021-03-16 00:39:05] Showing labelled: False (4725/5000 visible)
[2021-03-16 00:39:05] Found 4725 unlabelled features.
[2021-03-16 00:39:05] Showing labelled: True (300/5000 visible)
[2021-03-16 00:39:05] Training mlp across 300 data points in blobsthin...
[2021-03-16 00:39:06] Training loss: 0.0121
[2021-03-16 00:39:06] Testing on 5000 data points...
[2021-03-16 00:39:06] Test accuracy: 0.9858
[2021-03-16 00:39:06] Score for 300 labels: 0.9858
[2021-03-16 00:39:06] Experiment repeat 2/2
[2021-03-16 00:39:06] Randomly labelled 50/5000
[2021-03-16 00:39:06] Showing labelled: True (50/5000 visible)
[2021-03-16 00:39:06] Running: experiment 0
[2021-03-16 00:39:06] Showing labelled: True (50/5000 visible)
[2021-03-16 00:39:06] mlp: initialized 1482 parameters.
[2021-03-16 00:39:06] Creating trainer with model on device: cuda
[2021-03-16 00:39:06] Training mlp across 50 data points in blobsthin...
[2021-03-16 00:39:07] Training loss: 1.2924
[2021-03-16 00:39:07] Testing on 5000 data points...
[2021-03-16 00:39:07] Test accuracy: 0.3992
[2021-03-16 00:39:07] Score for 50 labels: 0.3992
[2021-03-16 00:39:07] Showing labelled: False (4950/5000 visible)
[2021-03-16 00:39:07] Found 4950 unlabelled features.
[2021-03-16 00:39:07] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:07] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:07] Showing labelled: True (75/5000 visible)
[2021-03-16 00:39:07] Training mlp across 75 data points in blobsthin...
[2021-03-16 00:39:08] Training loss: 0.6314
[2021-03-16 00:39:08] Testing on 5000 data points...
[2021-03-16 00:39:08] Test accuracy: 0.7104
[2021-03-16 00:39:08] Score for 75 labels: 0.7104
[2021-03-16 00:39:08] Showing labelled: False (4925/5000 visible)
[2021-03-16 00:39:08] Found 4925 unlabelled features.
[2021-03-16 00:39:08] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:08] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:08] Showing labelled: True (100/5000 visible)
[2021-03-16 00:39:08] Training mlp across 100 data points in blobsthin...
[2021-03-16 00:39:09] Training loss: 0.3122
[2021-03-16 00:39:09] Testing on 5000 data points...
[2021-03-16 00:39:09] Test accuracy: 0.7804
[2021-03-16 00:39:09] Score for 100 labels: 0.7804
[2021-03-16 00:39:09] Showing labelled: False (4900/5000 visible)
[2021-03-16 00:39:09] Found 4900 unlabelled features.
[2021-03-16 00:39:09] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:09] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:09] Showing labelled: True (125/5000 visible)
[2021-03-16 00:39:09] Training mlp across 125 data points in blobsthin...
[2021-03-16 00:39:10] Training loss: 0.2028
[2021-03-16 00:39:10] Testing on 5000 data points...
[2021-03-16 00:39:10] Test accuracy: 0.8866
[2021-03-16 00:39:10] Score for 125 labels: 0.8866
[2021-03-16 00:39:10] Showing labelled: False (4875/5000 visible)
[2021-03-16 00:39:10] Found 4875 unlabelled features.
[2021-03-16 00:39:10] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:10] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:11] Showing labelled: True (150/5000 visible)
[2021-03-16 00:39:11] Training mlp across 150 data points in blobsthin...
[2021-03-16 00:39:11] Training loss: 0.1373
[2021-03-16 00:39:11] Testing on 5000 data points...
[2021-03-16 00:39:11] Test accuracy: 0.8994
[2021-03-16 00:39:11] Score for 150 labels: 0.8994
[2021-03-16 00:39:11] Showing labelled: False (4850/5000 visible)
[2021-03-16 00:39:11] Found 4850 unlabelled features.
[2021-03-16 00:39:11] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:12] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:12] Showing labelled: True (175/5000 visible)
[2021-03-16 00:39:12] Training mlp across 175 data points in blobsthin...
[2021-03-16 00:39:13] Training loss: 0.0940
[2021-03-16 00:39:13] Testing on 5000 data points...
[2021-03-16 00:39:13] Test accuracy: 0.9238
[2021-03-16 00:39:13] Score for 175 labels: 0.9238
[2021-03-16 00:39:13] Showing labelled: False (4825/5000 visible)
[2021-03-16 00:39:13] Found 4825 unlabelled features.
[2021-03-16 00:39:13] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:13] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:14] Showing labelled: True (200/5000 visible)
[2021-03-16 00:39:14] Training mlp across 200 data points in blobsthin...
[2021-03-16 00:39:14] Training loss: 0.0692
[2021-03-16 00:39:14] Testing on 5000 data points...
[2021-03-16 00:39:14] Test accuracy: 0.9458
[2021-03-16 00:39:14] Score for 200 labels: 0.9458
[2021-03-16 00:39:14] Showing labelled: False (4800/5000 visible)
[2021-03-16 00:39:14] Found 4800 unlabelled features.
[2021-03-16 00:39:15] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:15] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:15] Showing labelled: True (225/5000 visible)
[2021-03-16 00:39:15] Training mlp across 225 data points in blobsthin...
[2021-03-16 00:39:16] Training loss: 0.0546
[2021-03-16 00:39:16] Testing on 5000 data points...
[2021-03-16 00:39:16] Test accuracy: 0.9660
[2021-03-16 00:39:16] Score for 225 labels: 0.9660
[2021-03-16 00:39:16] Showing labelled: False (4775/5000 visible)
[2021-03-16 00:39:16] Found 4775 unlabelled features.
[2021-03-16 00:39:16] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:17] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:17] Showing labelled: True (250/5000 visible)
[2021-03-16 00:39:17] Training mlp across 250 data points in blobsthin...
[2021-03-16 00:39:18] Training loss: 0.0393
[2021-03-16 00:39:18] Testing on 5000 data points...
[2021-03-16 00:39:18] Test accuracy: 0.9772
[2021-03-16 00:39:18] Score for 250 labels: 0.9772
[2021-03-16 00:39:18] Showing labelled: False (4750/5000 visible)
[2021-03-16 00:39:18] Found 4750 unlabelled features.
[2021-03-16 00:39:18] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:19] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:19] Showing labelled: True (275/5000 visible)
[2021-03-16 00:39:19] Training mlp across 275 data points in blobsthin...
[2021-03-16 00:39:21] Training loss: 0.0295
[2021-03-16 00:39:21] Testing on 5000 data points...
[2021-03-16 00:39:21] Test accuracy: 0.9844
[2021-03-16 00:39:21] Score for 275 labels: 0.9844
[2021-03-16 00:39:21] Showing labelled: False (4725/5000 visible)
[2021-03-16 00:39:21] Found 4725 unlabelled features.
[2021-03-16 00:39:21] Training GMM (k=50) on labelled features...
[2021-03-16 00:39:21] Training GMM (k=50) on unlabelled features...
[2021-03-16 00:39:21] Showing labelled: True (300/5000 visible)
[2021-03-16 00:39:21] Training mlp across 300 data points in blobsthin...
[2021-03-16 00:39:23] Training loss: 0.0214
[2021-03-16 00:39:23] Testing on 5000 data points...
[2021-03-16 00:39:23] Test accuracy: 0.9888
[2021-03-16 00:39:23] Score for 300 labels: 0.9888
[2021-03-16 00:39:23] Running: experiment 1
[2021-03-16 00:39:23] Showing labelled: True (50/5000 visible)
[2021-03-16 00:39:23] mlp: initialized 1482 parameters.
[2021-03-16 00:39:23] Creating trainer with model on device: cuda
[2021-03-16 00:39:23] Training mlp across 50 data points in blobsthin...
[2021-03-16 00:39:23] Training loss: 1.2687
[2021-03-16 00:39:23] Testing on 5000 data points...
[2021-03-16 00:39:23] Test accuracy: 0.4956
[2021-03-16 00:39:23] Score for 50 labels: 0.4956
[2021-03-16 00:39:23] Showing labelled: False (4950/5000 visible)
[2021-03-16 00:39:23] Found 4950 unlabelled features.
[2021-03-16 00:39:23] Searching core-set greedily...
[2021-03-16 00:39:25] Showing labelled: True (75/5000 visible)
[2021-03-16 00:39:25] Training mlp across 75 data points in blobsthin...
[2021-03-16 00:39:26] Training loss: 0.6352
[2021-03-16 00:39:26] Testing on 5000 data points...
[2021-03-16 00:39:26] Test accuracy: 0.8946
[2021-03-16 00:39:26] Score for 75 labels: 0.8946
[2021-03-16 00:39:26] Showing labelled: False (4925/5000 visible)
[2021-03-16 00:39:26] Found 4925 unlabelled features.
[2021-03-16 00:39:26] Searching core-set greedily...
[2021-03-16 00:39:28] Showing labelled: True (100/5000 visible)
[2021-03-16 00:39:28] Training mlp across 100 data points in blobsthin...
[2021-03-16 00:39:29] Training loss: 0.3782
[2021-03-16 00:39:29] Testing on 5000 data points...
[2021-03-16 00:39:29] Test accuracy: 0.8986
[2021-03-16 00:39:29] Score for 100 labels: 0.8986
[2021-03-16 00:39:29] Showing labelled: False (4900/5000 visible)
[2021-03-16 00:39:29] Found 4900 unlabelled features.
[2021-03-16 00:39:29] Searching core-set greedily...
[2021-03-16 00:39:31] Showing labelled: True (125/5000 visible)
[2021-03-16 00:39:31] Training mlp across 125 data points in blobsthin...
[2021-03-16 00:39:32] Training loss: 0.2485
[2021-03-16 00:39:32] Testing on 5000 data points...
[2021-03-16 00:39:32] Test accuracy: 0.9944
[2021-03-16 00:39:32] Score for 125 labels: 0.9944
[2021-03-16 00:39:32] Showing labelled: False (4875/5000 visible)
[2021-03-16 00:39:32] Found 4875 unlabelled features.
[2021-03-16 00:39:32] Searching core-set greedily...
[2021-03-16 00:39:35] Showing labelled: True (150/5000 visible)
[2021-03-16 00:39:35] Training mlp across 150 data points in blobsthin...
[2021-03-16 00:39:36] Training loss: 0.1676
[2021-03-16 00:39:36] Testing on 5000 data points...
[2021-03-16 00:39:36] Test accuracy: 0.9948
[2021-03-16 00:39:36] Score for 150 labels: 0.9948
[2021-03-16 00:39:36] Showing labelled: False (4850/5000 visible)
[2021-03-16 00:39:36] Found 4850 unlabelled features.
[2021-03-16 00:39:36] Searching core-set greedily...
[2021-03-16 00:39:40] Showing labelled: True (175/5000 visible)
[2021-03-16 00:39:40] Training mlp across 175 data points in blobsthin...
[2021-03-16 00:39:41] Training loss: 0.1221
[2021-03-16 00:39:41] Testing on 5000 data points...
[2021-03-16 00:39:41] Test accuracy: 0.9970
[2021-03-16 00:39:41] Score for 175 labels: 0.9970
[2021-03-16 00:39:41] Showing labelled: False (4825/5000 visible)
[2021-03-16 00:39:41] Found 4825 unlabelled features.
[2021-03-16 00:39:41] Searching core-set greedily...
[2021-03-16 00:39:45] Showing labelled: True (200/5000 visible)
[2021-03-16 00:39:45] Training mlp across 200 data points in blobsthin...
[2021-03-16 00:39:46] Training loss: 0.0838
[2021-03-16 00:39:46] Testing on 5000 data points...
[2021-03-16 00:39:46] Test accuracy: 0.9982
[2021-03-16 00:39:46] Score for 200 labels: 0.9982
[2021-03-16 00:39:46] Showing labelled: False (4800/5000 visible)
[2021-03-16 00:39:46] Found 4800 unlabelled features.
[2021-03-16 00:39:46] Searching core-set greedily...
[2021-03-16 00:39:52] Showing labelled: True (225/5000 visible)
[2021-03-16 00:39:52] Training mlp across 225 data points in blobsthin...
[2021-03-16 00:39:53] Training loss: 0.0590
[2021-03-16 00:39:53] Testing on 5000 data points...
[2021-03-16 00:39:53] Test accuracy: 0.9988
[2021-03-16 00:39:53] Score for 225 labels: 0.9988
[2021-03-16 00:39:53] Showing labelled: False (4775/5000 visible)
[2021-03-16 00:39:53] Found 4775 unlabelled features.
[2021-03-16 00:39:53] Searching core-set greedily...
[2021-03-16 00:39:58] Showing labelled: True (250/5000 visible)
[2021-03-16 00:39:58] Training mlp across 250 data points in blobsthin...
[2021-03-16 00:39:59] Training loss: 0.0445
[2021-03-16 00:39:59] Testing on 5000 data points...
[2021-03-16 00:39:59] Test accuracy: 0.9984
[2021-03-16 00:39:59] Score for 250 labels: 0.9984
[2021-03-16 00:39:59] Showing labelled: False (4750/5000 visible)
[2021-03-16 00:39:59] Found 4750 unlabelled features.
[2021-03-16 00:39:59] Searching core-set greedily...
[2021-03-16 00:40:04] Showing labelled: True (275/5000 visible)
[2021-03-16 00:40:04] Training mlp across 275 data points in blobsthin...
[2021-03-16 00:40:05] Training loss: 0.0370
[2021-03-16 00:40:05] Testing on 5000 data points...
[2021-03-16 00:40:05] Test accuracy: 0.9984
[2021-03-16 00:40:05] Score for 275 labels: 0.9984
[2021-03-16 00:40:05] Showing labelled: False (4725/5000 visible)
[2021-03-16 00:40:05] Found 4725 unlabelled features.
[2021-03-16 00:40:05] Searching core-set greedily...
[2021-03-16 00:40:10] Showing labelled: True (300/5000 visible)
[2021-03-16 00:40:10] Training mlp across 300 data points in blobsthin...
[2021-03-16 00:40:11] Training loss: 0.0313
[2021-03-16 00:40:11] Testing on 5000 data points...
[2021-03-16 00:40:12] Test accuracy: 0.9976
[2021-03-16 00:40:12] Score for 300 labels: 0.9976
[2021-03-16 00:40:12] Running: experiment 2
[2021-03-16 00:40:12] Showing labelled: True (50/5000 visible)
[2021-03-16 00:40:12] mlp: initialized 1482 parameters.
[2021-03-16 00:40:12] Creating trainer with model on device: cuda
[2021-03-16 00:40:12] Training mlp across 50 data points in blobsthin...
[2021-03-16 00:40:12] Training loss: 1.3530
[2021-03-16 00:40:12] Testing on 5000 data points...
[2021-03-16 00:40:12] Test accuracy: 0.3742
[2021-03-16 00:40:12] Score for 50 labels: 0.3742
[2021-03-16 00:40:12] Showing labelled: False (4950/5000 visible)
[2021-03-16 00:40:12] Found 4950 unlabelled features.
[2021-03-16 00:40:12] Showing labelled: True (75/5000 visible)
[2021-03-16 00:40:12] Training mlp across 75 data points in blobsthin...
[2021-03-16 00:40:12] Training loss: 0.6353
[2021-03-16 00:40:12] Testing on 5000 data points...
[2021-03-16 00:40:12] Test accuracy: 0.7000
[2021-03-16 00:40:12] Score for 75 labels: 0.7000
[2021-03-16 00:40:12] Showing labelled: False (4925/5000 visible)
[2021-03-16 00:40:12] Found 4925 unlabelled features.
[2021-03-16 00:40:12] Showing labelled: True (100/5000 visible)
[2021-03-16 00:40:12] Training mlp across 100 data points in blobsthin...
[2021-03-16 00:40:13] Training loss: 0.2629
[2021-03-16 00:40:13] Testing on 5000 data points...
[2021-03-16 00:40:13] Test accuracy: 0.7354
[2021-03-16 00:40:13] Score for 100 labels: 0.7354
[2021-03-16 00:40:13] Showing labelled: False (4900/5000 visible)
[2021-03-16 00:40:13] Found 4900 unlabelled features.
[2021-03-16 00:40:13] Showing labelled: True (125/5000 visible)
[2021-03-16 00:40:13] Training mlp across 125 data points in blobsthin...
[2021-03-16 00:40:14] Training loss: 0.1805
[2021-03-16 00:40:14] Testing on 5000 data points...
[2021-03-16 00:40:14] Test accuracy: 0.7950
[2021-03-16 00:40:14] Score for 125 labels: 0.7950
[2021-03-16 00:40:14] Showing labelled: False (4875/5000 visible)
[2021-03-16 00:40:14] Found 4875 unlabelled features.
[2021-03-16 00:40:14] Showing labelled: True (150/5000 visible)
[2021-03-16 00:40:14] Training mlp across 150 data points in blobsthin...
[2021-03-16 00:40:14] Training loss: 0.1109
[2021-03-16 00:40:14] Testing on 5000 data points...
[2021-03-16 00:40:15] Test accuracy: 0.8088
[2021-03-16 00:40:15] Score for 150 labels: 0.8088
[2021-03-16 00:40:15] Showing labelled: False (4850/5000 visible)
[2021-03-16 00:40:15] Found 4850 unlabelled features.
[2021-03-16 00:40:15] Showing labelled: True (175/5000 visible)
[2021-03-16 00:40:15] Training mlp across 175 data points in blobsthin...
[2021-03-16 00:40:15] Training loss: 0.0769
[2021-03-16 00:40:15] Testing on 5000 data points...
[2021-03-16 00:40:16] Test accuracy: 0.8694
[2021-03-16 00:40:16] Score for 175 labels: 0.8694
[2021-03-16 00:40:16] Showing labelled: False (4825/5000 visible)
[2021-03-16 00:40:16] Found 4825 unlabelled features.
[2021-03-16 00:40:16] Showing labelled: True (200/5000 visible)
[2021-03-16 00:40:16] Training mlp across 200 data points in blobsthin...
[2021-03-16 00:40:17] Training loss: 0.0540
[2021-03-16 00:40:17] Testing on 5000 data points...
[2021-03-16 00:40:17] Test accuracy: 0.8874
[2021-03-16 00:40:17] Score for 200 labels: 0.8874
[2021-03-16 00:40:17] Showing labelled: False (4800/5000 visible)
[2021-03-16 00:40:17] Found 4800 unlabelled features.
[2021-03-16 00:40:17] Showing labelled: True (225/5000 visible)
[2021-03-16 00:40:17] Training mlp across 225 data points in blobsthin...
[2021-03-16 00:40:18] Training loss: 0.0357
[2021-03-16 00:40:18] Testing on 5000 data points...
[2021-03-16 00:40:18] Test accuracy: 0.8926
[2021-03-16 00:40:18] Score for 225 labels: 0.8926
[2021-03-16 00:40:18] Showing labelled: False (4775/5000 visible)
[2021-03-16 00:40:18] Found 4775 unlabelled features.
[2021-03-16 00:40:18] Showing labelled: True (250/5000 visible)
[2021-03-16 00:40:18] Training mlp across 250 data points in blobsthin...
[2021-03-16 00:40:19] Training loss: 0.0287
[2021-03-16 00:40:19] Testing on 5000 data points...
[2021-03-16 00:40:19] Test accuracy: 0.8952
[2021-03-16 00:40:19] Score for 250 labels: 0.8952
[2021-03-16 00:40:19] Showing labelled: False (4750/5000 visible)
[2021-03-16 00:40:19] Found 4750 unlabelled features.
[2021-03-16 00:40:19] Showing labelled: True (275/5000 visible)
[2021-03-16 00:40:19] Training mlp across 275 data points in blobsthin...
[2021-03-16 00:40:21] Training loss: 0.0207
[2021-03-16 00:40:21] Testing on 5000 data points...
[2021-03-16 00:40:21] Test accuracy: 0.8958
[2021-03-16 00:40:21] Score for 275 labels: 0.8958
[2021-03-16 00:40:21] Showing labelled: False (4725/5000 visible)
[2021-03-16 00:40:21] Found 4725 unlabelled features.
[2021-03-16 00:40:21] Showing labelled: True (300/5000 visible)
[2021-03-16 00:40:21] Training mlp across 300 data points in blobsthin...
[2021-03-16 00:40:22] Training loss: 0.0163
[2021-03-16 00:40:22] Testing on 5000 data points...
[2021-03-16 00:40:22] Test accuracy: 0.8956
[2021-03-16 00:40:22] Score for 300 labels: 0.8956
[2021-03-16 00:40:22] Updated results: ../results/sp_coreset/results.json
