1
|
{"args": {"mode": "bp", "dataset": "hierarchical", "depth": 2, "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_d2_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": 2.6634795665740967}, {"step": 200, "epoch": 0, "train_loss": 1.7392361164093018}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.3322, "eval_loss": 1.8970150634765626, "val_acc": 0.3322, "val_loss": 1.8970150634765626}, {"step": 400, "epoch": 1, "train_loss": 1.7551640272140503}, {"step": 600, "epoch": 1, "train_loss": 1.7312787771224976}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.3508, "eval_loss": 1.87037197265625, "val_acc": 0.3508, "val_loss": 1.87037197265625}, {"step": 800, "epoch": 2, "train_loss": 1.6509016752243042}, {"step": 1000, "epoch": 2, "train_loss": 1.7625491619110107}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.3604, "eval_loss": 1.8333597412109375, "val_acc": 0.3604, "val_loss": 1.8333597412109375}, {"step": 1200, "epoch": 3, "train_loss": 1.5264045000076294}, {"step": 1400, "epoch": 3, "train_loss": 1.5917079448699951}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.3616, "eval_loss": 1.8280685302734374, "val_acc": 0.3616, "val_loss": 1.8280685302734374}, {"step": 1600, "epoch": 4, "train_loss": 1.6507923603057861}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.3738, "eval_loss": 1.8006971435546875, "val_acc": 0.3738, "val_loss": 1.8006971435546875}, {"step": 1800, "epoch": 5, "train_loss": 1.8038511276245117}, {"step": 2000, "epoch": 5, "train_loss": 1.5552788972854614}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.3748, "eval_loss": 1.7877079833984375, "val_acc": 0.3748, "val_loss": 1.7877079833984375}, {"step": 2200, "epoch": 6, "train_loss": 1.5938795804977417}, {"step": 2400, "epoch": 6, "train_loss": 1.5603779554367065}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.3842, "eval_loss": 1.7843631103515625, "val_acc": 0.3842, "val_loss": 1.7843631103515625}, {"step": 2600, "epoch": 7, "train_loss": 1.6169517040252686}, {"step": 2800, "epoch": 7, "train_loss": 1.7346149682998657}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.384, "eval_loss": 1.768867138671875, "val_acc": 0.384, "val_loss": 1.768867138671875}, {"step": 3000, "epoch": 8, "train_loss": 1.515174388885498}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.3788, "eval_loss": 1.767204150390625, "val_acc": 0.3788, "val_loss": 1.767204150390625}, {"step": 3200, "epoch": 9, "train_loss": 1.5812177658081055}, {"step": 3400, "epoch": 9, "train_loss": 1.5104117393493652}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.3828, "eval_loss": 1.7606010986328124, "val_acc": 0.3828, "val_loss": 1.7606010986328124}, {"step": 3600, "epoch": 10, "train_loss": 1.5871082544326782}, {"step": 3800, "epoch": 10, "train_loss": 1.4036861658096313}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.3834, "eval_loss": 1.7639999267578126, "val_acc": 0.3834, "val_loss": 1.7639999267578126}, {"step": 4000, "epoch": 11, "train_loss": 1.4818824529647827}, {"step": 4200, "epoch": 11, "train_loss": 1.6143189668655396}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.3918, "eval_loss": 1.7417701904296874, "val_acc": 0.3918, "val_loss": 1.7417701904296874}, {"step": 4400, "epoch": 12, "train_loss": 1.7102020978927612}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.386, "eval_loss": 1.7647007080078125, "val_acc": 0.386, "val_loss": 1.7647007080078125}, {"step": 4600, "epoch": 13, "train_loss": 1.40860116481781}, {"step": 4800, "epoch": 13, "train_loss": 1.6281629800796509}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.3888, "eval_loss": 1.758567529296875, "val_acc": 0.3888, "val_loss": 1.758567529296875}, {"step": 5000, "epoch": 14, "train_loss": 1.2127366065979004}, {"step": 5200, "epoch": 14, "train_loss": 1.4315799474716187}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.3938, "eval_loss": 1.75312724609375, "val_acc": 0.3938, "val_loss": 1.75312724609375}, {"step": 5400, "epoch": 15, "train_loss": 1.6419495344161987}, {"step": 5600, "epoch": 15, "train_loss": 1.4594193696975708}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.4016, "eval_loss": 1.729437646484375, "val_acc": 0.4016, "val_loss": 1.729437646484375}, {"step": 5800, "epoch": 16, "train_loss": 1.4091275930404663}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.4004, "eval_loss": 1.7404166259765625, "val_acc": 0.4004, "val_loss": 1.7404166259765625}, {"step": 6000, "epoch": 17, "train_loss": 1.320406198501587}, {"step": 6200, "epoch": 17, "train_loss": 1.3364851474761963}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.4044, "eval_loss": 1.743746240234375, "val_acc": 0.4044, "val_loss": 1.743746240234375}, {"step": 6400, "epoch": 18, "train_loss": 1.09772527217865}, {"step": 6600, "epoch": 18, "train_loss": 1.3545604944229126}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.4056, "eval_loss": 1.733443798828125, "val_acc": 0.4056, "val_loss": 1.733443798828125}, {"step": 6800, "epoch": 19, "train_loss": 1.2870097160339355}, {"step": 7000, "epoch": 19, "train_loss": 1.3358473777770996}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.4078, "eval_loss": 1.70913203125, "val_acc": 0.4078, "val_loss": 1.70913203125}, {"step": 7200, "epoch": 20, "train_loss": 1.4319428205490112}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.4054, "eval_loss": 1.7306862060546875, "val_acc": 0.4054, "val_loss": 1.7306862060546875}, {"step": 7400, "epoch": 21, "train_loss": 1.2578999996185303}, {"step": 7600, "epoch": 21, "train_loss": 1.2484898567199707}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.404, "eval_loss": 1.7344854736328126, "val_acc": 0.404, "val_loss": 1.7344854736328126}, {"step": 7800, "epoch": 22, "train_loss": 1.519671082496643}, {"step": 8000, "epoch": 22, "train_loss": 1.435581922531128}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.4064, "eval_loss": 1.707657080078125, "val_acc": 0.4064, "val_loss": 1.707657080078125}, {"step": 8200, "epoch": 23, "train_loss": 1.2397700548171997}, {"step": 8400, "epoch": 23, "train_loss": 1.4470657110214233}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.4064, "eval_loss": 1.728324169921875, "val_acc": 0.4064, "val_loss": 1.728324169921875}, {"step": 8600, "epoch": 24, "train_loss": 1.305275321006775}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.4058, "eval_loss": 1.7158028076171874, "val_acc": 0.4058, "val_loss": 1.7158028076171874}], "final": {"eval_split": "validation", "eval_acc": 0.4058, "eval_loss": 1.7158028076171874, "wall_s": 17.861095905303955, "val_acc": 0.4058, "val_loss": 1.7158028076171874}}
|
