1
|
{"args": {"mode": "bp", "dataset": "tentmap", "depth": 2, "width": 16, "act": "relu", "residual": 0, "epochs": 40, "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": 7, "task_n_in": 4, "task_classes": 10, "teacher_depth": 8, "teacher_width": 64, "teacher_residual": 1, "eta": 0.02, "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_tentmap_bp_w16_d2_t0_s0", "n_in": 4, "n_out": 2}, "split": {"dataset": "tentmap", "task_seed": 0, "train_examples": 45000, "test_examples": 10000, "evaluation_split": "validation", "synthetic_generator": true, "split_seed": 2027, "validation_examples": 5000, "validation_index_sha256": "291d95d8a591bae65a0250e7365655c3fc52324bdc34677026467ce13bf433b9", "validation_class_counts": {"0": 2500, "1": 2500}, "split_from_training_only": true}, "provenance": {"git_commit": "a0696a3d72cc5bf69876e1716b74f7e1ac525dd0", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 0.7234192490577698}, {"step": 200, "epoch": 0, "train_loss": 0.6983705163002014}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.4944, "eval_loss": 0.6948132934570312, "val_acc": 0.4944, "val_loss": 0.6948132934570312}, {"step": 400, "epoch": 1, "train_loss": 0.6936428546905518}, {"step": 600, "epoch": 1, "train_loss": 0.6950218677520752}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.5022, "eval_loss": 0.6933030883789062, "val_acc": 0.5022, "val_loss": 0.6933030883789062}, {"step": 800, "epoch": 2, "train_loss": 0.6947194337844849}, {"step": 1000, "epoch": 2, "train_loss": 0.6933380365371704}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.497, "eval_loss": 0.6936936645507813, "val_acc": 0.497, "val_loss": 0.6936936645507813}, {"step": 1200, "epoch": 3, "train_loss": 0.6922563314437866}, {"step": 1400, "epoch": 3, "train_loss": 0.6952409744262695}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.505, "eval_loss": 0.6932832397460937, "val_acc": 0.505, "val_loss": 0.6932832397460937}, {"step": 1600, "epoch": 4, "train_loss": 0.6934345364570618}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.4992, "eval_loss": 0.6934339965820312, "val_acc": 0.4992, "val_loss": 0.6934339965820312}, {"step": 1800, "epoch": 5, "train_loss": 0.6906495690345764}, {"step": 2000, "epoch": 5, "train_loss": 0.6929758191108704}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.4994, "eval_loss": 0.6934252075195313, "val_acc": 0.4994, "val_loss": 0.6934252075195313}, {"step": 2200, "epoch": 6, "train_loss": 0.6911870241165161}, {"step": 2400, "epoch": 6, "train_loss": 0.6939542889595032}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.4996, "eval_loss": 0.6934014404296875, "val_acc": 0.4996, "val_loss": 0.6934014404296875}, {"step": 2600, "epoch": 7, "train_loss": 0.6970890760421753}, {"step": 2800, "epoch": 7, "train_loss": 0.6932133436203003}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.4956, "eval_loss": 0.693291845703125, "val_acc": 0.4956, "val_loss": 0.693291845703125}, {"step": 3000, "epoch": 8, "train_loss": 0.6915990710258484}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.4962, "eval_loss": 0.6932441040039062, "val_acc": 0.4962, "val_loss": 0.6932441040039062}, {"step": 3200, "epoch": 9, "train_loss": 0.6991944909095764}, {"step": 3400, "epoch": 9, "train_loss": 0.695619523525238}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.4926, "eval_loss": 0.693361083984375, "val_acc": 0.4926, "val_loss": 0.693361083984375}, {"step": 3600, "epoch": 10, "train_loss": 0.69111567735672}, {"step": 3800, "epoch": 10, "train_loss": 0.6944504976272583}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.4962, "eval_loss": 0.693263232421875, "val_acc": 0.4962, "val_loss": 0.693263232421875}, {"step": 4000, "epoch": 11, "train_loss": 0.6895580887794495}, {"step": 4200, "epoch": 11, "train_loss": 0.6945664882659912}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.4988, "eval_loss": 0.6936599487304688, "val_acc": 0.4988, "val_loss": 0.6936599487304688}, {"step": 4400, "epoch": 12, "train_loss": 0.6921815276145935}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.4954, "eval_loss": 0.6932859375, "val_acc": 0.4954, "val_loss": 0.6932859375}, {"step": 4600, "epoch": 13, "train_loss": 0.6926528215408325}, {"step": 4800, "epoch": 13, "train_loss": 0.6943957805633545}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.5012, "eval_loss": 0.6933865234375, "val_acc": 0.5012, "val_loss": 0.6933865234375}, {"step": 5000, "epoch": 14, "train_loss": 0.6933137774467468}, {"step": 5200, "epoch": 14, "train_loss": 0.6908711194992065}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.5034, "eval_loss": 0.6933139770507812, "val_acc": 0.5034, "val_loss": 0.6933139770507812}, {"step": 5400, "epoch": 15, "train_loss": 0.6929924488067627}, {"step": 5600, "epoch": 15, "train_loss": 0.6907548904418945}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.4978, "eval_loss": 0.6932108764648437, "val_acc": 0.4978, "val_loss": 0.6932108764648437}, {"step": 5800, "epoch": 16, "train_loss": 0.689414381980896}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.4982, "eval_loss": 0.6935081420898438, "val_acc": 0.4982, "val_loss": 0.6935081420898438}, {"step": 6000, "epoch": 17, "train_loss": 0.692956268787384}, {"step": 6200, "epoch": 17, "train_loss": 0.6924958229064941}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.5006, "eval_loss": 0.6933629028320313, "val_acc": 0.5006, "val_loss": 0.6933629028320313}, {"step": 6400, "epoch": 18, "train_loss": 0.6946048140525818}, {"step": 6600, "epoch": 18, "train_loss": 0.6906641125679016}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.5006, "eval_loss": 0.6938782104492187, "val_acc": 0.5006, "val_loss": 0.6938782104492187}, {"step": 6800, "epoch": 19, "train_loss": 0.6922234892845154}, {"step": 7000, "epoch": 19, "train_loss": 0.6951857209205627}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.5024, "eval_loss": 0.693321923828125, "val_acc": 0.5024, "val_loss": 0.693321923828125}, {"step": 7200, "epoch": 20, "train_loss": 0.6930521130561829}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.4962, "eval_loss": 0.6934287841796875, "val_acc": 0.4962, "val_loss": 0.6934287841796875}, {"step": 7400, "epoch": 21, "train_loss": 0.693526029586792}, {"step": 7600, "epoch": 21, "train_loss": 0.6901332139968872}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.4986, "eval_loss": 0.693211572265625, "val_acc": 0.4986, "val_loss": 0.693211572265625}, {"step": 7800, "epoch": 22, "train_loss": 0.693686842918396}, {"step": 8000, "epoch": 22, "train_loss": 0.6924529671669006}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.5008, "eval_loss": 0.6936642700195312, "val_acc": 0.5008, "val_loss": 0.6936642700195312}, {"step": 8200, "epoch": 23, "train_loss": 0.6910799741744995}, {"step": 8400, "epoch": 23, "train_loss": 0.6932209730148315}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.503, "eval_loss": 0.6934892822265625, "val_acc": 0.503, "val_loss": 0.6934892822265625}, {"step": 8600, "epoch": 24, "train_loss": 0.6934233903884888}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.5002, "eval_loss": 0.6940283935546875, "val_acc": 0.5002, "val_loss": 0.6940283935546875}, {"step": 8800, "epoch": 25, "train_loss": 0.6923726797103882}, {"step": 9000, "epoch": 25, "train_loss": 0.6930972933769226}, {"epoch_end": 25, "step": 9152, "eval_split": "validation", "eval_acc": 0.5002, "eval_loss": 0.6940987915039063, "val_acc": 0.5002, "val_loss": 0.6940987915039063}, {"step": 9200, "epoch": 26, "train_loss": 0.6952126622200012}, {"step": 9400, "epoch": 26, "train_loss": 0.6929704546928406}, {"epoch_end": 26, "step": 9504, "eval_split": "validation", "eval_acc": 0.5034, "eval_loss": 0.69327421875, "val_acc": 0.5034, "val_loss": 0.69327421875}, {"step": 9600, "epoch": 27, "train_loss": 0.6948537826538086}, {"step": 9800, "epoch": 27, "train_loss": 0.6948608756065369}, {"epoch_end": 27, "step": 9856, "eval_split": "validation", "eval_acc": 0.4934, "eval_loss": 0.6933148315429688, "val_acc": 0.4934, "val_loss": 0.6933148315429688}, {"step": 10000, "epoch": 28, "train_loss": 0.6924528479576111}, {"step": 10200, "epoch": 28, "train_loss": 0.6937589049339294}, {"epoch_end": 28, "step": 10208, "eval_split": "validation", "eval_acc": 0.5062, "eval_loss": 0.6932617309570313, "val_acc": 0.5062, "val_loss": 0.6932617309570313}, {"step": 10400, "epoch": 29, "train_loss": 0.6898597478866577}, {"epoch_end": 29, "step": 10560, "eval_split": "validation", "eval_acc": 0.4998, "eval_loss": 0.6932134887695313, "val_acc": 0.4998, "val_loss": 0.6932134887695313}, {"step": 10600, "epoch": 30, "train_loss": 0.6981315016746521}, {"step": 10800, "epoch": 30, "train_loss": 0.702637791633606}, {"epoch_end": 30, "step": 10912, "eval_split": "validation", "eval_acc": 0.4996, "eval_loss": 0.6931874389648438, "val_acc": 0.4996, "val_loss": 0.6931874389648438}, {"step": 11000, "epoch": 31, "train_loss": 0.6947645545005798}, {"step": 11200, "epoch": 31, "train_loss": 0.6895925402641296}, {"epoch_end": 31, "step": 11264, "eval_split": "validation", "eval_acc": 0.4996, "eval_loss": 0.6947183959960938, "val_acc": 0.4996, "val_loss": 0.6947183959960938}, {"step": 11400, "epoch": 32, "train_loss": 0.6899244785308838}, {"step": 11600, "epoch": 32, "train_loss": 0.6939110159873962}, {"epoch_end": 32, "step": 11616, "eval_split": "validation", "eval_acc": 0.501, "eval_loss": 0.6935629638671875, "val_acc": 0.501, "val_loss": 0.6935629638671875}, {"step": 11800, "epoch": 33, "train_loss": 0.6895370483398438}, {"epoch_end": 33, "step": 11968, "eval_split": "validation", "eval_acc": 0.5004, "eval_loss": 0.69351015625, "val_acc": 0.5004, "val_loss": 0.69351015625}, {"step": 12000, "epoch": 34, "train_loss": 0.6940256357192993}, {"step": 12200, "epoch": 34, "train_loss": 0.6927611231803894}, {"epoch_end": 34, "step": 12320, "eval_split": "validation", "eval_acc": 0.5018, "eval_loss": 0.6933942260742187, "val_acc": 0.5018, "val_loss": 0.6933942260742187}, {"step": 12400, "epoch": 35, "train_loss": 0.6971215605735779}, {"step": 12600, "epoch": 35, "train_loss": 0.692359983921051}, {"epoch_end": 35, "step": 12672, "eval_split": "validation", "eval_acc": 0.502, "eval_loss": 0.6934452270507813, "val_acc": 0.502, "val_loss": 0.6934452270507813}, {"step": 12800, "epoch": 36, "train_loss": 0.6958345174789429}, {"step": 13000, "epoch": 36, "train_loss": 0.6922397017478943}, {"epoch_end": 36, "step": 13024, "eval_split": "validation", "eval_acc": 0.5024, "eval_loss": 0.6932959106445312, "val_acc": 0.5024, "val_loss": 0.6932959106445312}, {"step": 13200, "epoch": 37, "train_loss": 0.6923678517341614}, {"epoch_end": 37, "step": 13376, "eval_split": "validation", "eval_acc": 0.5024, "eval_loss": 0.6933904174804687, "val_acc": 0.5024, "val_loss": 0.6933904174804687}, {"step": 13400, "epoch": 38, "train_loss": 0.6921576261520386}, {"step": 13600, "epoch": 38, "train_loss": 0.6934102177619934}, {"epoch_end": 38, "step": 13728, "eval_split": "validation", "eval_acc": 0.504, "eval_loss": 0.6933526000976562, "val_acc": 0.504, "val_loss": 0.6933526000976562}, {"step": 13800, "epoch": 39, "train_loss": 0.6931068301200867}, {"step": 14000, "epoch": 39, "train_loss": 0.6963883638381958}, {"epoch_end": 39, "step": 14080, "eval_split": "validation", "eval_acc": 0.498, "eval_loss": 0.6932501098632813, "val_acc": 0.498, "val_loss": 0.6932501098632813}], "final": {"eval_split": "validation", "eval_acc": 0.498, "eval_loss": 0.6932501098632813, "wall_s": 26.63870906829834, "val_acc": 0.498, "val_loss": 0.6932501098632813}}
|
