1
|
{"args": {"mode": "sdil", "dataset": "mnist", "depth": 3, "width": 256, "act": "tanh", "residual": 1, "epochs": 15, "batch_size": 128, "train_examples": 0, "val_examples": 0, "split_seed": 2027, "eval_split": "test", "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": 0.05, "eta_A": 0.02, "eta_P": 0.05, "momentum": 0.9, "w_scale": 1.0, "a_scale": 1.0, "feedback_scale": 1.0, "pert_sigma": 0.01, "pert_every": 4, "pert_ndirs": 1, "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, "traffic_seed": 1234, "traffic_mode": "none", "predictor_mode": "diagonal", "normalize_delta": 0, "settle_steps": 0, "kappa": 0.0, "feedback": "error", "seed": 2, "device": "cuda", "log_every": 100000, "diagnostics": "alignment", "diagnostics_schedule": "final", "eval_every": 0, "max_steps": 0, "probe_bs": 512, "outdir": "results", "tag": "c1_confirm_v1_mnist_none_rho0_t1234_residual_d3_s2", "n_in": 784, "n_out": 10}, "split": {"dataset": "mnist", "split_seed": null, "validation_examples": 0, "validation_index_sha256": null, "validation_class_counts": {}, "train_examples": 60000, "test_examples": 10000, "split_from_training_only": true, "evaluation_split": "test"}, "provenance": {"git_commit": "18c9c1c1d1ec31560c78d41f4ff6957339c1e7c7", "git_dirty": false}, "diagnostic_protocol": {"probe_source": "training_prefix", "probe_examples": 512, "schedule": "final"}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.4491982460021973}, {"epoch_end": 0, "step": 469}, {"epoch_end": 1, "step": 938}, {"epoch_end": 2, "step": 1407}, {"epoch_end": 3, "step": 1876}, {"epoch_end": 4, "step": 2345}, {"epoch_end": 5, "step": 2814}, {"epoch_end": 6, "step": 3283}, {"epoch_end": 7, "step": 3752}, {"epoch_end": 8, "step": 4221}, {"epoch_end": 9, "step": 4690}, {"epoch_end": 10, "step": 5159}, {"epoch_end": 11, "step": 5628}, {"epoch_end": 12, "step": 6097}, {"epoch_end": 13, "step": 6566}, {"epoch_end": 14, "step": 7035}], "final": {"eval_split": "test", "eval_acc": 0.9737, "eval_loss": 0.11859232807159424, "wall_s": 22.313729524612427, "test_acc": 0.9737, "test_loss": 0.11859232807159424, "cos_r_negg": [0.6204646229743958, 0.3519253730773926, 0.32553520798683167], "cos_innovation_negg": [0.6204646229743958, 0.3519253730773926, 0.32553520798683167], "cos_apical_negg": [0.6204646229743958, 0.3519253730773926, 0.32553520798683167], "cos_Ac_negg": [0.6204646229743958, 0.3519253730773926, 0.32553520798683167], "r_norm": [0.00880928710103035, 0.00863414816558361, 0.00877653993666172], "g_norm": [2.5090203052968718e-05, 1.8679147615330294e-05, 1.479045477026375e-05], "traffic_norm": [0.0, 0.0, 0.0], "traffic_residual_norm": [0.0, 0.0, 0.0], "traffic_r2": [NaN, NaN, NaN]}, "timing": {"warmup_wall_s": 0.0, "training_loop_wall_s": 22.20801615715027, "diagnostics_wall_s": 0.0666956901550293, "inline_diagnostics_wall_s": 0.0, "optimizer_wall_s_excluding_inline_diagnostics": 22.20801615715027, "evaluation_wall_s": 0.0376126766204834}, "cost": {"train_steps": 7035, "ordinary_training_forward_examples": 900000, "predictor_warmup_forward_examples": 0, "perturbation_events": 1759, "calibration_batch_loss_evaluations": 3518, "calibration_example_loss_evaluations": 450048, "calibration_forward_equivalent_examples": 675072.0, "training_forward_equivalent_examples": 1575072.0, "max_perturbation_batch_expansion": 2}, "hardware": {"device": "cuda", "torch_version": "2.3.1+cu118", "cuda_device_name": "NVIDIA GeForce GTX 1080", "cuda_visible_devices": "5", "device_total_memory_bytes": 8507949056, "peak_memory_allocated_bytes": 250861568, "peak_memory_reserved_bytes": 255852544}}
|
