1
|
{"args": {"mode": "bp", "dataset": "hierarchical", "depth": 6, "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_d6_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": 2.573746919631958}, {"step": 200, "epoch": 0, "train_loss": 1.7104246616363525}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.3318, "eval_loss": 1.8961854248046874, "val_acc": 0.3318, "val_loss": 1.8961854248046874}, {"step": 400, "epoch": 1, "train_loss": 1.7290828227996826}, {"step": 600, "epoch": 1, "train_loss": 1.70455801486969}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.3526, "eval_loss": 1.86725263671875, "val_acc": 0.3526, "val_loss": 1.86725263671875}, {"step": 800, "epoch": 2, "train_loss": 1.6812281608581543}, {"step": 1000, "epoch": 2, "train_loss": 1.768471121788025}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.361, "eval_loss": 1.8266659912109375, "val_acc": 0.361, "val_loss": 1.8266659912109375}, {"step": 1200, "epoch": 3, "train_loss": 1.5495350360870361}, {"step": 1400, "epoch": 3, "train_loss": 1.5867180824279785}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.3622, "eval_loss": 1.821352001953125, "val_acc": 0.3622, "val_loss": 1.821352001953125}, {"step": 1600, "epoch": 4, "train_loss": 1.618988275527954}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.3738, "eval_loss": 1.7969952880859374, "val_acc": 0.3738, "val_loss": 1.7969952880859374}, {"step": 1800, "epoch": 5, "train_loss": 1.7750898599624634}, {"step": 2000, "epoch": 5, "train_loss": 1.5564403533935547}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.3806, "eval_loss": 1.7641541259765625, "val_acc": 0.3806, "val_loss": 1.7641541259765625}, {"step": 2200, "epoch": 6, "train_loss": 1.5268113613128662}, {"step": 2400, "epoch": 6, "train_loss": 1.5403099060058594}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.3882, "eval_loss": 1.7595339111328125, "val_acc": 0.3882, "val_loss": 1.7595339111328125}, {"step": 2600, "epoch": 7, "train_loss": 1.541327714920044}, {"step": 2800, "epoch": 7, "train_loss": 1.6355960369110107}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.386, "eval_loss": 1.7467285888671875, "val_acc": 0.386, "val_loss": 1.7467285888671875}, {"step": 3000, "epoch": 8, "train_loss": 1.5069880485534668}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.3858, "eval_loss": 1.753902001953125, "val_acc": 0.3858, "val_loss": 1.753902001953125}, {"step": 3200, "epoch": 9, "train_loss": 1.5253907442092896}, {"step": 3400, "epoch": 9, "train_loss": 1.5076396465301514}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.3918, "eval_loss": 1.75735751953125, "val_acc": 0.3918, "val_loss": 1.75735751953125}, {"step": 3600, "epoch": 10, "train_loss": 1.4507845640182495}, {"step": 3800, "epoch": 10, "train_loss": 1.304982304573059}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.3936, "eval_loss": 1.7436959228515625, "val_acc": 0.3936, "val_loss": 1.7436959228515625}, {"step": 4000, "epoch": 11, "train_loss": 1.3849855661392212}, {"step": 4200, "epoch": 11, "train_loss": 1.523558497428894}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.3978, "eval_loss": 1.7391307373046876, "val_acc": 0.3978, "val_loss": 1.7391307373046876}, {"step": 4400, "epoch": 12, "train_loss": 1.6115436553955078}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.39, "eval_loss": 1.76820224609375, "val_acc": 0.39, "val_loss": 1.76820224609375}, {"step": 4600, "epoch": 13, "train_loss": 1.3322609663009644}, {"step": 4800, "epoch": 13, "train_loss": 1.4619946479797363}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.395, "eval_loss": 1.7727217529296875, "val_acc": 0.395, "val_loss": 1.7727217529296875}, {"step": 5000, "epoch": 14, "train_loss": 1.2391079664230347}, {"step": 5200, "epoch": 14, "train_loss": 1.359143853187561}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.3922, "eval_loss": 1.7666169189453125, "val_acc": 0.3922, "val_loss": 1.7666169189453125}, {"step": 5400, "epoch": 15, "train_loss": 1.615647792816162}, {"step": 5600, "epoch": 15, "train_loss": 1.3558441400527954}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.3928, "eval_loss": 1.76622841796875, "val_acc": 0.3928, "val_loss": 1.76622841796875}, {"step": 5800, "epoch": 16, "train_loss": 1.206676721572876}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.3952, "eval_loss": 1.775632373046875, "val_acc": 0.3952, "val_loss": 1.775632373046875}, {"step": 6000, "epoch": 17, "train_loss": 1.1862266063690186}, {"step": 6200, "epoch": 17, "train_loss": 1.1934031248092651}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.392, "eval_loss": 1.8034113525390625, "val_acc": 0.392, "val_loss": 1.8034113525390625}, {"step": 6400, "epoch": 18, "train_loss": 1.1822999715805054}, {"step": 6600, "epoch": 18, "train_loss": 1.2095683813095093}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.3952, "eval_loss": 1.8068008056640625, "val_acc": 0.3952, "val_loss": 1.8068008056640625}, {"step": 6800, "epoch": 19, "train_loss": 1.319027304649353}, {"step": 7000, "epoch": 19, "train_loss": 1.1257448196411133}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.3884, "eval_loss": 1.7925139404296875, "val_acc": 0.3884, "val_loss": 1.7925139404296875}, {"step": 7200, "epoch": 20, "train_loss": 1.3518521785736084}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.3874, "eval_loss": 1.8201495849609375, "val_acc": 0.3874, "val_loss": 1.8201495849609375}, {"step": 7400, "epoch": 21, "train_loss": 1.2303465604782104}, {"step": 7600, "epoch": 21, "train_loss": 1.2465145587921143}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.3864, "eval_loss": 1.8330891845703126, "val_acc": 0.3864, "val_loss": 1.8330891845703126}, {"step": 7800, "epoch": 22, "train_loss": 1.385786771774292}, {"step": 8000, "epoch": 22, "train_loss": 1.2793265581130981}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.3876, "eval_loss": 1.8138976806640625, "val_acc": 0.3876, "val_loss": 1.8138976806640625}, {"step": 8200, "epoch": 23, "train_loss": 1.1240808963775635}, {"step": 8400, "epoch": 23, "train_loss": 1.2807129621505737}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.392, "eval_loss": 1.8204913818359374, "val_acc": 0.392, "val_loss": 1.8204913818359374}, {"step": 8600, "epoch": 24, "train_loss": 1.1727039813995361}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.3862, "eval_loss": 1.8511544189453124, "val_acc": 0.3862, "val_loss": 1.8511544189453124}], "final": {"eval_split": "validation", "eval_acc": 0.3862, "eval_loss": 1.8511544189453124, "wall_s": 37.33476114273071, "val_acc": 0.3862, "val_loss": 1.8511544189453124}}
|
