File size: 9,923 Bytes
58c21d4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 | [2026-09-04 11:52:08] queue out/queue/deeper_mnist.txt started, pid 14662
[2026-09-04 11:52:08] waiting: load 3.69, other jobs: 13135
[2026-09-04 12:02:09] waiting: load 3.19, other jobs: 13135
[2026-09-04 12:12:09] waiting: load 2.35, other jobs: 13135
[2026-09-04 12:22:10] waiting: load 2.34, other jobs:
[2026-09-04 12:23:10] idle after 1860s
[2026-09-04 12:23:10] start [1]: .venv/bin/python scripts/train_paper_mlp.py --steps 300 --hidden-layers 6 --out /private/tmp/claude-501/-Users-brianna-Cells-Developmental/75a5e546-8113-4317-8480-640996d23041/scratchpad/smoke_paper
[12:23:12] 12 checkpoints to 300 steps on 1000 digits, depth 7 (6 hidden layers)
[12:23:12] plain-h6 step 0: train 0.103 test 0.101 mse 1.055e-01 (1s)
[12:23:13] plain-h6 step 100: train 0.995 test 0.892 mse 3.824e-03 (1s)
[12:23:14] plain-h6 step 300: train 1.000 test 0.890 mse 4.179e-05 (2s)
[12:23:14] plain-h6: wrote 12 checkpoints; train 1.000 test 0.890 (3s)
[12:23:14] res-h6 step 0: train 0.142 test 0.141 mse 1.789e-01 (0s)
[12:23:15] res-h6 step 100: train 1.000 test 0.914 mse 3.125e-03 (1s)
[12:23:16] res-h6 step 300: train 1.000 test 0.912 mse 1.236e-04 (1s)
[12:23:16] res-h6: wrote 12 checkpoints; train 1.000 test 0.912 (2s)
[2026-09-04 12:23:17] exit [1]: status 0 after 7s
[2026-09-04 12:23:17] start [2]: .venv/bin/python scripts/train_mnist_dense.py --steps 300 --blocks 6 --out /private/tmp/claude-501/-Users-brianna-Cells-Developmental/75a5e546-8113-4317-8480-640996d23041/scratchpad/smoke_dense
[12:23:18] 12 checkpoints to 300 steps, 6 blocks of width 256, writing to /private/tmp/claude-501/-Users-brianna-Cells-Developmental/75a5e546-8113-4317-8480-640996d23041/scratchpad/smoke_dense
[12:23:19] step 0 test 0.0856 train CE 2.364e+00 (1s)
[12:23:19] step 1 test 0.1091 train CE 2.295e+00 (1s)
[12:23:20] step 2 test 0.2030 train CE 2.178e+00 (2s)
[12:23:20] step 3 test 0.3997 train CE 2.026e+00 (2s)
[12:23:21] step 6 test 0.6331 train CE 1.531e+00 (3s)
[12:23:21] step 10 test 0.7235 train CE 1.023e+00 (3s)
[12:23:21] step 18 test 0.8336 train CE 5.954e-01 (4s)
[12:23:22] step 32 test 0.8769 train CE 4.473e-01 (4s)
[12:23:22] step 56 test 0.8935 train CE 3.861e-01 (5s)
[12:23:23] step 100 test 0.8996 train CE 3.566e-01 (5s)
[12:23:24] step 178 test 0.9119 train CE 3.218e-01 (6s)
[12:23:25] step 300 test 0.9122 train CE 3.148e-01 (7s)
wrote 12 checkpoints to /private/tmp/claude-501/-Users-brianna-Cells-Developmental/75a5e546-8113-4317-8480-640996d23041/scratchpad/smoke_dense
[2026-09-04 12:23:27] exit [2]: status 0 after 10s
[2026-09-04 12:23:27] start [3]: .venv/bin/python scripts/train_paper_mlp.py --hidden-layers 4
[12:23:28] 41 checkpoints to 100200 steps on 1000 digits, depth 5 (4 hidden layers)
[12:23:28] plain-h4 step 0: train 0.107 test 0.096 mse 1.013e-01 (1s)
[12:23:29] plain-h4 step 100: train 0.996 test 0.910 mse 3.504e-03 (1s)
[12:23:36] plain-h4 step 3000: train 1.000 test 0.909 mse 2.350e-06 (8s)
[12:24:37] plain-h4 step 30000: train 1.000 test 0.893 mse 1.819e-06 (70s)
[12:27:15] plain-h4 step 100200: train 1.000 test 0.882 mse 2.092e-09 (227s)
[12:27:16] plain-h4: wrote 41 checkpoints; train 1.000 test 0.882 (229s)
[12:27:16] res-h4 step 0: train 0.105 test 0.097 mse 1.157e-01 (0s)
[12:27:17] res-h4 step 100: train 1.000 test 0.913 mse 3.053e-03 (1s)
[12:27:25] res-h4 step 3000: train 1.000 test 0.914 mse 1.595e-04 (9s)
[12:28:37] res-h4 step 30000: train 1.000 test 0.866 mse 4.458e-06 (81s)
[12:32:30] res-h4 step 100200: train 0.999 test 0.814 mse 1.995e-04 (314s)
[12:32:32] res-h4: wrote 41 checkpoints; train 0.999 test 0.814 (316s)
[2026-09-04 12:32:32] exit [3]: status 0 after 545s
[2026-09-04 12:32:32] waiting: load 2.50, other jobs:
[2026-09-04 12:37:51] queue out/queue/deeper_mnist_rest.txt started, pid 16287
[2026-09-04 12:37:51] start [1]: .venv/bin/python scripts/train_paper_mlp.py --hidden-layers 5
[12:37:53] 41 checkpoints to 100200 steps on 1000 digits, depth 6 (5 hidden layers)
[12:37:54] plain-h5 step 0: train 0.080 test 0.085 mse 1.085e-01 (1s)
[12:37:55] plain-h5 step 100: train 0.993 test 0.900 mse 4.220e-03 (2s)
[12:38:05] plain-h5 step 3000: train 1.000 test 0.901 mse 5.822e-06 (11s)
[12:39:23] plain-h5 step 30000: train 1.000 test 0.896 mse 1.084e-07 (89s)
[12:42:53] plain-h5 step 100200: train 1.000 test 0.896 mse 8.453e-08 (300s)
[12:42:55] plain-h5: wrote 41 checkpoints; train 1.000 test 0.896 (302s)
[12:42:55] res-h5 step 0: train 0.083 test 0.084 mse 1.539e-01 (0s)
[12:42:56] res-h5 step 100: train 0.999 test 0.907 mse 3.371e-03 (1s)
[12:43:07] res-h5 step 3000: train 1.000 test 0.908 mse 6.232e-05 (12s)
[12:44:43] res-h5 step 30000: train 0.999 test 0.861 mse 1.733e-04 (108s)
[12:48:45] res-h5 step 100200: train 0.998 test 0.840 mse 2.984e-04 (350s)
[12:48:47] res-h5: wrote 41 checkpoints; train 0.998 test 0.840 (352s)
[2026-09-04 12:48:48] exit [1]: status 0 after 657s
[2026-09-04 12:48:48] start [2]: .venv/bin/python scripts/train_paper_mlp.py --hidden-layers 6
[12:48:49] 41 checkpoints to 100200 steps on 1000 digits, depth 7 (6 hidden layers)
[12:48:50] plain-h6 step 0: train 0.103 test 0.101 mse 1.055e-01 (1s)
[12:48:51] plain-h6 step 100: train 0.995 test 0.892 mse 3.824e-03 (1s)
[12:49:00] plain-h6 step 3000: train 1.000 test 0.908 mse 3.052e-06 (11s)
[12:50:23] plain-h6 step 30000: train 1.000 test 0.899 mse 9.366e-08 (94s)
[12:53:54] plain-h6 step 100200: train 1.000 test 0.895 mse 2.648e-07 (305s)
[12:53:56] plain-h6: wrote 41 checkpoints; train 1.000 test 0.895 (306s)
[12:53:56] res-h6 step 0: train 0.142 test 0.141 mse 1.789e-01 (0s)
[12:53:57] res-h6 step 100: train 1.000 test 0.914 mse 3.125e-03 (1s)
[12:54:07] res-h6 step 3000: train 1.000 test 0.910 mse 1.573e-05 (12s)
[12:55:40] res-h6 step 30000: train 1.000 test 0.871 mse 3.015e-07 (104s)
[13:00:17] res-h6 step 100200: train 0.998 test 0.838 mse 3.961e-04 (382s)
[13:00:19] res-h6: wrote 41 checkpoints; train 0.998 test 0.838 (384s)
[2026-09-04 13:00:20] exit [2]: status 0 after 692s
[2026-09-04 13:00:20] start [3]: .venv/bin/python scripts/train_mnist_dense.py --blocks 5
[13:00:22] 23 checkpoints to 177828 steps, 5 blocks of width 256, writing to /Users/brianna/Cells_Developmental/out/mnist_dense-b5
[13:00:23] step 0 test 0.0859 train CE 2.387e+00 (1s)
[13:00:23] step 1 test 0.1076 train CE 2.326e+00 (1s)
[13:00:23] step 2 test 0.1797 train CE 2.219e+00 (2s)
[13:00:24] step 3 test 0.3101 train CE 2.075e+00 (2s)
[13:00:24] step 6 test 0.6090 train CE 1.576e+00 (3s)
[13:00:25] step 10 test 0.7255 train CE 1.044e+00 (3s)
[13:00:25] step 18 test 0.8147 train CE 6.254e-01 (4s)
[13:00:25] step 32 test 0.8735 train CE 4.559e-01 (4s)
[13:00:26] step 56 test 0.8935 train CE 3.898e-01 (4s)
[13:00:27] step 100 test 0.9010 train CE 3.592e-01 (5s)
[13:00:27] step 178 test 0.9098 train CE 3.237e-01 (6s)
[13:00:28] step 316 test 0.9052 train CE 3.218e-01 (7s)
[13:00:29] step 562 test 0.9128 train CE 2.958e-01 (8s)
[13:00:31] step 1000 test 0.9267 train CE 2.548e-01 (10s)
[13:00:34] step 1778 test 0.9419 train CE 1.872e-01 (13s)
[13:00:39] step 3162 test 0.9611 train CE 1.056e-01 (18s)
[13:00:48] step 5623 test 0.9704 train CE 4.883e-02 (26s)
[13:01:03] step 10000 test 0.9776 train CE 9.789e-03 (41s)
[13:01:29] step 17783 test 0.9804 train CE 6.703e-04 (68s)
[13:02:17] step 31623 test 0.9798 train CE 2.263e-04 (116s)
[13:03:41] step 56234 test 0.9801 train CE 9.273e-05 (200s)
[13:06:08] step 100000 test 0.9803 train CE 4.118e-05 (346s)
[13:10:28] step 177828 test 0.9803 train CE 1.913e-05 (607s)
wrote 23 checkpoints to /Users/brianna/Cells_Developmental/out/mnist_dense-b5
[2026-09-04 13:10:32] exit [3]: status 0 after 612s
[2026-09-04 13:10:32] start [4]: .venv/bin/python scripts/train_mnist_dense.py --blocks 6
[13:10:33] 23 checkpoints to 177828 steps, 6 blocks of width 256, writing to /Users/brianna/Cells_Developmental/out/mnist_dense-b6
[13:10:34] step 0 test 0.0856 train CE 2.364e+00 (1s)
[13:10:34] step 1 test 0.1091 train CE 2.295e+00 (2s)
[13:10:35] step 2 test 0.2030 train CE 2.178e+00 (2s)
[13:10:35] step 3 test 0.3997 train CE 2.026e+00 (3s)
[13:10:36] step 6 test 0.6331 train CE 1.531e+00 (3s)
[13:10:37] step 10 test 0.7235 train CE 1.023e+00 (4s)
[13:10:38] step 18 test 0.8336 train CE 5.954e-01 (5s)
[13:10:38] step 32 test 0.8769 train CE 4.473e-01 (5s)
[13:10:39] step 56 test 0.8935 train CE 3.861e-01 (6s)
[13:10:40] step 100 test 0.8996 train CE 3.566e-01 (7s)
[13:10:40] step 178 test 0.9119 train CE 3.218e-01 (8s)
[13:10:41] step 316 test 0.9058 train CE 3.207e-01 (9s)
[13:10:43] step 562 test 0.9133 train CE 2.964e-01 (10s)
[13:10:45] step 1000 test 0.9250 train CE 2.562e-01 (12s)
[13:10:49] step 1778 test 0.9416 train CE 1.892e-01 (16s)
[13:10:55] step 3162 test 0.9595 train CE 1.087e-01 (22s)
[13:11:05] step 5623 test 0.9704 train CE 4.865e-02 (32s)
[13:11:21] step 10000 test 0.9792 train CE 7.441e-03 (49s)
[13:11:51] step 17783 test 0.9803 train CE 6.737e-04 (78s)
[13:12:42] step 31623 test 0.9804 train CE 2.270e-04 (130s)
[13:14:12] step 56234 test 0.9810 train CE 9.301e-05 (219s)
[13:16:51] step 100000 test 0.9807 train CE 4.133e-05 (379s)
[13:21:43] step 177828 test 0.9807 train CE 1.919e-05 (670s)
wrote 23 checkpoints to /Users/brianna/Cells_Developmental/out/mnist_dense-b6
[2026-09-04 13:21:47] exit [4]: status 0 after 675s
[2026-09-04 13:21:47] queue finished: 4 commands
|