diff options
| author | YurenHao0426 <blackhao0426@gmail.com> | 2026-01-13 23:50:59 -0600 |
|---|---|---|
| committer | YurenHao0426 <blackhao0426@gmail.com> | 2026-01-13 23:50:59 -0600 |
| commit | 00cf667cee7ffacb144d5805fc7e0ef443f3583a (patch) | |
| tree | 77d20a3adaecf96bf3aff0612bdd3b5fa1a7dc7e /runs/slurm_logs/15261460_extreme4.out | |
| parent | c53c04aa1d6ff75cb478a9498c370baa929c74b6 (diff) | |
| parent | cd99d6b874d9d09b3bb87b8485cc787885af71f1 (diff) | |
Merge master into main
Diffstat (limited to 'runs/slurm_logs/15261460_extreme4.out')
| -rw-r--r-- | runs/slurm_logs/15261460_extreme4.out | 91 |
1 files changed, 91 insertions, 0 deletions
diff --git a/runs/slurm_logs/15261460_extreme4.out b/runs/slurm_logs/15261460_extreme4.out new file mode 100644 index 0000000..b8562fd --- /dev/null +++ b/runs/slurm_logs/15261460_extreme4.out @@ -0,0 +1,91 @@ +============================================================ +EXTREME THRESHOLD 4.0 Experiment +Job ID: 15261460 | Node: gpub099 +Start: Sat Jan 3 09:41:01 CST 2026 +============================================================ +NVIDIA A40, 46068 MiB +============================================================ +================================================================================ +DEPTH SCALING BENCHMARK +================================================================================ +Dataset: cifar100 +Depths: [4, 8, 12, 16] +Timesteps: 4 +Epochs: 150 +λ_reg: 0.3, λ_target: -0.1 +Reg type: extreme, Warmup epochs: 10 +Device: cuda +================================================================================ + +Loading cifar100... +Classes: 100, Input: (3, 32, 32) +Train: 50000, Test: 10000 + +Depth configurations: [(4, '4×1'), (8, '4×2'), (12, '4×3'), (16, '4×4')] +Regularization type: extreme +Warmup epochs: 10 +Stable init: False +Lyapunov threshold: 4.0 + +============================================================ +Depth = 4 conv layers (4 stages × 1 blocks) +============================================================ + Vanilla: depth=4, params=1,756,836 + Epoch 10: train=0.493 test=0.438 σ=9.40e-01/3.53e-08 + Epoch 20: train=0.629 test=0.493 σ=5.78e-01/2.42e-08 + Epoch 30: train=0.701 test=0.500 σ=4.69e-01/2.00e-08 + Epoch 40: train=0.754 test=0.540 σ=4.17e-01/1.73e-08 + Epoch 50: train=0.799 test=0.567 σ=3.66e-01/1.53e-08 + Epoch 60: train=0.830 test=0.587 σ=3.42e-01/1.40e-08 + Epoch 70: train=0.858 test=0.583 σ=3.21e-01/1.29e-08 + Epoch 80: train=0.881 test=0.577 σ=2.97e-01/1.20e-08 + Epoch 90: train=0.903 test=0.597 σ=2.85e-01/1.13e-08 + Epoch 100: train=0.920 test=0.599 σ=2.61e-01/1.03e-08 + Epoch 110: train=0.931 test=0.608 σ=2.64e-01/9.94e-09 + Epoch 120: train=0.941 test=0.602 σ=2.38e-01/9.00e-09 + Epoch 130: train=0.947 test=0.609 σ=2.41e-01/8.84e-09 + Epoch 140: train=0.949 test=0.613 σ=2.32e-01/8.45e-09 + Epoch 150: train=0.951 test=0.607 σ=2.44e-01/8.91e-09 + Best test acc: 0.613 + Lyapunov: depth=4, params=1,756,836 + Epoch 10: train=0.492 test=0.420 λ=2.049 σ=9.51e-01/3.58e-08 + Epoch 20: train=0.625 test=0.500 λ=1.999 σ=5.91e-01/2.46e-08 + Epoch 30: train=0.701 test=0.533 λ=1.975 σ=4.87e-01/2.04e-08 + Epoch 40: train=0.754 test=0.547 λ=1.964 σ=4.27e-01/1.73e-08 + Epoch 50: train=0.797 test=0.563 λ=1.957 σ=3.73e-01/1.54e-08 + Epoch 60: train=0.833 test=0.569 λ=1.964 σ=3.30e-01/1.38e-08 + Epoch 70: train=0.859 test=0.571 λ=1.968 σ=3.28e-01/1.31e-08 + Epoch 80: train=0.884 test=0.600 λ=1.961 σ=3.04e-01/1.20e-08 + Epoch 90: train=0.903 test=0.587 λ=1.967 σ=2.87e-01/1.11e-08 + Epoch 100: train=0.919 test=0.594 λ=1.964 σ=2.72e-01/1.04e-08 + Epoch 110: train=0.933 test=0.602 λ=1.966 σ=2.57e-01/9.82e-09 + Epoch 120: train=0.942 test=0.596 λ=1.972 σ=2.45e-01/9.57e-09 + Epoch 130: train=0.949 test=0.602 λ=1.971 σ=2.35e-01/8.64e-09 + Epoch 140: train=0.949 test=0.606 λ=1.975 σ=2.39e-01/8.83e-09 + Epoch 150: train=0.951 test=0.603 λ=1.972 σ=2.33e-01/8.59e-09 + Best test acc: 0.609 + +============================================================ +Depth = 8 conv layers (4 stages × 2 blocks) +============================================================ + Vanilla: depth=8, params=4,892,196 + Epoch 10: train=0.394 test=0.359 σ=8.68e-01/3.12e-08 + Epoch 20: train=0.541 test=0.435 σ=4.85e-01/2.20e-08 + Epoch 30: train=0.635 test=0.468 σ=3.82e-01/1.77e-08 + Epoch 40: train=0.698 test=0.479 σ=3.30e-01/1.56e-08 + Epoch 50: train=0.752 test=0.496 σ=3.12e-01/1.41e-08 + Epoch 60: train=0.798 test=0.511 σ=2.87e-01/1.26e-08 + Epoch 70: train=0.838 test=0.513 σ=2.71e-01/1.15e-08 + Epoch 80: train=0.869 test=0.519 σ=2.50e-01/1.06e-08 + Epoch 90: train=0.897 test=0.527 σ=2.39e-01/9.71e-09 + Epoch 100: train=0.919 test=0.515 σ=2.26e-01/9.04e-09 + Epoch 110: train=0.934 test=0.518 σ=2.08e-01/8.20e-09 + Epoch 120: train=0.944 test=0.525 σ=2.15e-01/8.02e-09 + Epoch 130: train=0.951 test=0.529 σ=2.06e-01/7.60e-09 + Epoch 140: train=0.955 test=0.536 σ=1.94e-01/7.42e-09 + Epoch 150: train=0.957 test=0.521 σ=2.00e-01/7.50e-09 + Best test acc: 0.539 + Lyapunov: depth=8, params=4,892,196 + Epoch 10: train=0.395 test=0.363 λ=2.893 σ=7.82e-01/3.03e-08 + Epoch 20: train=0.547 test=0.425 λ=2.878 σ=4.66e-01/2.14e-08 + Epoch 30: train=0.632 test=0.481 λ=2.864 σ=3.80e-01/1.77e-08 |
