1
|
{"args": {"mode": "bp", "dataset": "teacher", "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": 8, "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_teacher_bp_w64_d2_t0_s0", "n_in": 128, "n_out": 10}, "split": {"dataset": "teacher", "task_seed": 0, "train_examples": 45000, "test_examples": 10000, "evaluation_split": "validation", "synthetic_generator": true, "split_seed": 2027, "validation_examples": 5000, "validation_index_sha256": "0ec30707a69784626a2d17128e5408bb8a8a8964e2ea753ef8e54b35ffb36355", "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": 2.5378835201263428}, {"step": 200, "epoch": 0, "train_loss": 0.8362271785736084}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.6174, "eval_loss": 0.9981978393554688, "val_acc": 0.6174, "val_loss": 0.9981978393554688}, {"step": 400, "epoch": 1, "train_loss": 0.7236090302467346}, {"step": 600, "epoch": 1, "train_loss": 0.732539713382721}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.6244, "eval_loss": 0.99426201171875, "val_acc": 0.6244, "val_loss": 0.99426201171875}, {"step": 800, "epoch": 2, "train_loss": 0.7037859559059143}, {"step": 1000, "epoch": 2, "train_loss": 0.7181667685508728}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.6252, "eval_loss": 0.9997278076171875, "val_acc": 0.6252, "val_loss": 0.9997278076171875}, {"step": 1200, "epoch": 3, "train_loss": 0.8484516143798828}, {"step": 1400, "epoch": 3, "train_loss": 0.9854567646980286}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.6178, "eval_loss": 1.0343426635742188, "val_acc": 0.6178, "val_loss": 1.0343426635742188}, {"step": 1600, "epoch": 4, "train_loss": 0.7158283591270447}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.6178, "eval_loss": 1.0126439208984375, "val_acc": 0.6178, "val_loss": 1.0126439208984375}, {"step": 1800, "epoch": 5, "train_loss": 0.7714849710464478}, {"step": 2000, "epoch": 5, "train_loss": 0.6831921339035034}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.6112, "eval_loss": 1.0534729370117188, "val_acc": 0.6112, "val_loss": 1.0534729370117188}, {"step": 2200, "epoch": 6, "train_loss": 0.8236176371574402}, {"step": 2400, "epoch": 6, "train_loss": 0.759265124797821}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.6262, "eval_loss": 1.0188918823242188, "val_acc": 0.6262, "val_loss": 1.0188918823242188}, {"step": 2600, "epoch": 7, "train_loss": 0.7377707362174988}, {"step": 2800, "epoch": 7, "train_loss": 0.9144394993782043}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.6066, "eval_loss": 1.08231669921875, "val_acc": 0.6066, "val_loss": 1.08231669921875}, {"step": 3000, "epoch": 8, "train_loss": 0.6569148302078247}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.6148, "eval_loss": 1.0635174560546874, "val_acc": 0.6148, "val_loss": 1.0635174560546874}, {"step": 3200, "epoch": 9, "train_loss": 0.602764904499054}, {"step": 3400, "epoch": 9, "train_loss": 0.569445013999939}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.604, "eval_loss": 1.086694580078125, "val_acc": 0.604, "val_loss": 1.086694580078125}, {"step": 3600, "epoch": 10, "train_loss": 0.64704430103302}, {"step": 3800, "epoch": 10, "train_loss": 0.778519332408905}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.6044, "eval_loss": 1.0840155883789062, "val_acc": 0.6044, "val_loss": 1.0840155883789062}, {"step": 4000, "epoch": 11, "train_loss": 0.6884896755218506}, {"step": 4200, "epoch": 11, "train_loss": 0.8264381885528564}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.6038, "eval_loss": 1.1155463134765624, "val_acc": 0.6038, "val_loss": 1.1155463134765624}, {"step": 4400, "epoch": 12, "train_loss": 0.6245763301849365}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.6066, "eval_loss": 1.111953564453125, "val_acc": 0.6066, "val_loss": 1.111953564453125}, {"step": 4600, "epoch": 13, "train_loss": 0.7497974634170532}, {"step": 4800, "epoch": 13, "train_loss": 0.6287288665771484}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.6014, "eval_loss": 1.157133251953125, "val_acc": 0.6014, "val_loss": 1.157133251953125}, {"step": 5000, "epoch": 14, "train_loss": 0.6910437345504761}, {"step": 5200, "epoch": 14, "train_loss": 0.6045576930046082}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.602, "eval_loss": 1.143605908203125, "val_acc": 0.602, "val_loss": 1.143605908203125}, {"step": 5400, "epoch": 15, "train_loss": 0.704504132270813}, {"step": 5600, "epoch": 15, "train_loss": 0.7717189788818359}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.5932, "eval_loss": 1.188294921875, "val_acc": 0.5932, "val_loss": 1.188294921875}, {"step": 5800, "epoch": 16, "train_loss": 0.6528366208076477}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.5992, "eval_loss": 1.1741950439453126, "val_acc": 0.5992, "val_loss": 1.1741950439453126}, {"step": 6000, "epoch": 17, "train_loss": 0.6485159397125244}, {"step": 6200, "epoch": 17, "train_loss": 0.582676887512207}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.5962, "eval_loss": 1.2025969482421874, "val_acc": 0.5962, "val_loss": 1.2025969482421874}, {"step": 6400, "epoch": 18, "train_loss": 0.7955219149589539}, {"step": 6600, "epoch": 18, "train_loss": 0.6003790497779846}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.5924, "eval_loss": 1.214786328125, "val_acc": 0.5924, "val_loss": 1.214786328125}, {"step": 6800, "epoch": 19, "train_loss": 0.8257665634155273}, {"step": 7000, "epoch": 19, "train_loss": 0.4760558009147644}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.5956, "eval_loss": 1.2264597900390626, "val_acc": 0.5956, "val_loss": 1.2264597900390626}, {"step": 7200, "epoch": 20, "train_loss": 0.5883042216300964}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.5892, "eval_loss": 1.2485189697265624, "val_acc": 0.5892, "val_loss": 1.2485189697265624}, {"step": 7400, "epoch": 21, "train_loss": 0.6486283540725708}, {"step": 7600, "epoch": 21, "train_loss": 0.551514744758606}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.5868, "eval_loss": 1.2586528076171875, "val_acc": 0.5868, "val_loss": 1.2586528076171875}, {"step": 7800, "epoch": 22, "train_loss": 0.6804865598678589}, {"step": 8000, "epoch": 22, "train_loss": 0.5737948417663574}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.5838, "eval_loss": 1.274739111328125, "val_acc": 0.5838, "val_loss": 1.274739111328125}, {"step": 8200, "epoch": 23, "train_loss": 0.4766664505004883}, {"step": 8400, "epoch": 23, "train_loss": 0.7162339091300964}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.5796, "eval_loss": 1.299013330078125, "val_acc": 0.5796, "val_loss": 1.299013330078125}, {"step": 8600, "epoch": 24, "train_loss": 0.48768115043640137}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.5794, "eval_loss": 1.30275439453125, "val_acc": 0.5794, "val_loss": 1.30275439453125}], "final": {"eval_split": "validation", "eval_acc": 0.5794, "eval_loss": 1.30275439453125, "wall_s": 18.107185125350952, "val_acc": 0.5794, "val_loss": 1.30275439453125}}
|
