1
|
{"args": {"mode": "bp", "dataset": "hierarchical", "depth": 4, "width": 16, "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_w16_d4_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": 2.383604049682617}, {"step": 200, "epoch": 0, "train_loss": 1.7610948085784912}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.3338, "eval_loss": 1.912164501953125, "val_acc": 0.3338, "val_loss": 1.912164501953125}, {"step": 400, "epoch": 1, "train_loss": 1.7126342058181763}, {"step": 600, "epoch": 1, "train_loss": 1.7328016757965088}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.3556, "eval_loss": 1.8727889404296876, "val_acc": 0.3556, "val_loss": 1.8727889404296876}, {"step": 800, "epoch": 2, "train_loss": 1.6682446002960205}, {"step": 1000, "epoch": 2, "train_loss": 1.7850028276443481}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.3548, "eval_loss": 1.8369991455078125, "val_acc": 0.3548, "val_loss": 1.8369991455078125}, {"step": 1200, "epoch": 3, "train_loss": 1.6067845821380615}, {"step": 1400, "epoch": 3, "train_loss": 1.6644831895828247}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.3516, "eval_loss": 1.839272216796875, "val_acc": 0.3516, "val_loss": 1.839272216796875}, {"step": 1600, "epoch": 4, "train_loss": 1.671094536781311}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.3612, "eval_loss": 1.8231252685546875, "val_acc": 0.3612, "val_loss": 1.8231252685546875}, {"step": 1800, "epoch": 5, "train_loss": 1.8178144693374634}, {"step": 2000, "epoch": 5, "train_loss": 1.6754566431045532}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.3598, "eval_loss": 1.810904345703125, "val_acc": 0.3598, "val_loss": 1.810904345703125}, {"step": 2200, "epoch": 6, "train_loss": 1.610650897026062}, {"step": 2400, "epoch": 6, "train_loss": 1.6244782209396362}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.3618, "eval_loss": 1.8133049072265626, "val_acc": 0.3618, "val_loss": 1.8133049072265626}, {"step": 2600, "epoch": 7, "train_loss": 1.6868948936462402}, {"step": 2800, "epoch": 7, "train_loss": 1.7593493461608887}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.3684, "eval_loss": 1.7950828125, "val_acc": 0.3684, "val_loss": 1.7950828125}, {"step": 3000, "epoch": 8, "train_loss": 1.66415536403656}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.3674, "eval_loss": 1.79699560546875, "val_acc": 0.3674, "val_loss": 1.79699560546875}, {"step": 3200, "epoch": 9, "train_loss": 1.7509492635726929}, {"step": 3400, "epoch": 9, "train_loss": 1.5698741674423218}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.3674, "eval_loss": 1.79268994140625, "val_acc": 0.3674, "val_loss": 1.79268994140625}, {"step": 3600, "epoch": 10, "train_loss": 1.6385778188705444}, {"step": 3800, "epoch": 10, "train_loss": 1.5268993377685547}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.3666, "eval_loss": 1.7868793701171874, "val_acc": 0.3666, "val_loss": 1.7868793701171874}, {"step": 4000, "epoch": 11, "train_loss": 1.5952410697937012}, {"step": 4200, "epoch": 11, "train_loss": 1.6397472620010376}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.377, "eval_loss": 1.7727826416015624, "val_acc": 0.377, "val_loss": 1.7727826416015624}, {"step": 4400, "epoch": 12, "train_loss": 1.8183878660202026}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.3684, "eval_loss": 1.793745166015625, "val_acc": 0.3684, "val_loss": 1.793745166015625}, {"step": 4600, "epoch": 13, "train_loss": 1.5248687267303467}, {"step": 4800, "epoch": 13, "train_loss": 1.699881672859192}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.3644, "eval_loss": 1.7913708740234375, "val_acc": 0.3644, "val_loss": 1.7913708740234375}, {"step": 5000, "epoch": 14, "train_loss": 1.397655963897705}, {"step": 5200, "epoch": 14, "train_loss": 1.6332428455352783}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.3708, "eval_loss": 1.789706640625, "val_acc": 0.3708, "val_loss": 1.789706640625}, {"step": 5400, "epoch": 15, "train_loss": 1.8139365911483765}, {"step": 5600, "epoch": 15, "train_loss": 1.5195242166519165}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.3718, "eval_loss": 1.770397412109375, "val_acc": 0.3718, "val_loss": 1.770397412109375}, {"step": 5800, "epoch": 16, "train_loss": 1.5627028942108154}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.3706, "eval_loss": 1.784296728515625, "val_acc": 0.3706, "val_loss": 1.784296728515625}, {"step": 6000, "epoch": 17, "train_loss": 1.5166749954223633}, {"step": 6200, "epoch": 17, "train_loss": 1.5427132844924927}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.3742, "eval_loss": 1.7833609619140625, "val_acc": 0.3742, "val_loss": 1.7833609619140625}, {"step": 6400, "epoch": 18, "train_loss": 1.4202320575714111}, {"step": 6600, "epoch": 18, "train_loss": 1.5325849056243896}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.3778, "eval_loss": 1.7743547607421875, "val_acc": 0.3778, "val_loss": 1.7743547607421875}, {"step": 6800, "epoch": 19, "train_loss": 1.5907431840896606}, {"step": 7000, "epoch": 19, "train_loss": 1.464917778968811}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.3822, "eval_loss": 1.760672021484375, "val_acc": 0.3822, "val_loss": 1.760672021484375}, {"step": 7200, "epoch": 20, "train_loss": 1.6378196477890015}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.378, "eval_loss": 1.7727245849609374, "val_acc": 0.378, "val_loss": 1.7727245849609374}, {"step": 7400, "epoch": 21, "train_loss": 1.5653473138809204}, {"step": 7600, "epoch": 21, "train_loss": 1.5427693128585815}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.3744, "eval_loss": 1.7742513916015625, "val_acc": 0.3744, "val_loss": 1.7742513916015625}, {"step": 7800, "epoch": 22, "train_loss": 1.6414825916290283}, {"step": 8000, "epoch": 22, "train_loss": 1.6095434427261353}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.3796, "eval_loss": 1.769055322265625, "val_acc": 0.3796, "val_loss": 1.769055322265625}, {"step": 8200, "epoch": 23, "train_loss": 1.444868564605713}, {"step": 8400, "epoch": 23, "train_loss": 1.65962553024292}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.3826, "eval_loss": 1.7572109619140626, "val_acc": 0.3826, "val_loss": 1.7572109619140626}, {"step": 8600, "epoch": 24, "train_loss": 1.497135877609253}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.3822, "eval_loss": 1.752923876953125, "val_acc": 0.3822, "val_loss": 1.752923876953125}], "final": {"eval_split": "validation", "eval_acc": 0.3822, "eval_loss": 1.752923876953125, "wall_s": 29.056323766708374, "val_acc": 0.3822, "val_loss": 1.752923876953125}}
|
