diff options
| author | YurenHao0426 <blackhao0426@gmail.com> | 2026-01-13 23:49:05 -0600 |
|---|---|---|
| committer | YurenHao0426 <blackhao0426@gmail.com> | 2026-01-13 23:49:05 -0600 |
| commit | cd99d6b874d9d09b3bb87b8485cc787885af71f1 (patch) | |
| tree | 59a233959932ca0e4f12f196275e07fcf443b33f /runs/slurm_logs/14363511_smnist.out | |
init commit
Diffstat (limited to 'runs/slurm_logs/14363511_smnist.out')
| -rw-r--r-- | runs/slurm_logs/14363511_smnist.out | 47 |
1 files changed, 47 insertions, 0 deletions
diff --git a/runs/slurm_logs/14363511_smnist.out b/runs/slurm_logs/14363511_smnist.out new file mode 100644 index 0000000..9885a77 --- /dev/null +++ b/runs/slurm_logs/14363511_smnist.out @@ -0,0 +1,47 @@ +============================================================ +Sequential MNIST Experiment +Job ID: 14363511 | Node: gpub039 +Start: Mon Dec 29 00:20:58 CST 2025 +============================================================ +NVIDIA A40, 46068 MiB +============================================================ +====================================================================== +BENCHMARK EXPERIMENT: SMNIST +====================================================================== +Depths: [2, 4, 6, 8, 10] +Hidden dim: 128 +Epochs: 30 +λ_reg: 0.3, λ_target: -0.1 +Device: cuda +====================================================================== + +Loading smnist dataset... +Dataset info: {'T': 784, 'D': 1, 'classes': 10, 'description': 'Sequential MNIST - 784 timesteps, 1 pixel at a time'} +Train batches: 469, Val batches: 79 + +============================================================ +Depth = 2 layers +============================================================ + + Training VANILLA... + Final: loss=1.0723 acc=0.645 val_acc=0.645 λ=N/A + + Training LYAPUNOV... + Final: loss=1.1430 acc=0.630 val_acc=0.627 λ=0.245 + +============================================================ +Depth = 4 layers +============================================================ + + Training VANILLA... + Final: loss=0.7247 acc=0.761 val_acc=0.773 λ=N/A + + Training LYAPUNOV... + Training diverged at epoch 25 + Final: loss=nan acc=0.000 val_acc=0.000 λ=N/A + +============================================================ +Depth = 6 layers +============================================================ + + Training VANILLA... |