diff options
Diffstat (limited to 'results/scale_v3_cifar10_sdil_w64_d5_s0.json')
| -rw-r--r-- | results/scale_v3_cifar10_sdil_w64_d5_s0.json | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/results/scale_v3_cifar10_sdil_w64_d5_s0.json b/results/scale_v3_cifar10_sdil_w64_d5_s0.json new file mode 100644 index 0000000..3ab5e98 --- /dev/null +++ b/results/scale_v3_cifar10_sdil_w64_d5_s0.json @@ -0,0 +1 @@ +{"args": {"mode": "sdil", "dataset": "cifar10", "depth": 5, "width": 64, "act": "tanh", "residual": 1, "epochs": 5, "batch_size": 128, "train_examples": 0, "task_seed": 0, "task_train_examples": 50000, "task_test_examples": 10000, "task_levels": 8, "task_n_in": 128, "task_classes": 10, "teacher_depth": 8, "teacher_width": 64, "teacher_residual": 1, "eta": 0.02, "eta_A": 0.02, "eta_P": 0.002, "momentum": 0.9, "w_scale": 1.0, "a_scale": 1.0, "feedback_scale": 1.0, "pert_sigma": 0.01, "pert_every": 4, "pert_ndirs": 16, "pert_mode": "simultaneous", "use_residual": 1, "raw_scale_control": "none", "learn_A": 1, "learn_P": 1, "p_neutral": 1, "p_warmup_steps": 200, "p_warmup_eta": 0.05, "nuis_rho": 0.0, "predictor_mode": "diagonal", "normalize_delta": 0, "settle_steps": 0, "kappa": 0.0, "feedback": "error", "seed": 0, "device": "cuda", "log_every": 500, "max_steps": 0, "probe_bs": 256, "outdir": "results", "tag": "scale_v3_cifar10_sdil_w64_d5_s0", "n_in": 3072, "n_out": 10}, "provenance": {"git_commit": "d4aef1a571c21e6e4e4125c90aa58da7f14e62e3", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.403229236602783, "cos_r_negg": [0.02090781182050705, -0.010267659090459347, -0.08521074056625366, -0.044122613966464996, 0.026557618752121925], "cos_innovation_negg": [0.02090781182050705, -0.010267659090459347, -0.08521074056625366, -0.044122613966464996, 0.026557618752121925], "cos_apical_negg": [0.02090781182050705, -0.010267659090459347, -0.08521074056625366, -0.044122613966464996, 0.026557618752121925], "cos_Ac_negg": [0.02090781182050705, -0.010267659090459347, -0.08521074056625366, -0.044122613966464996, 0.026557618752121925], "r_norm": [2.381490707397461, 2.5456581115722656, 2.4876675605773926, 2.3456616401672363, 2.358691692352295], "ldr": {"dL_sdil": 0.00894474983215332, "dL_bp": -0.33072471618652344, "ratio": -0.027045906744715822}}, {"epoch_end": 0, "step": 391, "test_acc": 0.3724, "test_loss": 1.7897538940429687}, {"step": 500, "epoch": 1, "train_loss": 1.6050856113433838, "cos_r_negg": [0.674854040145874, 0.10878533124923706, 0.0727112740278244, 0.03240722045302391, 0.08048097044229507], "cos_innovation_negg": [0.674854040145874, 0.10878533124923706, 0.0727112740278244, 0.03240722045302391, 0.08048097044229507], "cos_apical_negg": [0.674854040145874, 0.10878533124923706, 0.0727112740278244, 0.03240722045302391, 0.08048097044229507], "cos_Ac_negg": [0.674854040145874, 0.10878533124923706, 0.0727112740278244, 0.03240722045302391, 0.08048097044229507], "r_norm": [1.7436394691467285, 1.7945313453674316, 1.7393816709518433, 1.6361432075500488, 1.6675000190734863]}, {"epoch_end": 1, "step": 782, "test_acc": 0.3985, "test_loss": 1.736846240234375}, {"step": 1000, "epoch": 2, "train_loss": 1.7890269756317139, "cos_r_negg": [0.6762683391571045, 0.20309486985206604, 0.1259402632713318, 0.17170320451259613, 0.1670699417591095], "cos_innovation_negg": [0.6762683391571045, 0.20309486985206604, 0.1259402632713318, 0.17170320451259613, 0.1670699417591095], "cos_apical_negg": [0.6762683391571045, 0.20309486985206604, 0.1259402632713318, 0.17170320451259613, 0.1670699417591095], "cos_Ac_negg": [0.6762683391571045, 0.20309486985206604, 0.1259402632713318, 0.17170320451259613, 0.1670699417591095], "r_norm": [1.512650728225708, 1.487385630607605, 1.434342861175537, 1.34562087059021, 1.3828237056732178]}, {"epoch_end": 2, "step": 1173, "test_acc": 0.3991, "test_loss": 1.7149715942382813}, {"step": 1500, "epoch": 3, "train_loss": 1.6189053058624268, "cos_r_negg": [0.7615132927894592, 0.2767869830131531, 0.24810822308063507, 0.2945393919944763, 0.28578731417655945], "cos_innovation_negg": [0.7615132927894592, 0.2767869830131531, 0.24810822308063507, 0.2945393919944763, 0.28578731417655945], "cos_apical_negg": [0.7615132927894592, 0.2767869830131531, 0.24810822308063507, 0.2945393919944763, 0.28578731417655945], "cos_Ac_negg": [0.7615132927894592, 0.2767869830131531, 0.24810822308063507, 0.2945393919944763, 0.28578731417655945], "r_norm": [1.3173727989196777, 1.231938123703003, 1.1819707155227661, 1.1142072677612305, 1.1422107219696045]}, {"epoch_end": 3, "step": 1564, "test_acc": 0.4076, "test_loss": 1.7007428833007812}, {"epoch_end": 4, "step": 1955, "test_acc": 0.4149, "test_loss": 1.6846970092773437}], "final": {"test_acc": 0.4149, "test_loss": 1.6846970092773437, "wall_s": 8.788044691085815, "cos_r_negg": [0.775210440158844, 0.3959091007709503, 0.3543007969856262, 0.4048098921775818, 0.3978502154350281], "cos_innovation_negg": [0.775210440158844, 0.3959091007709503, 0.3543007969856262, 0.4048098921775818, 0.3978502154350281], "cos_apical_negg": [0.775210440158844, 0.3959091007709503, 0.3543007969856262, 0.4048098921775818, 0.3978502154350281], "cos_Ac_negg": [0.775210440158844, 0.3959091007709503, 0.3543007969856262, 0.4048098921775818, 0.3978502154350281]}}
\ No newline at end of file |