1
|
{"args": {"mode": "bp", "dataset": "hierarchical", "depth": 12, "width": 64, "act": "tanh", "residual": 1, "epochs": 25, "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": 6, "task_n_in": 128, "task_classes": 10, "teacher_depth": 8, "teacher_width": 64, "teacher_residual": 1, "eta": 0.03, "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, "traffic_mode": "soma", "predictor_mode": "diagonal", "normalize_delta": 0, "settle_steps": 0, "kappa": 0.0, "feedback": "error", "seed": 0, "device": "cuda", "log_every": 200, "max_steps": 0, "probe_bs": 512, "outdir": "results", "tag": "depthtask_dev_v1_hierarchical_bp_w64_d12_t0_s0", "n_in": 64, "n_out": 10}, "split": {"dataset": "hierarchical", "task_seed": 0, "train_examples": 45000, "test_examples": 10000, "evaluation_split": "validation", "synthetic_generator": true, "split_seed": 2027, "validation_examples": 5000, "validation_index_sha256": "f5282ed201086357dbe24c23e70cc9eb86deaaf70d99c41553ffc78626ee8d6c", "validation_class_counts": {"0": 500, "1": 500, "2": 500, "3": 500, "4": 500, "5": 500, "6": 500, "7": 500, "8": 500, "9": 500}, "split_from_training_only": true}, "provenance": {"git_commit": "a0696a3d72cc5bf69876e1716b74f7e1ac525dd0", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.5225329399108887}, {"step": 200, "epoch": 0, "train_loss": 1.7906372547149658}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.332, "eval_loss": 1.904244287109375, "val_acc": 0.332, "val_loss": 1.904244287109375}, {"step": 400, "epoch": 1, "train_loss": 1.7433497905731201}, {"step": 600, "epoch": 1, "train_loss": 1.717389702796936}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.3442, "eval_loss": 1.8789136962890625, "val_acc": 0.3442, "val_loss": 1.8789136962890625}, {"step": 800, "epoch": 2, "train_loss": 1.6778650283813477}, {"step": 1000, "epoch": 2, "train_loss": 1.7652276754379272}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.3556, "eval_loss": 1.8470677001953124, "val_acc": 0.3556, "val_loss": 1.8470677001953124}, {"step": 1200, "epoch": 3, "train_loss": 1.523913860321045}, {"step": 1400, "epoch": 3, "train_loss": 1.5748168230056763}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.3574, "eval_loss": 1.8402134033203126, "val_acc": 0.3574, "val_loss": 1.8402134033203126}, {"step": 1600, "epoch": 4, "train_loss": 1.5846136808395386}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.372, "eval_loss": 1.808778759765625, "val_acc": 0.372, "val_loss": 1.808778759765625}, {"step": 1800, "epoch": 5, "train_loss": 1.8003953695297241}, {"step": 2000, "epoch": 5, "train_loss": 1.5881260633468628}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.375, "eval_loss": 1.7751354248046876, "val_acc": 0.375, "val_loss": 1.7751354248046876}, {"step": 2200, "epoch": 6, "train_loss": 1.5656712055206299}, {"step": 2400, "epoch": 6, "train_loss": 1.550453782081604}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.3904, "eval_loss": 1.771417041015625, "val_acc": 0.3904, "val_loss": 1.771417041015625}, {"step": 2600, "epoch": 7, "train_loss": 1.5640873908996582}, {"step": 2800, "epoch": 7, "train_loss": 1.6734482049942017}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.3942, "eval_loss": 1.7509275146484375, "val_acc": 0.3942, "val_loss": 1.7509275146484375}, {"step": 3000, "epoch": 8, "train_loss": 1.5729361772537231}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.3864, "eval_loss": 1.770663134765625, "val_acc": 0.3864, "val_loss": 1.770663134765625}, {"step": 3200, "epoch": 9, "train_loss": 1.5623794794082642}, {"step": 3400, "epoch": 9, "train_loss": 1.500632882118225}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.3804, "eval_loss": 1.77856767578125, "val_acc": 0.3804, "val_loss": 1.77856767578125}, {"step": 3600, "epoch": 10, "train_loss": 1.486441969871521}, {"step": 3800, "epoch": 10, "train_loss": 1.365587592124939}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.3812, "eval_loss": 1.7705543701171875, "val_acc": 0.3812, "val_loss": 1.7705543701171875}, {"step": 4000, "epoch": 11, "train_loss": 1.4871476888656616}, {"step": 4200, "epoch": 11, "train_loss": 1.5890864133834839}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.388, "eval_loss": 1.7552383056640626, "val_acc": 0.388, "val_loss": 1.7552383056640626}, {"step": 4400, "epoch": 12, "train_loss": 1.6518833637237549}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.3834, "eval_loss": 1.778572607421875, "val_acc": 0.3834, "val_loss": 1.778572607421875}, {"step": 4600, "epoch": 13, "train_loss": 1.382397174835205}, {"step": 4800, "epoch": 13, "train_loss": 1.422805666923523}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.3864, "eval_loss": 1.792176416015625, "val_acc": 0.3864, "val_loss": 1.792176416015625}, {"step": 5000, "epoch": 14, "train_loss": 1.1823656558990479}, {"step": 5200, "epoch": 14, "train_loss": 1.418053150177002}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.3936, "eval_loss": 1.7863834716796876, "val_acc": 0.3936, "val_loss": 1.7863834716796876}, {"step": 5400, "epoch": 15, "train_loss": 1.6090099811553955}, {"step": 5600, "epoch": 15, "train_loss": 1.3677173852920532}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.3938, "eval_loss": 1.771215478515625, "val_acc": 0.3938, "val_loss": 1.771215478515625}, {"step": 5800, "epoch": 16, "train_loss": 1.1730642318725586}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.391, "eval_loss": 1.77024072265625, "val_acc": 0.391, "val_loss": 1.77024072265625}, {"step": 6000, "epoch": 17, "train_loss": 1.1396112442016602}, {"step": 6200, "epoch": 17, "train_loss": 1.1576471328735352}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.3838, "eval_loss": 1.826544287109375, "val_acc": 0.3838, "val_loss": 1.826544287109375}, {"step": 6400, "epoch": 18, "train_loss": 1.109771966934204}, {"step": 6600, "epoch": 18, "train_loss": 1.2505069971084595}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.3872, "eval_loss": 1.8254947265625, "val_acc": 0.3872, "val_loss": 1.8254947265625}, {"step": 6800, "epoch": 19, "train_loss": 1.2415235042572021}, {"step": 7000, "epoch": 19, "train_loss": 1.1805624961853027}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.3862, "eval_loss": 1.8145503173828126, "val_acc": 0.3862, "val_loss": 1.8145503173828126}, {"step": 7200, "epoch": 20, "train_loss": 1.3972448110580444}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.3864, "eval_loss": 1.8674730712890626, "val_acc": 0.3864, "val_loss": 1.8674730712890626}, {"step": 7400, "epoch": 21, "train_loss": 1.1311677694320679}, {"step": 7600, "epoch": 21, "train_loss": 1.1309796571731567}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.3832, "eval_loss": 1.8787137939453125, "val_acc": 0.3832, "val_loss": 1.8787137939453125}, {"step": 7800, "epoch": 22, "train_loss": 1.3694546222686768}, {"step": 8000, "epoch": 22, "train_loss": 1.1945430040359497}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.3886, "eval_loss": 1.8810318603515626, "val_acc": 0.3886, "val_loss": 1.8810318603515626}, {"step": 8200, "epoch": 23, "train_loss": 1.033811330795288}, {"step": 8400, "epoch": 23, "train_loss": 1.1782039403915405}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.3818, "eval_loss": 1.927628955078125, "val_acc": 0.3818, "val_loss": 1.927628955078125}, {"step": 8600, "epoch": 24, "train_loss": 1.1393238306045532}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.373, "eval_loss": 1.9600396728515626, "val_acc": 0.373, "val_loss": 1.9600396728515626}], "final": {"eval_split": "validation", "eval_acc": 0.373, "eval_loss": 1.9600396728515626, "wall_s": 61.94551920890808, "val_acc": 0.373, "val_loss": 1.9600396728515626}}
|
