1
|
{"args": {"mode": "sdil", "dataset": "fmnist", "depth": 3, "width": 256, "act": "tanh", "residual": 1, "epochs": 15, "batch_size": 128, "train_examples": 0, "val_examples": 5000, "split_seed": 2027, "eval_split": "validation", "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": 0, "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": 0, "device": "cuda", "log_every": 100000, "diagnostics": "alignment", "diagnostics_schedule": "final", "eval_every": 0, "max_steps": 0, "probe_bs": 512, "outdir": "results", "tag": "c1_fmnist_val_v1_none_rho0_t1234_raw_d3_s0", "n_in": 784, "n_out": 10}, "split": {"dataset": "fmnist", "split_seed": 2027, "validation_examples": 5000, "validation_index_sha256": "c5f293471287d93b4149c9ca0f9aec57df1ddb6d499234fcf6fd45f4f80ddf58", "validation_class_counts": {"0": 500, "1": 500, "2": 500, "3": 500, "4": 500, "5": 500, "6": 500, "7": 500, "8": 500, "9": 500}, "train_examples": 55000, "test_examples": 10000, "split_from_training_only": true, "evaluation_split": "validation"}, "provenance": {"git_commit": "26407ca3e958ef3701cc69afae4973f023b5b19e", "git_dirty": false}, "diagnostic_protocol": {"probe_source": "training_prefix", "probe_examples": 512, "schedule": "final"}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.502072334289551}, {"epoch_end": 0, "step": 430}, {"epoch_end": 1, "step": 860}, {"epoch_end": 2, "step": 1290}, {"epoch_end": 3, "step": 1720}, {"epoch_end": 4, "step": 2150}, {"epoch_end": 5, "step": 2580}, {"epoch_end": 6, "step": 3010}, {"epoch_end": 7, "step": 3440}, {"epoch_end": 8, "step": 3870}, {"epoch_end": 9, "step": 4300}, {"epoch_end": 10, "step": 4730}, {"epoch_end": 11, "step": 5160}, {"epoch_end": 12, "step": 5590}, {"epoch_end": 13, "step": 6020}, {"epoch_end": 14, "step": 6450}], "final": {"eval_split": "validation", "eval_acc": 0.88, "eval_loss": 0.41392838134765625, "wall_s": 20.670273065567017, "val_acc": 0.88, "val_loss": 0.41392838134765625, "cos_r_negg": [0.8215733766555786, 0.5133273601531982, 0.4824512004852295], "cos_innovation_negg": [0.8215733766555786, 0.5133273601531982, 0.4824512004852295], "cos_apical_negg": [0.8215733766555786, 0.5133273601531982, 0.4824512004852295], "cos_Ac_negg": [0.8215733766555786, 0.5133273601531982, 0.4824512004852295], "r_norm": [0.49292951822280884, 0.4109225571155548, 0.41594964265823364], "g_norm": [0.0011965048033744097, 0.0011563380248844624, 0.0011454486520960927], "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": 20.555959939956665, "diagnostics_wall_s": 0.07414674758911133, "inline_diagnostics_wall_s": 0.0, "optimizer_wall_s_excluding_inline_diagnostics": 20.555959939956665, "evaluation_wall_s": 0.031392812728881836}, "cost": {"train_steps": 6450, "ordinary_training_forward_examples": 825000, "predictor_warmup_forward_examples": 0, "perturbation_events": 1613, "calibration_batch_loss_evaluations": 3226, "calibration_example_loss_evaluations": 412928, "calibration_forward_equivalent_examples": 619392.0, "training_forward_equivalent_examples": 1444392.0, "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": 218714624, "peak_memory_reserved_bytes": 236978176}}
|
