[2021-03-15 01:42:57] Seeded: 5
[2021-03-15 01:42:57] Created experiment 0:
[2021-03-15 01:42:57]  - Model: mlp
[2021-03-15 01:42:57]  - Acquisition function: greedy-coreset
[2021-03-15 01:42:57] Created experiment 1:
[2021-03-15 01:42:57]  - Model: mlp
[2021-03-15 01:42:57]  - Acquisition function: random
[2021-03-15 01:42:57] Created experiment 2:
[2021-03-15 01:42:57]  - Model: mlp
[2021-03-15 01:42:57]  - Acquisition function: least-confidence
[2021-03-15 01:42:57] Created experiment 3:
[2021-03-15 01:42:57]  - Model: mlp
[2021-03-15 01:42:57]  - Acquisition function: max-entropy
[2021-03-15 01:42:57] Created experiment 4:
[2021-03-15 01:42:57]  - Model: mlp
[2021-03-15 01:42:57]  - Acquisition function: p-coreset
[2021-03-15 01:42:57] Loading blobsthin test set...
[2021-03-15 01:42:57] Generated 5000 blobs:
[2021-03-15 01:42:57] Class 0: 200
[2021-03-15 01:42:57] Class 1: 300
[2021-03-15 01:42:57] Class 2: 1000
[2021-03-15 01:42:57] Class 3: 100
[2021-03-15 01:42:57] Class 4: 50
[2021-03-15 01:42:57] Class 5: 200
[2021-03-15 01:42:57] Class 6: 400
[2021-03-15 01:42:57] Class 7: 1800
[2021-03-15 01:42:57] Class 8: 50
[2021-03-15 01:42:57] Class 9: 900
[2021-03-15 01:42:58] Saved figure to: ../data/blobsthin_data.png
[2021-03-15 01:42:58] Generated 5000 test points.
[2021-03-15 01:42:58] Saved figure to: ../data/blobsthin_test.png
[2021-03-15 01:42:58] Experiment repeat 1/5
[2021-03-15 01:42:58] Randomly labelled 50/5000
[2021-03-15 01:42:58] Showing labelled: True (50/5000 visible)
[2021-03-15 01:42:58] Running: experiment 0
[2021-03-15 01:42:58] Showing labelled: True (50/5000 visible)
[2021-03-15 01:42:58] mlp: initialized 1482 parameters.
[2021-03-15 01:42:58] Creating trainer with model on device: cuda
[2021-03-15 01:43:03] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:43:03] Training loss: 1.5627
[2021-03-15 01:43:03] Testing on 5000 data points...
[2021-03-15 01:43:03] Test accuracy: 0.5000
[2021-03-15 01:43:03] Score for 50 labels: 0.5000
[2021-03-15 01:43:03] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:43:03] Found 4950 unlabelled features.
[2021-03-15 01:43:03] Searching core-set greedily...
[2021-03-15 01:43:09] Showing labelled: True (100/5000 visible)
[2021-03-15 01:43:09] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:43:09] Training loss: 0.7615
[2021-03-15 01:43:09] Testing on 5000 data points...
[2021-03-15 01:43:10] Test accuracy: 0.8958
[2021-03-15 01:43:10] Score for 100 labels: 0.8958
[2021-03-15 01:43:10] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:43:10] Found 4900 unlabelled features.
[2021-03-15 01:43:10] Searching core-set greedily...
[2021-03-15 01:43:15] Showing labelled: True (150/5000 visible)
[2021-03-15 01:43:15] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:43:16] Training loss: 0.3642
[2021-03-15 01:43:16] Testing on 5000 data points...
[2021-03-15 01:43:16] Test accuracy: 0.8994
[2021-03-15 01:43:16] Score for 150 labels: 0.8994
[2021-03-15 01:43:16] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:43:16] Found 4850 unlabelled features.
[2021-03-15 01:43:16] Searching core-set greedily...
[2021-03-15 01:43:25] Showing labelled: True (200/5000 visible)
[2021-03-15 01:43:25] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:43:26] Training loss: 0.2114
[2021-03-15 01:43:26] Testing on 5000 data points...
[2021-03-15 01:43:26] Test accuracy: 0.9712
[2021-03-15 01:43:26] Score for 200 labels: 0.9712
[2021-03-15 01:43:26] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:43:26] Found 4800 unlabelled features.
[2021-03-15 01:43:26] Searching core-set greedily...
[2021-03-15 01:43:35] Showing labelled: True (250/5000 visible)
[2021-03-15 01:43:35] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:43:37] Training loss: 0.1256
[2021-03-15 01:43:37] Testing on 5000 data points...
[2021-03-15 01:43:37] Test accuracy: 0.9972
[2021-03-15 01:43:37] Score for 250 labels: 0.9972
[2021-03-15 01:43:37] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:43:37] Found 4750 unlabelled features.
[2021-03-15 01:43:37] Searching core-set greedily...
[2021-03-15 01:43:47] Showing labelled: True (300/5000 visible)
[2021-03-15 01:43:47] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:43:49] Training loss: 0.0709
[2021-03-15 01:43:49] Testing on 5000 data points...
[2021-03-15 01:43:49] Test accuracy: 0.9972
[2021-03-15 01:43:49] Score for 300 labels: 0.9972
[2021-03-15 01:43:49] Running: experiment 1
[2021-03-15 01:43:49] Showing labelled: True (50/5000 visible)
[2021-03-15 01:43:49] mlp: initialized 1482 parameters.
[2021-03-15 01:43:49] Creating trainer with model on device: cuda
[2021-03-15 01:43:49] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:43:49] Training loss: 1.4116
[2021-03-15 01:43:49] Testing on 5000 data points...
[2021-03-15 01:43:49] Test accuracy: 0.5000
[2021-03-15 01:43:49] Score for 50 labels: 0.5000
[2021-03-15 01:43:49] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:43:49] Found 4950 unlabelled features.
[2021-03-15 01:43:49] Showing labelled: True (100/5000 visible)
[2021-03-15 01:43:49] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:43:50] Training loss: 0.5394
[2021-03-15 01:43:50] Testing on 5000 data points...
[2021-03-15 01:43:50] Test accuracy: 0.7656
[2021-03-15 01:43:50] Score for 100 labels: 0.7656
[2021-03-15 01:43:50] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:43:50] Found 4900 unlabelled features.
[2021-03-15 01:43:50] Showing labelled: True (150/5000 visible)
[2021-03-15 01:43:50] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:43:51] Training loss: 0.2231
[2021-03-15 01:43:51] Testing on 5000 data points...
[2021-03-15 01:43:51] Test accuracy: 0.8964
[2021-03-15 01:43:51] Score for 150 labels: 0.8964
[2021-03-15 01:43:51] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:43:51] Found 4850 unlabelled features.
[2021-03-15 01:43:51] Showing labelled: True (200/5000 visible)
[2021-03-15 01:43:51] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:43:53] Training loss: 0.1196
[2021-03-15 01:43:53] Testing on 5000 data points...
[2021-03-15 01:43:53] Test accuracy: 0.9004
[2021-03-15 01:43:53] Score for 200 labels: 0.9004
[2021-03-15 01:43:53] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:43:53] Found 4800 unlabelled features.
[2021-03-15 01:43:53] Showing labelled: True (250/5000 visible)
[2021-03-15 01:43:53] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:43:54] Training loss: 0.0720
[2021-03-15 01:43:54] Testing on 5000 data points...
[2021-03-15 01:43:54] Test accuracy: 0.9472
[2021-03-15 01:43:54] Score for 250 labels: 0.9472
[2021-03-15 01:43:54] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:43:54] Found 4750 unlabelled features.
[2021-03-15 01:43:54] Showing labelled: True (300/5000 visible)
[2021-03-15 01:43:54] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:43:56] Training loss: 0.0465
[2021-03-15 01:43:56] Testing on 5000 data points...
[2021-03-15 01:43:56] Test accuracy: 0.9912
[2021-03-15 01:43:56] Score for 300 labels: 0.9912
[2021-03-15 01:43:56] Running: experiment 2
[2021-03-15 01:43:56] Showing labelled: True (50/5000 visible)
[2021-03-15 01:43:56] mlp: initialized 1482 parameters.
[2021-03-15 01:43:56] Creating trainer with model on device: cuda
[2021-03-15 01:43:56] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:43:57] Training loss: 1.4136
[2021-03-15 01:43:57] Testing on 5000 data points...
[2021-03-15 01:43:57] Test accuracy: 0.5000
[2021-03-15 01:43:57] Score for 50 labels: 0.5000
[2021-03-15 01:43:57] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:43:57] Found 4950 unlabelled features.
[2021-03-15 01:43:57] Showing labelled: True (100/5000 visible)
[2021-03-15 01:43:57] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:43:57] Training loss: 0.5146
[2021-03-15 01:43:57] Testing on 5000 data points...
[2021-03-15 01:43:58] Test accuracy: 0.6852
[2021-03-15 01:43:58] Score for 100 labels: 0.6852
[2021-03-15 01:43:58] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:43:58] Found 4900 unlabelled features.
[2021-03-15 01:43:58] Showing labelled: True (150/5000 visible)
[2021-03-15 01:43:58] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:43:59] Training loss: 0.1935
[2021-03-15 01:43:59] Testing on 5000 data points...
[2021-03-15 01:43:59] Test accuracy: 0.7854
[2021-03-15 01:43:59] Score for 150 labels: 0.7854
[2021-03-15 01:43:59] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:43:59] Found 4850 unlabelled features.
[2021-03-15 01:43:59] Showing labelled: True (200/5000 visible)
[2021-03-15 01:43:59] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:44:00] Training loss: 0.1306
[2021-03-15 01:44:00] Testing on 5000 data points...
[2021-03-15 01:44:00] Test accuracy: 0.7970
[2021-03-15 01:44:00] Score for 200 labels: 0.7970
[2021-03-15 01:44:00] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:44:00] Found 4800 unlabelled features.
[2021-03-15 01:44:00] Showing labelled: True (250/5000 visible)
[2021-03-15 01:44:00] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:44:02] Training loss: 0.0942
[2021-03-15 01:44:02] Testing on 5000 data points...
[2021-03-15 01:44:02] Test accuracy: 0.7992
[2021-03-15 01:44:02] Score for 250 labels: 0.7992
[2021-03-15 01:44:02] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:44:02] Found 4750 unlabelled features.
[2021-03-15 01:44:02] Showing labelled: True (300/5000 visible)
[2021-03-15 01:44:02] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:44:04] Training loss: 0.0916
[2021-03-15 01:44:04] Testing on 5000 data points...
[2021-03-15 01:44:04] Test accuracy: 0.8968
[2021-03-15 01:44:04] Score for 300 labels: 0.8968
[2021-03-15 01:44:04] Running: experiment 3
[2021-03-15 01:44:04] Showing labelled: True (50/5000 visible)
[2021-03-15 01:44:04] mlp: initialized 1482 parameters.
[2021-03-15 01:44:04] Creating trainer with model on device: cuda
[2021-03-15 01:44:04] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:44:04] Training loss: 1.3943
[2021-03-15 01:44:04] Testing on 5000 data points...
[2021-03-15 01:44:04] Test accuracy: 0.5000
[2021-03-15 01:44:04] Score for 50 labels: 0.5000
[2021-03-15 01:44:04] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:44:04] Found 4950 unlabelled features.
[2021-03-15 01:44:04] Showing labelled: True (100/5000 visible)
[2021-03-15 01:44:04] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:44:05] Training loss: 0.4365
[2021-03-15 01:44:05] Testing on 5000 data points...
[2021-03-15 01:44:05] Test accuracy: 0.6490
[2021-03-15 01:44:05] Score for 100 labels: 0.6490
[2021-03-15 01:44:05] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:44:05] Found 4900 unlabelled features.
[2021-03-15 01:44:05] Showing labelled: True (150/5000 visible)
[2021-03-15 01:44:05] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:44:06] Training loss: 0.1468
[2021-03-15 01:44:06] Testing on 5000 data points...
[2021-03-15 01:44:06] Test accuracy: 0.6918
[2021-03-15 01:44:06] Score for 150 labels: 0.6918
[2021-03-15 01:44:06] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:44:06] Found 4850 unlabelled features.
[2021-03-15 01:44:06] Showing labelled: True (200/5000 visible)
[2021-03-15 01:44:06] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:44:07] Training loss: 0.1184
[2021-03-15 01:44:07] Testing on 5000 data points...
[2021-03-15 01:44:07] Test accuracy: 0.7956
[2021-03-15 01:44:07] Score for 200 labels: 0.7956
[2021-03-15 01:44:07] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:44:07] Found 4800 unlabelled features.
[2021-03-15 01:44:07] Showing labelled: True (250/5000 visible)
[2021-03-15 01:44:07] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:44:09] Training loss: 0.1125
[2021-03-15 01:44:09] Testing on 5000 data points...
[2021-03-15 01:44:09] Test accuracy: 0.8968
[2021-03-15 01:44:09] Score for 250 labels: 0.8968
[2021-03-15 01:44:09] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:44:09] Found 4750 unlabelled features.
[2021-03-15 01:44:09] Showing labelled: True (300/5000 visible)
[2021-03-15 01:44:09] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:44:11] Training loss: 0.0837
[2021-03-15 01:44:11] Testing on 5000 data points...
[2021-03-15 01:44:11] Test accuracy: 0.8982
[2021-03-15 01:44:11] Score for 300 labels: 0.8982
[2021-03-15 01:44:11] Running: experiment 4
[2021-03-15 01:44:11] Showing labelled: True (50/5000 visible)
[2021-03-15 01:44:11] mlp: initialized 1482 parameters.
[2021-03-15 01:44:11] Creating trainer with model on device: cuda
[2021-03-15 01:44:11] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:44:12] Training loss: 1.4412
[2021-03-15 01:44:12] Testing on 5000 data points...
[2021-03-15 01:44:12] Test accuracy: 0.5000
[2021-03-15 01:44:12] Score for 50 labels: 0.5000
[2021-03-15 01:44:12] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:44:12] Found 4950 unlabelled features.
[2021-03-15 01:44:12] Training GMM (k=50) on labelled features...
[2021-03-15 01:44:12] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:44:12] Scoring random batches...
[2021-03-15 01:44:12] Showing labelled: True (100/5000 visible)
[2021-03-15 01:44:12] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:44:13] Training loss: 0.5515
[2021-03-15 01:44:13] Testing on 5000 data points...
[2021-03-15 01:44:13] Test accuracy: 0.7056
[2021-03-15 01:44:13] Score for 100 labels: 0.7056
[2021-03-15 01:44:13] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:44:13] Found 4900 unlabelled features.
[2021-03-15 01:44:13] Training GMM (k=50) on labelled features...
[2021-03-15 01:44:13] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:44:13] Scoring random batches...
[2021-03-15 01:44:13] Showing labelled: True (150/5000 visible)
[2021-03-15 01:44:13] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:44:14] Training loss: 0.2536
[2021-03-15 01:44:14] Testing on 5000 data points...
[2021-03-15 01:44:14] Test accuracy: 0.8960
[2021-03-15 01:44:14] Score for 150 labels: 0.8960
[2021-03-15 01:44:14] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:44:14] Found 4850 unlabelled features.
[2021-03-15 01:44:14] Training GMM (k=50) on labelled features...
[2021-03-15 01:44:14] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:44:15] Scoring random batches...
[2021-03-15 01:44:15] Showing labelled: True (200/5000 visible)
[2021-03-15 01:44:15] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:44:16] Training loss: 0.1297
[2021-03-15 01:44:16] Testing on 5000 data points...
[2021-03-15 01:44:16] Test accuracy: 0.9678
[2021-03-15 01:44:16] Score for 200 labels: 0.9678
[2021-03-15 01:44:16] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:44:16] Found 4800 unlabelled features.
[2021-03-15 01:44:16] Training GMM (k=50) on labelled features...
[2021-03-15 01:44:16] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:44:17] Scoring random batches...
[2021-03-15 01:44:17] Showing labelled: True (250/5000 visible)
[2021-03-15 01:44:17] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:44:18] Training loss: 0.0683
[2021-03-15 01:44:18] Testing on 5000 data points...
[2021-03-15 01:44:18] Test accuracy: 0.9920
[2021-03-15 01:44:18] Score for 250 labels: 0.9920
[2021-03-15 01:44:18] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:44:18] Found 4750 unlabelled features.
[2021-03-15 01:44:18] Training GMM (k=50) on labelled features...
[2021-03-15 01:44:19] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:44:19] Scoring random batches...
[2021-03-15 01:44:19] Showing labelled: True (300/5000 visible)
[2021-03-15 01:44:19] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:44:21] Training loss: 0.0394
[2021-03-15 01:44:21] Testing on 5000 data points...
[2021-03-15 01:44:21] Test accuracy: 0.9946
[2021-03-15 01:44:21] Score for 300 labels: 0.9946
[2021-03-15 01:44:21] Experiment repeat 2/5
[2021-03-15 01:44:21] Randomly labelled 50/5000
[2021-03-15 01:44:21] Showing labelled: True (50/5000 visible)
[2021-03-15 01:44:21] Running: experiment 0
[2021-03-15 01:44:21] Showing labelled: True (50/5000 visible)
[2021-03-15 01:44:21] mlp: initialized 1482 parameters.
[2021-03-15 01:44:21] Creating trainer with model on device: cuda
[2021-03-15 01:44:21] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:44:21] Training loss: 1.1875
[2021-03-15 01:44:21] Testing on 5000 data points...
[2021-03-15 01:44:21] Test accuracy: 0.4948
[2021-03-15 01:44:21] Score for 50 labels: 0.4948
[2021-03-15 01:44:21] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:44:21] Found 4950 unlabelled features.
[2021-03-15 01:44:22] Searching core-set greedily...
[2021-03-15 01:44:28] Showing labelled: True (100/5000 visible)
[2021-03-15 01:44:28] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:44:29] Training loss: 0.6911
[2021-03-15 01:44:29] Testing on 5000 data points...
[2021-03-15 01:44:29] Test accuracy: 0.8330
[2021-03-15 01:44:29] Score for 100 labels: 0.8330
[2021-03-15 01:44:29] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:44:29] Found 4900 unlabelled features.
[2021-03-15 01:44:29] Searching core-set greedily...
[2021-03-15 01:44:36] Showing labelled: True (150/5000 visible)
[2021-03-15 01:44:36] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:44:37] Training loss: 0.3425
[2021-03-15 01:44:37] Testing on 5000 data points...
[2021-03-15 01:44:37] Test accuracy: 0.9134
[2021-03-15 01:44:37] Score for 150 labels: 0.9134
[2021-03-15 01:44:37] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:44:37] Found 4850 unlabelled features.
[2021-03-15 01:44:37] Searching core-set greedily...
[2021-03-15 01:44:46] Showing labelled: True (200/5000 visible)
[2021-03-15 01:44:46] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:44:47] Training loss: 0.1911
[2021-03-15 01:44:47] Testing on 5000 data points...
[2021-03-15 01:44:47] Test accuracy: 0.9908
[2021-03-15 01:44:47] Score for 200 labels: 0.9908
[2021-03-15 01:44:47] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:44:47] Found 4800 unlabelled features.
[2021-03-15 01:44:47] Searching core-set greedily...
[2021-03-15 01:44:57] Showing labelled: True (250/5000 visible)
[2021-03-15 01:44:57] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:44:59] Training loss: 0.1121
[2021-03-15 01:44:59] Testing on 5000 data points...
[2021-03-15 01:44:59] Test accuracy: 0.9984
[2021-03-15 01:44:59] Score for 250 labels: 0.9984
[2021-03-15 01:44:59] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:44:59] Found 4750 unlabelled features.
[2021-03-15 01:44:59] Searching core-set greedily...
[2021-03-15 01:45:09] Showing labelled: True (300/5000 visible)
[2021-03-15 01:45:09] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:45:11] Training loss: 0.0691
[2021-03-15 01:45:11] Testing on 5000 data points...
[2021-03-15 01:45:11] Test accuracy: 0.9982
[2021-03-15 01:45:11] Score for 300 labels: 0.9982
[2021-03-15 01:45:11] Running: experiment 1
[2021-03-15 01:45:11] Showing labelled: True (50/5000 visible)
[2021-03-15 01:45:11] mlp: initialized 1482 parameters.
[2021-03-15 01:45:11] Creating trainer with model on device: cuda
[2021-03-15 01:45:11] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:45:12] Training loss: 1.2118
[2021-03-15 01:45:12] Testing on 5000 data points...
[2021-03-15 01:45:12] Test accuracy: 0.3832
[2021-03-15 01:45:12] Score for 50 labels: 0.3832
[2021-03-15 01:45:12] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:45:12] Found 4950 unlabelled features.
[2021-03-15 01:45:12] Showing labelled: True (100/5000 visible)
[2021-03-15 01:45:12] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:45:12] Training loss: 0.5972
[2021-03-15 01:45:12] Testing on 5000 data points...
[2021-03-15 01:45:12] Test accuracy: 0.7830
[2021-03-15 01:45:12] Score for 100 labels: 0.7830
[2021-03-15 01:45:12] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:45:12] Found 4900 unlabelled features.
[2021-03-15 01:45:12] Showing labelled: True (150/5000 visible)
[2021-03-15 01:45:12] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:45:13] Training loss: 0.2630
[2021-03-15 01:45:13] Testing on 5000 data points...
[2021-03-15 01:45:13] Test accuracy: 0.8300
[2021-03-15 01:45:13] Score for 150 labels: 0.8300
[2021-03-15 01:45:13] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:45:13] Found 4850 unlabelled features.
[2021-03-15 01:45:13] Showing labelled: True (200/5000 visible)
[2021-03-15 01:45:13] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:45:15] Training loss: 0.1355
[2021-03-15 01:45:15] Testing on 5000 data points...
[2021-03-15 01:45:15] Test accuracy: 0.8948
[2021-03-15 01:45:15] Score for 200 labels: 0.8948
[2021-03-15 01:45:15] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:45:15] Found 4800 unlabelled features.
[2021-03-15 01:45:15] Showing labelled: True (250/5000 visible)
[2021-03-15 01:45:15] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:45:16] Training loss: 0.0843
[2021-03-15 01:45:16] Testing on 5000 data points...
[2021-03-15 01:45:16] Test accuracy: 0.9824
[2021-03-15 01:45:16] Score for 250 labels: 0.9824
[2021-03-15 01:45:16] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:45:16] Found 4750 unlabelled features.
[2021-03-15 01:45:16] Showing labelled: True (300/5000 visible)
[2021-03-15 01:45:16] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:45:18] Training loss: 0.0453
[2021-03-15 01:45:18] Testing on 5000 data points...
[2021-03-15 01:45:18] Test accuracy: 0.9912
[2021-03-15 01:45:18] Score for 300 labels: 0.9912
[2021-03-15 01:45:18] Running: experiment 2
[2021-03-15 01:45:18] Showing labelled: True (50/5000 visible)
[2021-03-15 01:45:18] mlp: initialized 1482 parameters.
[2021-03-15 01:45:18] Creating trainer with model on device: cuda
[2021-03-15 01:45:18] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:45:19] Training loss: 1.2220
[2021-03-15 01:45:19] Testing on 5000 data points...
[2021-03-15 01:45:19] Test accuracy: 0.5912
[2021-03-15 01:45:19] Score for 50 labels: 0.5912
[2021-03-15 01:45:19] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:45:19] Found 4950 unlabelled features.
[2021-03-15 01:45:19] Showing labelled: True (100/5000 visible)
[2021-03-15 01:45:19] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:45:20] Training loss: 0.4599
[2021-03-15 01:45:20] Testing on 5000 data points...
[2021-03-15 01:45:20] Test accuracy: 0.6000
[2021-03-15 01:45:20] Score for 100 labels: 0.6000
[2021-03-15 01:45:20] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:45:20] Found 4900 unlabelled features.
[2021-03-15 01:45:20] Showing labelled: True (150/5000 visible)
[2021-03-15 01:45:20] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:45:21] Training loss: 0.1589
[2021-03-15 01:45:21] Testing on 5000 data points...
[2021-03-15 01:45:21] Test accuracy: 0.6980
[2021-03-15 01:45:21] Score for 150 labels: 0.6980
[2021-03-15 01:45:21] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:45:21] Found 4850 unlabelled features.
[2021-03-15 01:45:21] Showing labelled: True (200/5000 visible)
[2021-03-15 01:45:21] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:45:22] Training loss: 0.0893
[2021-03-15 01:45:22] Testing on 5000 data points...
[2021-03-15 01:45:22] Test accuracy: 0.7000
[2021-03-15 01:45:22] Score for 200 labels: 0.7000
[2021-03-15 01:45:22] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:45:22] Found 4800 unlabelled features.
[2021-03-15 01:45:22] Showing labelled: True (250/5000 visible)
[2021-03-15 01:45:22] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:45:24] Training loss: 0.0924
[2021-03-15 01:45:24] Testing on 5000 data points...
[2021-03-15 01:45:24] Test accuracy: 0.8000
[2021-03-15 01:45:24] Score for 250 labels: 0.8000
[2021-03-15 01:45:24] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:45:24] Found 4750 unlabelled features.
[2021-03-15 01:45:24] Showing labelled: True (300/5000 visible)
[2021-03-15 01:45:24] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:45:26] Training loss: 0.0624
[2021-03-15 01:45:26] Testing on 5000 data points...
[2021-03-15 01:45:26] Test accuracy: 0.9000
[2021-03-15 01:45:26] Score for 300 labels: 0.9000
[2021-03-15 01:45:26] Running: experiment 3
[2021-03-15 01:45:26] Showing labelled: True (50/5000 visible)
[2021-03-15 01:45:26] mlp: initialized 1482 parameters.
[2021-03-15 01:45:26] Creating trainer with model on device: cuda
[2021-03-15 01:45:26] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:45:26] Training loss: 1.2875
[2021-03-15 01:45:26] Testing on 5000 data points...
[2021-03-15 01:45:26] Test accuracy: 0.5320
[2021-03-15 01:45:26] Score for 50 labels: 0.5320
[2021-03-15 01:45:26] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:45:26] Found 4950 unlabelled features.
[2021-03-15 01:45:26] Showing labelled: True (100/5000 visible)
[2021-03-15 01:45:26] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:45:27] Training loss: 0.4971
[2021-03-15 01:45:27] Testing on 5000 data points...
[2021-03-15 01:45:27] Test accuracy: 0.5998
[2021-03-15 01:45:27] Score for 100 labels: 0.5998
[2021-03-15 01:45:27] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:45:27] Found 4900 unlabelled features.
[2021-03-15 01:45:27] Showing labelled: True (150/5000 visible)
[2021-03-15 01:45:27] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:45:28] Training loss: 0.1510
[2021-03-15 01:45:28] Testing on 5000 data points...
[2021-03-15 01:45:28] Test accuracy: 0.6990
[2021-03-15 01:45:28] Score for 150 labels: 0.6990
[2021-03-15 01:45:28] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:45:28] Found 4850 unlabelled features.
[2021-03-15 01:45:28] Showing labelled: True (200/5000 visible)
[2021-03-15 01:45:28] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:45:30] Training loss: 0.0821
[2021-03-15 01:45:30] Testing on 5000 data points...
[2021-03-15 01:45:30] Test accuracy: 0.7000
[2021-03-15 01:45:30] Score for 200 labels: 0.7000
[2021-03-15 01:45:30] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:45:30] Found 4800 unlabelled features.
[2021-03-15 01:45:30] Showing labelled: True (250/5000 visible)
[2021-03-15 01:45:30] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:45:31] Training loss: 0.0587
[2021-03-15 01:45:31] Testing on 5000 data points...
[2021-03-15 01:45:31] Test accuracy: 0.7992
[2021-03-15 01:45:31] Score for 250 labels: 0.7992
[2021-03-15 01:45:31] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:45:31] Found 4750 unlabelled features.
[2021-03-15 01:45:31] Showing labelled: True (300/5000 visible)
[2021-03-15 01:45:31] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:45:33] Training loss: 0.0711
[2021-03-15 01:45:33] Testing on 5000 data points...
[2021-03-15 01:45:33] Test accuracy: 0.7998
[2021-03-15 01:45:33] Score for 300 labels: 0.7998
[2021-03-15 01:45:33] Running: experiment 4
[2021-03-15 01:45:33] Showing labelled: True (50/5000 visible)
[2021-03-15 01:45:33] mlp: initialized 1482 parameters.
[2021-03-15 01:45:33] Creating trainer with model on device: cuda
[2021-03-15 01:45:33] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:45:34] Training loss: 1.2395
[2021-03-15 01:45:34] Testing on 5000 data points...
[2021-03-15 01:45:34] Test accuracy: 0.4820
[2021-03-15 01:45:34] Score for 50 labels: 0.4820
[2021-03-15 01:45:34] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:45:34] Found 4950 unlabelled features.
[2021-03-15 01:45:34] Training GMM (k=50) on labelled features...
[2021-03-15 01:45:34] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:45:34] Scoring random batches...
[2021-03-15 01:45:34] Showing labelled: True (100/5000 visible)
[2021-03-15 01:45:34] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:45:35] Training loss: 0.6807
[2021-03-15 01:45:35] Testing on 5000 data points...
[2021-03-15 01:45:35] Test accuracy: 0.6862
[2021-03-15 01:45:35] Score for 100 labels: 0.6862
[2021-03-15 01:45:35] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:45:35] Found 4900 unlabelled features.
[2021-03-15 01:45:35] Training GMM (k=50) on labelled features...
[2021-03-15 01:45:35] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:45:35] Scoring random batches...
[2021-03-15 01:45:35] Showing labelled: True (150/5000 visible)
[2021-03-15 01:45:35] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:45:36] Training loss: 0.2916
[2021-03-15 01:45:36] Testing on 5000 data points...
[2021-03-15 01:45:36] Test accuracy: 0.8862
[2021-03-15 01:45:36] Score for 150 labels: 0.8862
[2021-03-15 01:45:36] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:45:36] Found 4850 unlabelled features.
[2021-03-15 01:45:36] Training GMM (k=50) on labelled features...
[2021-03-15 01:45:36] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:45:37] Scoring random batches...
[2021-03-15 01:45:37] Showing labelled: True (200/5000 visible)
[2021-03-15 01:45:37] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:45:38] Training loss: 0.1426
[2021-03-15 01:45:38] Testing on 5000 data points...
[2021-03-15 01:45:38] Test accuracy: 0.8984
[2021-03-15 01:45:38] Score for 200 labels: 0.8984
[2021-03-15 01:45:38] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:45:38] Found 4800 unlabelled features.
[2021-03-15 01:45:38] Training GMM (k=50) on labelled features...
[2021-03-15 01:45:38] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:45:39] Scoring random batches...
[2021-03-15 01:45:39] Showing labelled: True (250/5000 visible)
[2021-03-15 01:45:39] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:45:40] Training loss: 0.0915
[2021-03-15 01:45:40] Testing on 5000 data points...
[2021-03-15 01:45:40] Test accuracy: 0.9032
[2021-03-15 01:45:40] Score for 250 labels: 0.9032
[2021-03-15 01:45:40] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:45:40] Found 4750 unlabelled features.
[2021-03-15 01:45:40] Training GMM (k=50) on labelled features...
[2021-03-15 01:45:41] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:45:41] Scoring random batches...
[2021-03-15 01:45:41] Showing labelled: True (300/5000 visible)
[2021-03-15 01:45:41] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:45:43] Training loss: 0.0560
[2021-03-15 01:45:43] Testing on 5000 data points...
[2021-03-15 01:45:43] Test accuracy: 0.9528
[2021-03-15 01:45:43] Score for 300 labels: 0.9528
[2021-03-15 01:45:43] Experiment repeat 3/5
[2021-03-15 01:45:43] Randomly labelled 50/5000
[2021-03-15 01:45:43] Showing labelled: True (50/5000 visible)
[2021-03-15 01:45:43] Running: experiment 0
[2021-03-15 01:45:43] Showing labelled: True (50/5000 visible)
[2021-03-15 01:45:43] mlp: initialized 1482 parameters.
[2021-03-15 01:45:43] Creating trainer with model on device: cuda
[2021-03-15 01:45:43] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:45:44] Training loss: 1.3660
[2021-03-15 01:45:44] Testing on 5000 data points...
[2021-03-15 01:45:44] Test accuracy: 0.4000
[2021-03-15 01:45:44] Score for 50 labels: 0.4000
[2021-03-15 01:45:44] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:45:44] Found 4950 unlabelled features.
[2021-03-15 01:45:44] Searching core-set greedily...
[2021-03-15 01:45:48] Showing labelled: True (100/5000 visible)
[2021-03-15 01:45:48] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:45:49] Training loss: 0.7380
[2021-03-15 01:45:49] Testing on 5000 data points...
[2021-03-15 01:45:49] Test accuracy: 0.8932
[2021-03-15 01:45:49] Score for 100 labels: 0.8932
[2021-03-15 01:45:49] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:45:49] Found 4900 unlabelled features.
[2021-03-15 01:45:49] Searching core-set greedily...
[2021-03-15 01:45:57] Showing labelled: True (150/5000 visible)
[2021-03-15 01:45:57] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:45:58] Training loss: 0.3715
[2021-03-15 01:45:58] Testing on 5000 data points...
[2021-03-15 01:45:58] Test accuracy: 0.9006
[2021-03-15 01:45:58] Score for 150 labels: 0.9006
[2021-03-15 01:45:58] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:45:58] Found 4850 unlabelled features.
[2021-03-15 01:45:58] Searching core-set greedily...
[2021-03-15 01:46:07] Showing labelled: True (200/5000 visible)
[2021-03-15 01:46:07] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:46:09] Training loss: 0.2167
[2021-03-15 01:46:09] Testing on 5000 data points...
[2021-03-15 01:46:09] Test accuracy: 0.9894
[2021-03-15 01:46:09] Score for 200 labels: 0.9894
[2021-03-15 01:46:09] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:46:09] Found 4800 unlabelled features.
[2021-03-15 01:46:09] Searching core-set greedily...
[2021-03-15 01:46:20] Showing labelled: True (250/5000 visible)
[2021-03-15 01:46:20] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:46:21] Training loss: 0.1264
[2021-03-15 01:46:21] Testing on 5000 data points...
[2021-03-15 01:46:22] Test accuracy: 0.9970
[2021-03-15 01:46:22] Score for 250 labels: 0.9970
[2021-03-15 01:46:22] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:46:22] Found 4750 unlabelled features.
[2021-03-15 01:46:22] Searching core-set greedily...
[2021-03-15 01:46:33] Showing labelled: True (300/5000 visible)
[2021-03-15 01:46:33] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:46:35] Training loss: 0.0781
[2021-03-15 01:46:35] Testing on 5000 data points...
[2021-03-15 01:46:35] Test accuracy: 0.9972
[2021-03-15 01:46:35] Score for 300 labels: 0.9972
[2021-03-15 01:46:35] Running: experiment 1
[2021-03-15 01:46:35] Showing labelled: True (50/5000 visible)
[2021-03-15 01:46:35] mlp: initialized 1482 parameters.
[2021-03-15 01:46:35] Creating trainer with model on device: cuda
[2021-03-15 01:46:35] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:46:35] Training loss: 1.3196
[2021-03-15 01:46:35] Testing on 5000 data points...
[2021-03-15 01:46:35] Test accuracy: 0.5000
[2021-03-15 01:46:35] Score for 50 labels: 0.5000
[2021-03-15 01:46:35] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:46:35] Found 4950 unlabelled features.
[2021-03-15 01:46:35] Showing labelled: True (100/5000 visible)
[2021-03-15 01:46:35] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:46:36] Training loss: 0.5102
[2021-03-15 01:46:36] Testing on 5000 data points...
[2021-03-15 01:46:36] Test accuracy: 0.7276
[2021-03-15 01:46:36] Score for 100 labels: 0.7276
[2021-03-15 01:46:36] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:46:36] Found 4900 unlabelled features.
[2021-03-15 01:46:36] Showing labelled: True (150/5000 visible)
[2021-03-15 01:46:36] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:46:37] Training loss: 0.3121
[2021-03-15 01:46:37] Testing on 5000 data points...
[2021-03-15 01:46:37] Test accuracy: 0.8936
[2021-03-15 01:46:37] Score for 150 labels: 0.8936
[2021-03-15 01:46:37] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:46:37] Found 4850 unlabelled features.
[2021-03-15 01:46:37] Showing labelled: True (200/5000 visible)
[2021-03-15 01:46:37] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:46:39] Training loss: 0.1628
[2021-03-15 01:46:39] Testing on 5000 data points...
[2021-03-15 01:46:39] Test accuracy: 0.8986
[2021-03-15 01:46:39] Score for 200 labels: 0.8986
[2021-03-15 01:46:39] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:46:39] Found 4800 unlabelled features.
[2021-03-15 01:46:39] Showing labelled: True (250/5000 visible)
[2021-03-15 01:46:39] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:46:40] Training loss: 0.1183
[2021-03-15 01:46:40] Testing on 5000 data points...
[2021-03-15 01:46:40] Test accuracy: 0.9504
[2021-03-15 01:46:40] Score for 250 labels: 0.9504
[2021-03-15 01:46:40] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:46:40] Found 4750 unlabelled features.
[2021-03-15 01:46:40] Showing labelled: True (300/5000 visible)
[2021-03-15 01:46:40] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:46:42] Training loss: 0.0745
[2021-03-15 01:46:42] Testing on 5000 data points...
[2021-03-15 01:46:42] Test accuracy: 0.9928
[2021-03-15 01:46:42] Score for 300 labels: 0.9928
[2021-03-15 01:46:42] Running: experiment 2
[2021-03-15 01:46:42] Showing labelled: True (50/5000 visible)
[2021-03-15 01:46:42] mlp: initialized 1482 parameters.
[2021-03-15 01:46:42] Creating trainer with model on device: cuda
[2021-03-15 01:46:42] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:46:43] Training loss: 1.3525
[2021-03-15 01:46:43] Testing on 5000 data points...
[2021-03-15 01:46:43] Test accuracy: 0.3852
[2021-03-15 01:46:43] Score for 50 labels: 0.3852
[2021-03-15 01:46:43] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:46:43] Found 4950 unlabelled features.
[2021-03-15 01:46:43] Showing labelled: True (100/5000 visible)
[2021-03-15 01:46:43] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:46:43] Training loss: 0.4012
[2021-03-15 01:46:43] Testing on 5000 data points...
[2021-03-15 01:46:43] Test accuracy: 0.5820
[2021-03-15 01:46:43] Score for 100 labels: 0.5820
[2021-03-15 01:46:43] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:46:43] Found 4900 unlabelled features.
[2021-03-15 01:46:44] Showing labelled: True (150/5000 visible)
[2021-03-15 01:46:44] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:46:45] Training loss: 0.3082
[2021-03-15 01:46:45] Testing on 5000 data points...
[2021-03-15 01:46:45] Test accuracy: 0.7862
[2021-03-15 01:46:45] Score for 150 labels: 0.7862
[2021-03-15 01:46:45] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:46:45] Found 4850 unlabelled features.
[2021-03-15 01:46:45] Showing labelled: True (200/5000 visible)
[2021-03-15 01:46:45] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:46:46] Training loss: 0.1365
[2021-03-15 01:46:46] Testing on 5000 data points...
[2021-03-15 01:46:46] Test accuracy: 0.7994
[2021-03-15 01:46:46] Score for 200 labels: 0.7994
[2021-03-15 01:46:46] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:46:46] Found 4800 unlabelled features.
[2021-03-15 01:46:46] Showing labelled: True (250/5000 visible)
[2021-03-15 01:46:46] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:46:48] Training loss: 0.1000
[2021-03-15 01:46:48] Testing on 5000 data points...
[2021-03-15 01:46:48] Test accuracy: 0.7992
[2021-03-15 01:46:48] Score for 250 labels: 0.7992
[2021-03-15 01:46:48] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:46:48] Found 4750 unlabelled features.
[2021-03-15 01:46:48] Showing labelled: True (300/5000 visible)
[2021-03-15 01:46:48] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:46:50] Training loss: 0.0942
[2021-03-15 01:46:50] Testing on 5000 data points...
[2021-03-15 01:46:50] Test accuracy: 0.8980
[2021-03-15 01:46:50] Score for 300 labels: 0.8980
[2021-03-15 01:46:50] Running: experiment 3
[2021-03-15 01:46:50] Showing labelled: True (50/5000 visible)
[2021-03-15 01:46:50] mlp: initialized 1482 parameters.
[2021-03-15 01:46:50] Creating trainer with model on device: cuda
[2021-03-15 01:46:50] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:46:50] Training loss: 1.2595
[2021-03-15 01:46:50] Testing on 5000 data points...
[2021-03-15 01:46:50] Test accuracy: 0.4980
[2021-03-15 01:46:50] Score for 50 labels: 0.4980
[2021-03-15 01:46:50] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:46:50] Found 4950 unlabelled features.
[2021-03-15 01:46:50] Showing labelled: True (100/5000 visible)
[2021-03-15 01:46:50] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:46:51] Training loss: 0.4596
[2021-03-15 01:46:51] Testing on 5000 data points...
[2021-03-15 01:46:51] Test accuracy: 0.7426
[2021-03-15 01:46:51] Score for 100 labels: 0.7426
[2021-03-15 01:46:51] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:46:51] Found 4900 unlabelled features.
[2021-03-15 01:46:51] Showing labelled: True (150/5000 visible)
[2021-03-15 01:46:51] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:46:52] Training loss: 0.2938
[2021-03-15 01:46:52] Testing on 5000 data points...
[2021-03-15 01:46:52] Test accuracy: 0.8952
[2021-03-15 01:46:52] Score for 150 labels: 0.8952
[2021-03-15 01:46:52] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:46:52] Found 4850 unlabelled features.
[2021-03-15 01:46:52] Showing labelled: True (200/5000 visible)
[2021-03-15 01:46:52] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:46:53] Training loss: 0.1091
[2021-03-15 01:46:53] Testing on 5000 data points...
[2021-03-15 01:46:53] Test accuracy: 0.8984
[2021-03-15 01:46:53] Score for 200 labels: 0.8984
[2021-03-15 01:46:53] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:46:53] Found 4800 unlabelled features.
[2021-03-15 01:46:54] Showing labelled: True (250/5000 visible)
[2021-03-15 01:46:54] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:46:55] Training loss: 0.1110
[2021-03-15 01:46:55] Testing on 5000 data points...
[2021-03-15 01:46:55] Test accuracy: 0.9972
[2021-03-15 01:46:55] Score for 250 labels: 0.9972
[2021-03-15 01:46:55] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:46:55] Found 4750 unlabelled features.
[2021-03-15 01:46:55] Showing labelled: True (300/5000 visible)
[2021-03-15 01:46:55] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:46:57] Training loss: 0.0782
[2021-03-15 01:46:57] Testing on 5000 data points...
[2021-03-15 01:46:57] Test accuracy: 0.9978
[2021-03-15 01:46:57] Score for 300 labels: 0.9978
[2021-03-15 01:46:57] Running: experiment 4
[2021-03-15 01:46:57] Showing labelled: True (50/5000 visible)
[2021-03-15 01:46:57] mlp: initialized 1482 parameters.
[2021-03-15 01:46:57] Creating trainer with model on device: cuda
[2021-03-15 01:46:57] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:46:57] Training loss: 1.2853
[2021-03-15 01:46:57] Testing on 5000 data points...
[2021-03-15 01:46:57] Test accuracy: 0.3094
[2021-03-15 01:46:57] Score for 50 labels: 0.3094
[2021-03-15 01:46:57] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:46:57] Found 4950 unlabelled features.
[2021-03-15 01:46:58] Training GMM (k=50) on labelled features...
[2021-03-15 01:46:58] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:46:58] Scoring random batches...
[2021-03-15 01:46:58] Showing labelled: True (100/5000 visible)
[2021-03-15 01:46:58] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:46:59] Training loss: 0.5927
[2021-03-15 01:46:59] Testing on 5000 data points...
[2021-03-15 01:46:59] Test accuracy: 0.7738
[2021-03-15 01:46:59] Score for 100 labels: 0.7738
[2021-03-15 01:46:59] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:46:59] Found 4900 unlabelled features.
[2021-03-15 01:46:59] Training GMM (k=50) on labelled features...
[2021-03-15 01:46:59] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:46:59] Scoring random batches...
[2021-03-15 01:46:59] Showing labelled: True (150/5000 visible)
[2021-03-15 01:46:59] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:47:00] Training loss: 0.2819
[2021-03-15 01:47:00] Testing on 5000 data points...
[2021-03-15 01:47:00] Test accuracy: 0.7980
[2021-03-15 01:47:00] Score for 150 labels: 0.7980
[2021-03-15 01:47:00] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:47:00] Found 4850 unlabelled features.
[2021-03-15 01:47:00] Training GMM (k=50) on labelled features...
[2021-03-15 01:47:00] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:47:01] Scoring random batches...
[2021-03-15 01:47:01] Showing labelled: True (200/5000 visible)
[2021-03-15 01:47:01] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:47:02] Training loss: 0.1455
[2021-03-15 01:47:02] Testing on 5000 data points...
[2021-03-15 01:47:02] Test accuracy: 0.7992
[2021-03-15 01:47:02] Score for 200 labels: 0.7992
[2021-03-15 01:47:02] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:47:02] Found 4800 unlabelled features.
[2021-03-15 01:47:02] Training GMM (k=50) on labelled features...
[2021-03-15 01:47:03] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:47:03] Scoring random batches...
[2021-03-15 01:47:03] Showing labelled: True (250/5000 visible)
[2021-03-15 01:47:03] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:47:04] Training loss: 0.1136
[2021-03-15 01:47:04] Testing on 5000 data points...
[2021-03-15 01:47:05] Test accuracy: 0.8342
[2021-03-15 01:47:05] Score for 250 labels: 0.8342
[2021-03-15 01:47:05] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:47:05] Found 4750 unlabelled features.
[2021-03-15 01:47:05] Training GMM (k=50) on labelled features...
[2021-03-15 01:47:05] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:47:05] Scoring random batches...
[2021-03-15 01:47:05] Showing labelled: True (300/5000 visible)
[2021-03-15 01:47:05] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:47:07] Training loss: 0.0733
[2021-03-15 01:47:07] Testing on 5000 data points...
[2021-03-15 01:47:07] Test accuracy: 0.9816
[2021-03-15 01:47:07] Score for 300 labels: 0.9816
[2021-03-15 01:47:07] Experiment repeat 4/5
[2021-03-15 01:47:07] Randomly labelled 50/5000
[2021-03-15 01:47:07] Showing labelled: True (50/5000 visible)
[2021-03-15 01:47:07] Running: experiment 0
[2021-03-15 01:47:07] Showing labelled: True (50/5000 visible)
[2021-03-15 01:47:07] mlp: initialized 1482 parameters.
[2021-03-15 01:47:07] Creating trainer with model on device: cuda
[2021-03-15 01:47:07] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:47:07] Training loss: 1.1301
[2021-03-15 01:47:07] Testing on 5000 data points...
[2021-03-15 01:47:08] Test accuracy: 0.3866
[2021-03-15 01:47:08] Score for 50 labels: 0.3866
[2021-03-15 01:47:08] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:47:08] Found 4950 unlabelled features.
[2021-03-15 01:47:08] Searching core-set greedily...
[2021-03-15 01:47:12] Showing labelled: True (100/5000 visible)
[2021-03-15 01:47:12] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:47:13] Training loss: 0.7899
[2021-03-15 01:47:13] Testing on 5000 data points...
[2021-03-15 01:47:13] Test accuracy: 0.8996
[2021-03-15 01:47:13] Score for 100 labels: 0.8996
[2021-03-15 01:47:13] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:47:13] Found 4900 unlabelled features.
[2021-03-15 01:47:13] Searching core-set greedily...
[2021-03-15 01:47:19] Showing labelled: True (150/5000 visible)
[2021-03-15 01:47:19] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:47:20] Training loss: 0.3224
[2021-03-15 01:47:20] Testing on 5000 data points...
[2021-03-15 01:47:20] Test accuracy: 0.9004
[2021-03-15 01:47:20] Score for 150 labels: 0.9004
[2021-03-15 01:47:20] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:47:20] Found 4850 unlabelled features.
[2021-03-15 01:47:20] Searching core-set greedily...
[2021-03-15 01:47:29] Showing labelled: True (200/5000 visible)
[2021-03-15 01:47:29] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:47:30] Training loss: 0.1938
[2021-03-15 01:47:30] Testing on 5000 data points...
[2021-03-15 01:47:30] Test accuracy: 0.9956
[2021-03-15 01:47:30] Score for 200 labels: 0.9956
[2021-03-15 01:47:30] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:47:30] Found 4800 unlabelled features.
[2021-03-15 01:47:30] Searching core-set greedily...
[2021-03-15 01:47:41] Showing labelled: True (250/5000 visible)
[2021-03-15 01:47:41] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:47:43] Training loss: 0.1104
[2021-03-15 01:47:43] Testing on 5000 data points...
[2021-03-15 01:47:43] Test accuracy: 0.9970
[2021-03-15 01:47:43] Score for 250 labels: 0.9970
[2021-03-15 01:47:43] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:47:43] Found 4750 unlabelled features.
[2021-03-15 01:47:43] Searching core-set greedily...
[2021-03-15 01:47:54] Showing labelled: True (300/5000 visible)
[2021-03-15 01:47:54] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:47:56] Training loss: 0.0642
[2021-03-15 01:47:56] Testing on 5000 data points...
[2021-03-15 01:47:56] Test accuracy: 0.9986
[2021-03-15 01:47:56] Score for 300 labels: 0.9986
[2021-03-15 01:47:56] Running: experiment 1
[2021-03-15 01:47:56] Showing labelled: True (50/5000 visible)
[2021-03-15 01:47:56] mlp: initialized 1482 parameters.
[2021-03-15 01:47:56] Creating trainer with model on device: cuda
[2021-03-15 01:47:56] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:47:56] Training loss: 1.0007
[2021-03-15 01:47:56] Testing on 5000 data points...
[2021-03-15 01:47:57] Test accuracy: 0.4476
[2021-03-15 01:47:57] Score for 50 labels: 0.4476
[2021-03-15 01:47:57] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:47:57] Found 4950 unlabelled features.
[2021-03-15 01:47:57] Showing labelled: True (100/5000 visible)
[2021-03-15 01:47:57] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:47:57] Training loss: 0.6244
[2021-03-15 01:47:57] Testing on 5000 data points...
[2021-03-15 01:47:57] Test accuracy: 0.6966
[2021-03-15 01:47:57] Score for 100 labels: 0.6966
[2021-03-15 01:47:57] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:47:57] Found 4900 unlabelled features.
[2021-03-15 01:47:57] Showing labelled: True (150/5000 visible)
[2021-03-15 01:47:57] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:47:58] Training loss: 0.2656
[2021-03-15 01:47:58] Testing on 5000 data points...
[2021-03-15 01:47:58] Test accuracy: 0.8002
[2021-03-15 01:47:58] Score for 150 labels: 0.8002
[2021-03-15 01:47:58] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:47:58] Found 4850 unlabelled features.
[2021-03-15 01:47:58] Showing labelled: True (200/5000 visible)
[2021-03-15 01:47:58] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:48:00] Training loss: 0.2198
[2021-03-15 01:48:00] Testing on 5000 data points...
[2021-03-15 01:48:00] Test accuracy: 0.9282
[2021-03-15 01:48:00] Score for 200 labels: 0.9282
[2021-03-15 01:48:00] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:48:00] Found 4800 unlabelled features.
[2021-03-15 01:48:00] Showing labelled: True (250/5000 visible)
[2021-03-15 01:48:00] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:48:01] Training loss: 0.0976
[2021-03-15 01:48:01] Testing on 5000 data points...
[2021-03-15 01:48:01] Test accuracy: 0.9852
[2021-03-15 01:48:01] Score for 250 labels: 0.9852
[2021-03-15 01:48:01] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:48:01] Found 4750 unlabelled features.
[2021-03-15 01:48:01] Showing labelled: True (300/5000 visible)
[2021-03-15 01:48:01] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:48:03] Training loss: 0.0530
[2021-03-15 01:48:03] Testing on 5000 data points...
[2021-03-15 01:48:03] Test accuracy: 0.9938
[2021-03-15 01:48:03] Score for 300 labels: 0.9938
[2021-03-15 01:48:03] Running: experiment 2
[2021-03-15 01:48:03] Showing labelled: True (50/5000 visible)
[2021-03-15 01:48:03] mlp: initialized 1482 parameters.
[2021-03-15 01:48:03] Creating trainer with model on device: cuda
[2021-03-15 01:48:03] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:48:04] Training loss: 1.0568
[2021-03-15 01:48:04] Testing on 5000 data points...
[2021-03-15 01:48:04] Test accuracy: 0.3484
[2021-03-15 01:48:04] Score for 50 labels: 0.3484
[2021-03-15 01:48:04] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:48:04] Found 4950 unlabelled features.
[2021-03-15 01:48:04] Showing labelled: True (100/5000 visible)
[2021-03-15 01:48:04] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:48:05] Training loss: 0.4176
[2021-03-15 01:48:05] Testing on 5000 data points...
[2021-03-15 01:48:05] Test accuracy: 0.5556
[2021-03-15 01:48:05] Score for 100 labels: 0.5556
[2021-03-15 01:48:05] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:48:05] Found 4900 unlabelled features.
[2021-03-15 01:48:05] Showing labelled: True (150/5000 visible)
[2021-03-15 01:48:05] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:48:06] Training loss: 0.1491
[2021-03-15 01:48:06] Testing on 5000 data points...
[2021-03-15 01:48:06] Test accuracy: 0.6508
[2021-03-15 01:48:06] Score for 150 labels: 0.6508
[2021-03-15 01:48:06] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:48:06] Found 4850 unlabelled features.
[2021-03-15 01:48:06] Showing labelled: True (200/5000 visible)
[2021-03-15 01:48:06] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:48:07] Training loss: 0.1369
[2021-03-15 01:48:07] Testing on 5000 data points...
[2021-03-15 01:48:07] Test accuracy: 0.7840
[2021-03-15 01:48:07] Score for 200 labels: 0.7840
[2021-03-15 01:48:07] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:48:07] Found 4800 unlabelled features.
[2021-03-15 01:48:07] Showing labelled: True (250/5000 visible)
[2021-03-15 01:48:07] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:48:09] Training loss: 0.1556
[2021-03-15 01:48:09] Testing on 5000 data points...
[2021-03-15 01:48:09] Test accuracy: 0.8894
[2021-03-15 01:48:09] Score for 250 labels: 0.8894
[2021-03-15 01:48:09] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:48:09] Found 4750 unlabelled features.
[2021-03-15 01:48:09] Showing labelled: True (300/5000 visible)
[2021-03-15 01:48:09] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:48:11] Training loss: 0.1719
[2021-03-15 01:48:11] Testing on 5000 data points...
[2021-03-15 01:48:11] Test accuracy: 0.9882
[2021-03-15 01:48:11] Score for 300 labels: 0.9882
[2021-03-15 01:48:11] Running: experiment 3
[2021-03-15 01:48:11] Showing labelled: True (50/5000 visible)
[2021-03-15 01:48:11] mlp: initialized 1482 parameters.
[2021-03-15 01:48:11] Creating trainer with model on device: cuda
[2021-03-15 01:48:11] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:48:11] Training loss: 1.0097
[2021-03-15 01:48:11] Testing on 5000 data points...
[2021-03-15 01:48:11] Test accuracy: 0.4982
[2021-03-15 01:48:11] Score for 50 labels: 0.4982
[2021-03-15 01:48:11] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:48:11] Found 4950 unlabelled features.
[2021-03-15 01:48:11] Showing labelled: True (100/5000 visible)
[2021-03-15 01:48:11] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:48:12] Training loss: 0.4041
[2021-03-15 01:48:12] Testing on 5000 data points...
[2021-03-15 01:48:12] Test accuracy: 0.4970
[2021-03-15 01:48:12] Score for 100 labels: 0.4970
[2021-03-15 01:48:12] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:48:12] Found 4900 unlabelled features.
[2021-03-15 01:48:12] Showing labelled: True (150/5000 visible)
[2021-03-15 01:48:12] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:48:13] Training loss: 0.1893
[2021-03-15 01:48:13] Testing on 5000 data points...
[2021-03-15 01:48:13] Test accuracy: 0.5998
[2021-03-15 01:48:13] Score for 150 labels: 0.5998
[2021-03-15 01:48:13] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:48:13] Found 4850 unlabelled features.
[2021-03-15 01:48:13] Showing labelled: True (200/5000 visible)
[2021-03-15 01:48:13] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:48:14] Training loss: 0.1231
[2021-03-15 01:48:14] Testing on 5000 data points...
[2021-03-15 01:48:14] Test accuracy: 0.6208
[2021-03-15 01:48:15] Score for 200 labels: 0.6208
[2021-03-15 01:48:15] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:48:15] Found 4800 unlabelled features.
[2021-03-15 01:48:15] Showing labelled: True (250/5000 visible)
[2021-03-15 01:48:15] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:48:16] Training loss: 0.0806
[2021-03-15 01:48:16] Testing on 5000 data points...
[2021-03-15 01:48:16] Test accuracy: 0.6996
[2021-03-15 01:48:16] Score for 250 labels: 0.6996
[2021-03-15 01:48:16] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:48:16] Found 4750 unlabelled features.
[2021-03-15 01:48:16] Showing labelled: True (300/5000 visible)
[2021-03-15 01:48:16] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:48:18] Training loss: 0.1397
[2021-03-15 01:48:18] Testing on 5000 data points...
[2021-03-15 01:48:18] Test accuracy: 0.8654
[2021-03-15 01:48:18] Score for 300 labels: 0.8654
[2021-03-15 01:48:18] Running: experiment 4
[2021-03-15 01:48:18] Showing labelled: True (50/5000 visible)
[2021-03-15 01:48:18] mlp: initialized 1482 parameters.
[2021-03-15 01:48:18] Creating trainer with model on device: cuda
[2021-03-15 01:48:18] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:48:19] Training loss: 1.0427
[2021-03-15 01:48:19] Testing on 5000 data points...
[2021-03-15 01:48:19] Test accuracy: 0.3000
[2021-03-15 01:48:19] Score for 50 labels: 0.3000
[2021-03-15 01:48:19] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:48:19] Found 4950 unlabelled features.
[2021-03-15 01:48:19] Training GMM (k=50) on labelled features...
[2021-03-15 01:48:19] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:48:19] Scoring random batches...
[2021-03-15 01:48:19] Showing labelled: True (100/5000 visible)
[2021-03-15 01:48:19] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:48:20] Training loss: 0.6322
[2021-03-15 01:48:20] Testing on 5000 data points...
[2021-03-15 01:48:20] Test accuracy: 0.5974
[2021-03-15 01:48:20] Score for 100 labels: 0.5974
[2021-03-15 01:48:20] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:48:20] Found 4900 unlabelled features.
[2021-03-15 01:48:20] Training GMM (k=50) on labelled features...
[2021-03-15 01:48:20] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:48:20] Scoring random batches...
[2021-03-15 01:48:20] Showing labelled: True (150/5000 visible)
[2021-03-15 01:48:20] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:48:21] Training loss: 0.3428
[2021-03-15 01:48:21] Testing on 5000 data points...
[2021-03-15 01:48:21] Test accuracy: 0.7422
[2021-03-15 01:48:21] Score for 150 labels: 0.7422
[2021-03-15 01:48:21] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:48:21] Found 4850 unlabelled features.
[2021-03-15 01:48:21] Training GMM (k=50) on labelled features...
[2021-03-15 01:48:22] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:48:22] Scoring random batches...
[2021-03-15 01:48:22] Showing labelled: True (200/5000 visible)
[2021-03-15 01:48:22] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:48:23] Training loss: 0.1684
[2021-03-15 01:48:23] Testing on 5000 data points...
[2021-03-15 01:48:23] Test accuracy: 0.8130
[2021-03-15 01:48:23] Score for 200 labels: 0.8130
[2021-03-15 01:48:23] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:48:23] Found 4800 unlabelled features.
[2021-03-15 01:48:23] Training GMM (k=50) on labelled features...
[2021-03-15 01:48:24] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:48:24] Scoring random batches...
[2021-03-15 01:48:24] Showing labelled: True (250/5000 visible)
[2021-03-15 01:48:24] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:48:26] Training loss: 0.0952
[2021-03-15 01:48:26] Testing on 5000 data points...
[2021-03-15 01:48:26] Test accuracy: 0.8930
[2021-03-15 01:48:26] Score for 250 labels: 0.8930
[2021-03-15 01:48:26] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:48:26] Found 4750 unlabelled features.
[2021-03-15 01:48:26] Training GMM (k=50) on labelled features...
[2021-03-15 01:48:26] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:48:26] Scoring random batches...
[2021-03-15 01:48:26] Showing labelled: True (300/5000 visible)
[2021-03-15 01:48:26] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:48:28] Training loss: 0.0538
[2021-03-15 01:48:28] Testing on 5000 data points...
[2021-03-15 01:48:28] Test accuracy: 0.9936
[2021-03-15 01:48:28] Score for 300 labels: 0.9936
[2021-03-15 01:48:28] Experiment repeat 5/5
[2021-03-15 01:48:28] Randomly labelled 50/5000
[2021-03-15 01:48:28] Showing labelled: True (50/5000 visible)
[2021-03-15 01:48:28] Running: experiment 0
[2021-03-15 01:48:28] Showing labelled: True (50/5000 visible)
[2021-03-15 01:48:28] mlp: initialized 1482 parameters.
[2021-03-15 01:48:28] Creating trainer with model on device: cuda
[2021-03-15 01:48:28] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:48:29] Training loss: 1.1627
[2021-03-15 01:48:29] Testing on 5000 data points...
[2021-03-15 01:48:29] Test accuracy: 0.3848
[2021-03-15 01:48:29] Score for 50 labels: 0.3848
[2021-03-15 01:48:29] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:48:29] Found 4950 unlabelled features.
[2021-03-15 01:48:29] Searching core-set greedily...
[2021-03-15 01:48:35] Showing labelled: True (100/5000 visible)
[2021-03-15 01:48:35] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:48:35] Training loss: 0.7282
[2021-03-15 01:48:35] Testing on 5000 data points...
[2021-03-15 01:48:36] Test accuracy: 0.8000
[2021-03-15 01:48:36] Score for 100 labels: 0.8000
[2021-03-15 01:48:36] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:48:36] Found 4900 unlabelled features.
[2021-03-15 01:48:36] Searching core-set greedily...
[2021-03-15 01:48:41] Showing labelled: True (150/5000 visible)
[2021-03-15 01:48:41] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:48:42] Training loss: 0.3670
[2021-03-15 01:48:42] Testing on 5000 data points...
[2021-03-15 01:48:42] Test accuracy: 0.9386
[2021-03-15 01:48:42] Score for 150 labels: 0.9386
[2021-03-15 01:48:42] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:48:42] Found 4850 unlabelled features.
[2021-03-15 01:48:42] Searching core-set greedily...
[2021-03-15 01:48:52] Showing labelled: True (200/5000 visible)
[2021-03-15 01:48:52] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:48:53] Training loss: 0.2157
[2021-03-15 01:48:53] Testing on 5000 data points...
[2021-03-15 01:48:54] Test accuracy: 0.9964
[2021-03-15 01:48:54] Score for 200 labels: 0.9964
[2021-03-15 01:48:54] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:48:54] Found 4800 unlabelled features.
[2021-03-15 01:48:54] Searching core-set greedily...
[2021-03-15 01:49:03] Showing labelled: True (250/5000 visible)
[2021-03-15 01:49:03] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:49:05] Training loss: 0.1142
[2021-03-15 01:49:05] Testing on 5000 data points...
[2021-03-15 01:49:05] Test accuracy: 0.9982
[2021-03-15 01:49:05] Score for 250 labels: 0.9982
[2021-03-15 01:49:05] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:49:05] Found 4750 unlabelled features.
[2021-03-15 01:49:05] Searching core-set greedily...
[2021-03-15 01:49:15] Showing labelled: True (300/5000 visible)
[2021-03-15 01:49:15] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:49:17] Training loss: 0.0741
[2021-03-15 01:49:17] Testing on 5000 data points...
[2021-03-15 01:49:17] Test accuracy: 0.9962
[2021-03-15 01:49:17] Score for 300 labels: 0.9962
[2021-03-15 01:49:17] Running: experiment 1
[2021-03-15 01:49:17] Showing labelled: True (50/5000 visible)
[2021-03-15 01:49:17] mlp: initialized 1482 parameters.
[2021-03-15 01:49:17] Creating trainer with model on device: cuda
[2021-03-15 01:49:17] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:49:17] Training loss: 1.1842
[2021-03-15 01:49:17] Testing on 5000 data points...
[2021-03-15 01:49:17] Test accuracy: 0.4064
[2021-03-15 01:49:17] Score for 50 labels: 0.4064
[2021-03-15 01:49:17] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:49:17] Found 4950 unlabelled features.
[2021-03-15 01:49:17] Showing labelled: True (100/5000 visible)
[2021-03-15 01:49:17] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:49:18] Training loss: 0.6262
[2021-03-15 01:49:18] Testing on 5000 data points...
[2021-03-15 01:49:18] Test accuracy: 0.6954
[2021-03-15 01:49:18] Score for 100 labels: 0.6954
[2021-03-15 01:49:18] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:49:18] Found 4900 unlabelled features.
[2021-03-15 01:49:18] Showing labelled: True (150/5000 visible)
[2021-03-15 01:49:18] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:49:19] Training loss: 0.3275
[2021-03-15 01:49:19] Testing on 5000 data points...
[2021-03-15 01:49:19] Test accuracy: 0.7766
[2021-03-15 01:49:19] Score for 150 labels: 0.7766
[2021-03-15 01:49:19] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:49:19] Found 4850 unlabelled features.
[2021-03-15 01:49:19] Showing labelled: True (200/5000 visible)
[2021-03-15 01:49:19] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:49:20] Training loss: 0.2005
[2021-03-15 01:49:20] Testing on 5000 data points...
[2021-03-15 01:49:20] Test accuracy: 0.9494
[2021-03-15 01:49:20] Score for 200 labels: 0.9494
[2021-03-15 01:49:20] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:49:20] Found 4800 unlabelled features.
[2021-03-15 01:49:20] Showing labelled: True (250/5000 visible)
[2021-03-15 01:49:20] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:49:22] Training loss: 0.0981
[2021-03-15 01:49:22] Testing on 5000 data points...
[2021-03-15 01:49:22] Test accuracy: 0.9734
[2021-03-15 01:49:22] Score for 250 labels: 0.9734
[2021-03-15 01:49:22] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:49:22] Found 4750 unlabelled features.
[2021-03-15 01:49:22] Showing labelled: True (300/5000 visible)
[2021-03-15 01:49:22] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:49:24] Training loss: 0.0622
[2021-03-15 01:49:24] Testing on 5000 data points...
[2021-03-15 01:49:24] Test accuracy: 0.9882
[2021-03-15 01:49:24] Score for 300 labels: 0.9882
[2021-03-15 01:49:24] Running: experiment 2
[2021-03-15 01:49:24] Showing labelled: True (50/5000 visible)
[2021-03-15 01:49:24] mlp: initialized 1482 parameters.
[2021-03-15 01:49:24] Creating trainer with model on device: cuda
[2021-03-15 01:49:24] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:49:24] Training loss: 1.0272
[2021-03-15 01:49:24] Testing on 5000 data points...
[2021-03-15 01:49:24] Test accuracy: 0.4992
[2021-03-15 01:49:24] Score for 50 labels: 0.4992
[2021-03-15 01:49:24] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:49:24] Found 4950 unlabelled features.
[2021-03-15 01:49:24] Showing labelled: True (100/5000 visible)
[2021-03-15 01:49:24] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:49:25] Training loss: 0.5116
[2021-03-15 01:49:25] Testing on 5000 data points...
[2021-03-15 01:49:25] Test accuracy: 0.5000
[2021-03-15 01:49:25] Score for 100 labels: 0.5000
[2021-03-15 01:49:25] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:49:25] Found 4900 unlabelled features.
[2021-03-15 01:49:25] Showing labelled: True (150/5000 visible)
[2021-03-15 01:49:25] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:49:26] Training loss: 0.1997
[2021-03-15 01:49:26] Testing on 5000 data points...
[2021-03-15 01:49:26] Test accuracy: 0.5998
[2021-03-15 01:49:26] Score for 150 labels: 0.5998
[2021-03-15 01:49:26] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:49:26] Found 4850 unlabelled features.
[2021-03-15 01:49:27] Showing labelled: True (200/5000 visible)
[2021-03-15 01:49:27] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:49:28] Training loss: 0.1265
[2021-03-15 01:49:28] Testing on 5000 data points...
[2021-03-15 01:49:28] Test accuracy: 0.7040
[2021-03-15 01:49:28] Score for 200 labels: 0.7040
[2021-03-15 01:49:28] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:49:28] Found 4800 unlabelled features.
[2021-03-15 01:49:28] Showing labelled: True (250/5000 visible)
[2021-03-15 01:49:28] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:49:29] Training loss: 0.0800
[2021-03-15 01:49:29] Testing on 5000 data points...
[2021-03-15 01:49:29] Test accuracy: 0.7984
[2021-03-15 01:49:29] Score for 250 labels: 0.7984
[2021-03-15 01:49:29] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:49:29] Found 4750 unlabelled features.
[2021-03-15 01:49:30] Showing labelled: True (300/5000 visible)
[2021-03-15 01:49:30] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:49:31] Training loss: 0.0708
[2021-03-15 01:49:31] Testing on 5000 data points...
[2021-03-15 01:49:31] Test accuracy: 0.8970
[2021-03-15 01:49:31] Score for 300 labels: 0.8970
[2021-03-15 01:49:31] Running: experiment 3
[2021-03-15 01:49:31] Showing labelled: True (50/5000 visible)
[2021-03-15 01:49:31] mlp: initialized 1482 parameters.
[2021-03-15 01:49:31] Creating trainer with model on device: cuda
[2021-03-15 01:49:31] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:49:32] Training loss: 1.1655
[2021-03-15 01:49:32] Testing on 5000 data points...
[2021-03-15 01:49:32] Test accuracy: 0.4730
[2021-03-15 01:49:32] Score for 50 labels: 0.4730
[2021-03-15 01:49:32] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:49:32] Found 4950 unlabelled features.
[2021-03-15 01:49:32] Showing labelled: True (100/5000 visible)
[2021-03-15 01:49:32] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:49:33] Training loss: 0.5292
[2021-03-15 01:49:33] Testing on 5000 data points...
[2021-03-15 01:49:33] Test accuracy: 0.4992
[2021-03-15 01:49:33] Score for 100 labels: 0.4992
[2021-03-15 01:49:33] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:49:33] Found 4900 unlabelled features.
[2021-03-15 01:49:33] Showing labelled: True (150/5000 visible)
[2021-03-15 01:49:33] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:49:34] Training loss: 0.1660
[2021-03-15 01:49:34] Testing on 5000 data points...
[2021-03-15 01:49:34] Test accuracy: 0.5736
[2021-03-15 01:49:34] Score for 150 labels: 0.5736
[2021-03-15 01:49:34] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:49:34] Found 4850 unlabelled features.
[2021-03-15 01:49:34] Showing labelled: True (200/5000 visible)
[2021-03-15 01:49:34] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:49:35] Training loss: 0.1447
[2021-03-15 01:49:35] Testing on 5000 data points...
[2021-03-15 01:49:35] Test accuracy: 0.6474
[2021-03-15 01:49:35] Score for 200 labels: 0.6474
[2021-03-15 01:49:35] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:49:35] Found 4800 unlabelled features.
[2021-03-15 01:49:35] Showing labelled: True (250/5000 visible)
[2021-03-15 01:49:35] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:49:37] Training loss: 0.0989
[2021-03-15 01:49:37] Testing on 5000 data points...
[2021-03-15 01:49:37] Test accuracy: 0.6664
[2021-03-15 01:49:37] Score for 250 labels: 0.6664
[2021-03-15 01:49:37] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:49:37] Found 4750 unlabelled features.
[2021-03-15 01:49:37] Showing labelled: True (300/5000 visible)
[2021-03-15 01:49:37] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:49:39] Training loss: 0.0665
[2021-03-15 01:49:39] Testing on 5000 data points...
[2021-03-15 01:49:39] Test accuracy: 0.6986
[2021-03-15 01:49:39] Score for 300 labels: 0.6986
[2021-03-15 01:49:39] Running: experiment 4
[2021-03-15 01:49:39] Showing labelled: True (50/5000 visible)
[2021-03-15 01:49:39] mlp: initialized 1482 parameters.
[2021-03-15 01:49:39] Creating trainer with model on device: cuda
[2021-03-15 01:49:39] Training mlp across 50 data points in blobsthin...
[2021-03-15 01:49:39] Training loss: 1.1383
[2021-03-15 01:49:39] Testing on 5000 data points...
[2021-03-15 01:49:39] Test accuracy: 0.4934
[2021-03-15 01:49:39] Score for 50 labels: 0.4934
[2021-03-15 01:49:39] Showing labelled: False (4950/5000 visible)
[2021-03-15 01:49:39] Found 4950 unlabelled features.
[2021-03-15 01:49:39] Training GMM (k=50) on labelled features...
[2021-03-15 01:49:39] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:49:40] Scoring random batches...
[2021-03-15 01:49:40] Showing labelled: True (100/5000 visible)
[2021-03-15 01:49:40] Training mlp across 100 data points in blobsthin...
[2021-03-15 01:49:40] Training loss: 0.6468
[2021-03-15 01:49:40] Testing on 5000 data points...
[2021-03-15 01:49:40] Test accuracy: 0.5974
[2021-03-15 01:49:40] Score for 100 labels: 0.5974
[2021-03-15 01:49:40] Showing labelled: False (4900/5000 visible)
[2021-03-15 01:49:40] Found 4900 unlabelled features.
[2021-03-15 01:49:40] Training GMM (k=50) on labelled features...
[2021-03-15 01:49:41] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:49:41] Scoring random batches...
[2021-03-15 01:49:41] Showing labelled: True (150/5000 visible)
[2021-03-15 01:49:41] Training mlp across 150 data points in blobsthin...
[2021-03-15 01:49:42] Training loss: 0.3288
[2021-03-15 01:49:42] Testing on 5000 data points...
[2021-03-15 01:49:42] Test accuracy: 0.8866
[2021-03-15 01:49:42] Score for 150 labels: 0.8866
[2021-03-15 01:49:42] Showing labelled: False (4850/5000 visible)
[2021-03-15 01:49:42] Found 4850 unlabelled features.
[2021-03-15 01:49:42] Training GMM (k=50) on labelled features...
[2021-03-15 01:49:42] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:49:43] Scoring random batches...
[2021-03-15 01:49:43] Showing labelled: True (200/5000 visible)
[2021-03-15 01:49:43] Training mlp across 200 data points in blobsthin...
[2021-03-15 01:49:44] Training loss: 0.1463
[2021-03-15 01:49:44] Testing on 5000 data points...
[2021-03-15 01:49:44] Test accuracy: 0.9286
[2021-03-15 01:49:44] Score for 200 labels: 0.9286
[2021-03-15 01:49:44] Showing labelled: False (4800/5000 visible)
[2021-03-15 01:49:44] Found 4800 unlabelled features.
[2021-03-15 01:49:44] Training GMM (k=50) on labelled features...
[2021-03-15 01:49:44] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:49:45] Scoring random batches...
[2021-03-15 01:49:45] Showing labelled: True (250/5000 visible)
[2021-03-15 01:49:45] Training mlp across 250 data points in blobsthin...
[2021-03-15 01:49:46] Training loss: 0.0894
[2021-03-15 01:49:46] Testing on 5000 data points...
[2021-03-15 01:49:46] Test accuracy: 0.9770
[2021-03-15 01:49:46] Score for 250 labels: 0.9770
[2021-03-15 01:49:46] Showing labelled: False (4750/5000 visible)
[2021-03-15 01:49:46] Found 4750 unlabelled features.
[2021-03-15 01:49:46] Training GMM (k=50) on labelled features...
[2021-03-15 01:49:47] Training GMM (k=50) on unlabelled features...
[2021-03-15 01:49:47] Scoring random batches...
[2021-03-15 01:49:47] Showing labelled: True (300/5000 visible)
[2021-03-15 01:49:47] Training mlp across 300 data points in blobsthin...
[2021-03-15 01:49:49] Training loss: 0.0511
[2021-03-15 01:49:49] Testing on 5000 data points...
[2021-03-15 01:49:49] Test accuracy: 0.9890
[2021-03-15 01:49:49] Score for 300 labels: 0.9890
[2021-03-15 01:49:49] Updated results: ../results/replicate/results.json
