1
|
{"args": {"mode": "bp", "dataset": "hierarchical", "depth": 8, "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_d8_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": 2.485805034637451}, {"step": 200, "epoch": 0, "train_loss": 1.7921242713928223}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.3256, "eval_loss": 1.9081274658203125, "val_acc": 0.3256, "val_loss": 1.9081274658203125}, {"step": 400, "epoch": 1, "train_loss": 1.7679376602172852}, {"step": 600, "epoch": 1, "train_loss": 1.7464025020599365}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.3376, "eval_loss": 1.87998232421875, "val_acc": 0.3376, "val_loss": 1.87998232421875}, {"step": 800, "epoch": 2, "train_loss": 1.685981273651123}, {"step": 1000, "epoch": 2, "train_loss": 1.8076378107070923}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.3488, "eval_loss": 1.8490821044921875, "val_acc": 0.3488, "val_loss": 1.8490821044921875}, {"step": 1200, "epoch": 3, "train_loss": 1.5900728702545166}, {"step": 1400, "epoch": 3, "train_loss": 1.6340383291244507}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.3554, "eval_loss": 1.8328413330078126, "val_acc": 0.3554, "val_loss": 1.8328413330078126}, {"step": 1600, "epoch": 4, "train_loss": 1.62771475315094}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.3726, "eval_loss": 1.809384033203125, "val_acc": 0.3726, "val_loss": 1.809384033203125}, {"step": 1800, "epoch": 5, "train_loss": 1.8734294176101685}, {"step": 2000, "epoch": 5, "train_loss": 1.6111594438552856}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.3706, "eval_loss": 1.7925618408203126, "val_acc": 0.3706, "val_loss": 1.7925618408203126}, {"step": 2200, "epoch": 6, "train_loss": 1.6069446802139282}, {"step": 2400, "epoch": 6, "train_loss": 1.6950554847717285}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.377, "eval_loss": 1.7837630126953126, "val_acc": 0.377, "val_loss": 1.7837630126953126}, {"step": 2600, "epoch": 7, "train_loss": 1.6443514823913574}, {"step": 2800, "epoch": 7, "train_loss": 1.6822401285171509}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.3752, "eval_loss": 1.7716834716796874, "val_acc": 0.3752, "val_loss": 1.7716834716796874}, {"step": 3000, "epoch": 8, "train_loss": 1.6433720588684082}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.3742, "eval_loss": 1.7764485595703126, "val_acc": 0.3742, "val_loss": 1.7764485595703126}, {"step": 3200, "epoch": 9, "train_loss": 1.6560735702514648}, {"step": 3400, "epoch": 9, "train_loss": 1.5867209434509277}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.3854, "eval_loss": 1.7663809326171875, "val_acc": 0.3854, "val_loss": 1.7663809326171875}, {"step": 3600, "epoch": 10, "train_loss": 1.5836684703826904}, {"step": 3800, "epoch": 10, "train_loss": 1.5297778844833374}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.383, "eval_loss": 1.764136328125, "val_acc": 0.383, "val_loss": 1.764136328125}, {"step": 4000, "epoch": 11, "train_loss": 1.6158984899520874}, {"step": 4200, "epoch": 11, "train_loss": 1.6285854578018188}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.3874, "eval_loss": 1.7495160400390626, "val_acc": 0.3874, "val_loss": 1.7495160400390626}, {"step": 4400, "epoch": 12, "train_loss": 1.7406468391418457}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.3754, "eval_loss": 1.7745019775390625, "val_acc": 0.3754, "val_loss": 1.7745019775390625}, {"step": 4600, "epoch": 13, "train_loss": 1.5345377922058105}, {"step": 4800, "epoch": 13, "train_loss": 1.6579797267913818}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.3772, "eval_loss": 1.766105712890625, "val_acc": 0.3772, "val_loss": 1.766105712890625}, {"step": 5000, "epoch": 14, "train_loss": 1.3781059980392456}, {"step": 5200, "epoch": 14, "train_loss": 1.6215590238571167}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.3758, "eval_loss": 1.768886669921875, "val_acc": 0.3758, "val_loss": 1.768886669921875}, {"step": 5400, "epoch": 15, "train_loss": 1.788153052330017}, {"step": 5600, "epoch": 15, "train_loss": 1.4719194173812866}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.379, "eval_loss": 1.7546478759765625, "val_acc": 0.379, "val_loss": 1.7546478759765625}, {"step": 5800, "epoch": 16, "train_loss": 1.490365982055664}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.3806, "eval_loss": 1.772960888671875, "val_acc": 0.3806, "val_loss": 1.772960888671875}, {"step": 6000, "epoch": 17, "train_loss": 1.494633674621582}, {"step": 6200, "epoch": 17, "train_loss": 1.4664727449417114}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.3838, "eval_loss": 1.76002763671875, "val_acc": 0.3838, "val_loss": 1.76002763671875}, {"step": 6400, "epoch": 18, "train_loss": 1.4253993034362793}, {"step": 6600, "epoch": 18, "train_loss": 1.581397294998169}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.3832, "eval_loss": 1.7663487548828125, "val_acc": 0.3832, "val_loss": 1.7663487548828125}, {"step": 6800, "epoch": 19, "train_loss": 1.5670100450515747}, {"step": 7000, "epoch": 19, "train_loss": 1.392325520515442}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.3876, "eval_loss": 1.746642041015625, "val_acc": 0.3876, "val_loss": 1.746642041015625}, {"step": 7200, "epoch": 20, "train_loss": 1.627132773399353}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.3826, "eval_loss": 1.755614697265625, "val_acc": 0.3826, "val_loss": 1.755614697265625}, {"step": 7400, "epoch": 21, "train_loss": 1.5662859678268433}, {"step": 7600, "epoch": 21, "train_loss": 1.5469049215316772}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.3868, "eval_loss": 1.76341796875, "val_acc": 0.3868, "val_loss": 1.76341796875}, {"step": 7800, "epoch": 22, "train_loss": 1.6311407089233398}, {"step": 8000, "epoch": 22, "train_loss": 1.5676263570785522}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.3848, "eval_loss": 1.7579181884765624, "val_acc": 0.3848, "val_loss": 1.7579181884765624}, {"step": 8200, "epoch": 23, "train_loss": 1.4638445377349854}, {"step": 8400, "epoch": 23, "train_loss": 1.6865875720977783}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.3832, "eval_loss": 1.7470204345703124, "val_acc": 0.3832, "val_loss": 1.7470204345703124}, {"step": 8600, "epoch": 24, "train_loss": 1.4909882545471191}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.3822, "eval_loss": 1.7505083984375, "val_acc": 0.3822, "val_loss": 1.7505083984375}], "final": {"eval_split": "validation", "eval_acc": 0.3822, "eval_loss": 1.7505083984375, "wall_s": 39.662893295288086, "val_acc": 0.3822, "val_loss": 1.7505083984375}}
|
