1
|
{"args": {"mode": "bp", "dataset": "tentmap", "depth": 6, "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_d6_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": 0.6928590536117554}, {"step": 200, "epoch": 0, "train_loss": 0.6938450336456299}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.4968, "eval_loss": 0.6932424926757812, "val_acc": 0.4968, "val_loss": 0.6932424926757812}, {"step": 400, "epoch": 1, "train_loss": 0.6921735405921936}, {"step": 600, "epoch": 1, "train_loss": 0.6930134296417236}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.497, "eval_loss": 0.6932382446289063, "val_acc": 0.497, "val_loss": 0.6932382446289063}, {"step": 800, "epoch": 2, "train_loss": 0.6925055980682373}, {"step": 1000, "epoch": 2, "train_loss": 0.6947975158691406}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.5044, "eval_loss": 0.693189306640625, "val_acc": 0.5044, "val_loss": 0.693189306640625}, {"step": 1200, "epoch": 3, "train_loss": 0.6924083828926086}, {"step": 1400, "epoch": 3, "train_loss": 0.6937776803970337}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.6933202026367188, "val_acc": 0.5, "val_loss": 0.6933202026367188}, {"step": 1600, "epoch": 4, "train_loss": 0.6917003393173218}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.6933621948242188, "val_acc": 0.5, "val_loss": 0.6933621948242188}, {"step": 1800, "epoch": 5, "train_loss": 0.6928338408470154}, {"step": 2000, "epoch": 5, "train_loss": 0.6931189894676208}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.5008, "eval_loss": 0.6933397338867188, "val_acc": 0.5008, "val_loss": 0.6933397338867188}, {"step": 2200, "epoch": 6, "train_loss": 0.6914609670639038}, {"step": 2400, "epoch": 6, "train_loss": 0.6944755911827087}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.4936, "eval_loss": 0.6932679321289063, "val_acc": 0.4936, "val_loss": 0.6932679321289063}, {"step": 2600, "epoch": 7, "train_loss": 0.6937891840934753}, {"step": 2800, "epoch": 7, "train_loss": 0.6930403709411621}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.5002, "eval_loss": 0.6932306396484375, "val_acc": 0.5002, "val_loss": 0.6932306396484375}, {"step": 3000, "epoch": 8, "train_loss": 0.6917064189910889}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.4996, "eval_loss": 0.6932857543945312, "val_acc": 0.4996, "val_loss": 0.6932857543945312}, {"step": 3200, "epoch": 9, "train_loss": 0.6983595490455627}, {"step": 3400, "epoch": 9, "train_loss": 0.6949564218521118}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.69331455078125, "val_acc": 0.5, "val_loss": 0.69331455078125}, {"step": 3600, "epoch": 10, "train_loss": 0.6908610463142395}, {"step": 3800, "epoch": 10, "train_loss": 0.6944295167922974}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.5006, "eval_loss": 0.6932333740234375, "val_acc": 0.5006, "val_loss": 0.6932333740234375}, {"step": 4000, "epoch": 11, "train_loss": 0.6910235285758972}, {"step": 4200, "epoch": 11, "train_loss": 0.6929867267608643}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.5008, "eval_loss": 0.6934860107421875, "val_acc": 0.5008, "val_loss": 0.6934860107421875}, {"step": 4400, "epoch": 12, "train_loss": 0.6926133036613464}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.693273583984375, "val_acc": 0.5, "val_loss": 0.693273583984375}, {"step": 4600, "epoch": 13, "train_loss": 0.6919313073158264}, {"step": 4800, "epoch": 13, "train_loss": 0.6933034658432007}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.4998, "eval_loss": 0.6933156372070313, "val_acc": 0.4998, "val_loss": 0.6933156372070313}, {"step": 5000, "epoch": 14, "train_loss": 0.6936420202255249}, {"step": 5200, "epoch": 14, "train_loss": 0.6913618445396423}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.69323125, "val_acc": 0.5, "val_loss": 0.69323125}, {"step": 5400, "epoch": 15, "train_loss": 0.6928245425224304}, {"step": 5600, "epoch": 15, "train_loss": 0.6929822564125061}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.69324013671875, "val_acc": 0.5, "val_loss": 0.69324013671875}, {"step": 5800, "epoch": 16, "train_loss": 0.6921864151954651}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.5018, "eval_loss": 0.693311328125, "val_acc": 0.5018, "val_loss": 0.693311328125}, {"step": 6000, "epoch": 17, "train_loss": 0.694209098815918}, {"step": 6200, "epoch": 17, "train_loss": 0.6931802034378052}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.693387939453125, "val_acc": 0.5, "val_loss": 0.693387939453125}, {"step": 6400, "epoch": 18, "train_loss": 0.6933950781822205}, {"step": 6600, "epoch": 18, "train_loss": 0.6909139752388}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.502, "eval_loss": 0.6936298583984375, "val_acc": 0.502, "val_loss": 0.6936298583984375}, {"step": 6800, "epoch": 19, "train_loss": 0.6926027536392212}, {"step": 7000, "epoch": 19, "train_loss": 0.6945126056671143}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.4988, "eval_loss": 0.6933175048828125, "val_acc": 0.4988, "val_loss": 0.6933175048828125}, {"step": 7200, "epoch": 20, "train_loss": 0.6935617327690125}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.4936, "eval_loss": 0.6933082763671875, "val_acc": 0.4936, "val_loss": 0.6933082763671875}, {"step": 7400, "epoch": 21, "train_loss": 0.6923162341117859}, {"step": 7600, "epoch": 21, "train_loss": 0.6906459331512451}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.5002, "eval_loss": 0.69322353515625, "val_acc": 0.5002, "val_loss": 0.69322353515625}, {"step": 7800, "epoch": 22, "train_loss": 0.6934617757797241}, {"step": 8000, "epoch": 22, "train_loss": 0.6912260055541992}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.4954, "eval_loss": 0.6938241821289063, "val_acc": 0.4954, "val_loss": 0.6938241821289063}, {"step": 8200, "epoch": 23, "train_loss": 0.6896759867668152}, {"step": 8400, "epoch": 23, "train_loss": 0.6918059587478638}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.5004, "eval_loss": 0.6933515869140625, "val_acc": 0.5004, "val_loss": 0.6933515869140625}, {"step": 8600, "epoch": 24, "train_loss": 0.6928461790084839}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.6941350341796875, "val_acc": 0.5, "val_loss": 0.6941350341796875}, {"step": 8800, "epoch": 25, "train_loss": 0.6902450323104858}, {"step": 9000, "epoch": 25, "train_loss": 0.6906232237815857}, {"epoch_end": 25, "step": 9152, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.6939969604492188, "val_acc": 0.5, "val_loss": 0.6939969604492188}, {"step": 9200, "epoch": 26, "train_loss": 0.6954138875007629}, {"step": 9400, "epoch": 26, "train_loss": 0.6905629634857178}, {"epoch_end": 26, "step": 9504, "eval_split": "validation", "eval_acc": 0.5052, "eval_loss": 0.693370068359375, "val_acc": 0.5052, "val_loss": 0.693370068359375}, {"step": 9600, "epoch": 27, "train_loss": 0.6932179927825928}, {"step": 9800, "epoch": 27, "train_loss": 0.6923965215682983}, {"epoch_end": 27, "step": 9856, "eval_split": "validation", "eval_acc": 0.4958, "eval_loss": 0.693275634765625, "val_acc": 0.4958, "val_loss": 0.693275634765625}, {"step": 10000, "epoch": 28, "train_loss": 0.6931954622268677}, {"step": 10200, "epoch": 28, "train_loss": 0.692685067653656}, {"epoch_end": 28, "step": 10208, "eval_split": "validation", "eval_acc": 0.4962, "eval_loss": 0.6933490478515625, "val_acc": 0.4962, "val_loss": 0.6933490478515625}, {"step": 10400, "epoch": 29, "train_loss": 0.6921876668930054}, {"epoch_end": 29, "step": 10560, "eval_split": "validation", "eval_acc": 0.4958, "eval_loss": 0.6932742919921875, "val_acc": 0.4958, "val_loss": 0.6932742919921875}, {"step": 10600, "epoch": 30, "train_loss": 0.6979708075523376}, {"step": 10800, "epoch": 30, "train_loss": 0.7037988305091858}, {"epoch_end": 30, "step": 10912, "eval_split": "validation", "eval_acc": 0.4926, "eval_loss": 0.693386962890625, "val_acc": 0.4926, "val_loss": 0.693386962890625}, {"step": 11000, "epoch": 31, "train_loss": 0.6972269415855408}, {"step": 11200, "epoch": 31, "train_loss": 0.6931207180023193}, {"epoch_end": 31, "step": 11264, "eval_split": "validation", "eval_acc": 0.5, "eval_loss": 0.6945745361328125, "val_acc": 0.5, "val_loss": 0.6945745361328125}, {"step": 11400, "epoch": 32, "train_loss": 0.6919587254524231}, {"step": 11600, "epoch": 32, "train_loss": 0.693355143070221}, {"epoch_end": 32, "step": 11616, "eval_split": "validation", "eval_acc": 0.5008, "eval_loss": 0.6936034790039063, "val_acc": 0.5008, "val_loss": 0.6936034790039063}, {"step": 11800, "epoch": 33, "train_loss": 0.68900066614151}, {"epoch_end": 33, "step": 11968, "eval_split": "validation", "eval_acc": 0.5018, "eval_loss": 0.693598974609375, "val_acc": 0.5018, "val_loss": 0.693598974609375}, {"step": 12000, "epoch": 34, "train_loss": 0.6928737759590149}, {"step": 12200, "epoch": 34, "train_loss": 0.6938790082931519}, {"epoch_end": 34, "step": 12320, "eval_split": "validation", "eval_acc": 0.5016, "eval_loss": 0.6934698364257813, "val_acc": 0.5016, "val_loss": 0.6934698364257813}, {"step": 12400, "epoch": 35, "train_loss": 0.6947739720344543}, {"step": 12600, "epoch": 35, "train_loss": 0.6918246746063232}, {"epoch_end": 35, "step": 12672, "eval_split": "validation", "eval_acc": 0.4996, "eval_loss": 0.6935453735351562, "val_acc": 0.4996, "val_loss": 0.6935453735351562}, {"step": 12800, "epoch": 36, "train_loss": 0.6953124403953552}, {"step": 13000, "epoch": 36, "train_loss": 0.6913015246391296}, {"epoch_end": 36, "step": 13024, "eval_split": "validation", "eval_acc": 0.4946, "eval_loss": 0.6934827392578125, "val_acc": 0.4946, "val_loss": 0.6934827392578125}, {"step": 13200, "epoch": 37, "train_loss": 0.6909523606300354}, {"epoch_end": 37, "step": 13376, "eval_split": "validation", "eval_acc": 0.5038, "eval_loss": 0.6933150756835937, "val_acc": 0.5038, "val_loss": 0.6933150756835937}, {"step": 13400, "epoch": 38, "train_loss": 0.6924707889556885}, {"step": 13600, "epoch": 38, "train_loss": 0.6919642686843872}, {"epoch_end": 38, "step": 13728, "eval_split": "validation", "eval_acc": 0.502, "eval_loss": 0.6932087036132812, "val_acc": 0.502, "val_loss": 0.6932087036132812}, {"step": 13800, "epoch": 39, "train_loss": 0.6929119825363159}, {"step": 14000, "epoch": 39, "train_loss": 0.695218026638031}, {"epoch_end": 39, "step": 14080, "eval_split": "validation", "eval_acc": 0.5004, "eval_loss": 0.6934271606445312, "val_acc": 0.5004, "val_loss": 0.6934271606445312}], "final": {"eval_split": "validation", "eval_acc": 0.5004, "eval_loss": 0.6934271606445312, "wall_s": 46.88235116004944, "val_acc": 0.5004, "val_loss": 0.6934271606445312}}
|
