1
|
{"args": {"mode": "bp", "dataset": "teacher", "depth": 8, "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_d8_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": 2.564910411834717}, {"step": 200, "epoch": 0, "train_loss": 0.9054244756698608}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.6184, "eval_loss": 1.0137853515625, "val_acc": 0.6184, "val_loss": 1.0137853515625}, {"step": 400, "epoch": 1, "train_loss": 0.7534605264663696}, {"step": 600, "epoch": 1, "train_loss": 0.7791372537612915}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.616, "eval_loss": 1.0175526611328125, "val_acc": 0.616, "val_loss": 1.0175526611328125}, {"step": 800, "epoch": 2, "train_loss": 0.7227439284324646}, {"step": 1000, "epoch": 2, "train_loss": 0.7281097769737244}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.6092, "eval_loss": 1.0258677368164062, "val_acc": 0.6092, "val_loss": 1.0258677368164062}, {"step": 1200, "epoch": 3, "train_loss": 0.8531402349472046}, {"step": 1400, "epoch": 3, "train_loss": 0.8890358209609985}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.6096, "eval_loss": 1.0535975341796875, "val_acc": 0.6096, "val_loss": 1.0535975341796875}, {"step": 1600, "epoch": 4, "train_loss": 0.6758511662483215}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.6186, "eval_loss": 1.0217825439453125, "val_acc": 0.6186, "val_loss": 1.0217825439453125}, {"step": 1800, "epoch": 5, "train_loss": 0.784747838973999}, {"step": 2000, "epoch": 5, "train_loss": 0.6705027222633362}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.608, "eval_loss": 1.0609456176757812, "val_acc": 0.608, "val_loss": 1.0609456176757812}, {"step": 2200, "epoch": 6, "train_loss": 0.7481552958488464}, {"step": 2400, "epoch": 6, "train_loss": 0.6858530044555664}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.6126, "eval_loss": 1.0318828491210938, "val_acc": 0.6126, "val_loss": 1.0318828491210938}, {"step": 2600, "epoch": 7, "train_loss": 0.7171128392219543}, {"step": 2800, "epoch": 7, "train_loss": 0.7544036507606506}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.6042, "eval_loss": 1.1035579833984375, "val_acc": 0.6042, "val_loss": 1.1035579833984375}, {"step": 3000, "epoch": 8, "train_loss": 0.7477214932441711}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.6062, "eval_loss": 1.0809581787109375, "val_acc": 0.6062, "val_loss": 1.0809581787109375}, {"step": 3200, "epoch": 9, "train_loss": 0.6675366163253784}, {"step": 3400, "epoch": 9, "train_loss": 0.5769951343536377}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.6004, "eval_loss": 1.1271369873046875, "val_acc": 0.6004, "val_loss": 1.1271369873046875}, {"step": 3600, "epoch": 10, "train_loss": 0.5943953990936279}, {"step": 3800, "epoch": 10, "train_loss": 0.6529951095581055}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.6064, "eval_loss": 1.1046400146484374, "val_acc": 0.6064, "val_loss": 1.1046400146484374}, {"step": 4000, "epoch": 11, "train_loss": 0.6365727186203003}, {"step": 4200, "epoch": 11, "train_loss": 0.733120858669281}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.601, "eval_loss": 1.1599071044921876, "val_acc": 0.601, "val_loss": 1.1599071044921876}, {"step": 4400, "epoch": 12, "train_loss": 0.6451032161712646}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.6016, "eval_loss": 1.1620322021484375, "val_acc": 0.6016, "val_loss": 1.1620322021484375}, {"step": 4600, "epoch": 13, "train_loss": 0.6921226978302002}, {"step": 4800, "epoch": 13, "train_loss": 0.5964938998222351}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.5942, "eval_loss": 1.1940201171875, "val_acc": 0.5942, "val_loss": 1.1940201171875}, {"step": 5000, "epoch": 14, "train_loss": 0.5242152214050293}, {"step": 5200, "epoch": 14, "train_loss": 0.6223666667938232}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.5888, "eval_loss": 1.2247827880859374, "val_acc": 0.5888, "val_loss": 1.2247827880859374}, {"step": 5400, "epoch": 15, "train_loss": 0.6306442022323608}, {"step": 5600, "epoch": 15, "train_loss": 0.5820850729942322}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.5844, "eval_loss": 1.2776964111328124, "val_acc": 0.5844, "val_loss": 1.2776964111328124}, {"step": 5800, "epoch": 16, "train_loss": 0.4614133834838867}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.593, "eval_loss": 1.2561759765625, "val_acc": 0.593, "val_loss": 1.2561759765625}, {"step": 6000, "epoch": 17, "train_loss": 0.6130412220954895}, {"step": 6200, "epoch": 17, "train_loss": 0.6101951003074646}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.5792, "eval_loss": 1.3334126953125, "val_acc": 0.5792, "val_loss": 1.3334126953125}, {"step": 6400, "epoch": 18, "train_loss": 0.5869726538658142}, {"step": 6600, "epoch": 18, "train_loss": 0.5814520716667175}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.5816, "eval_loss": 1.334586962890625, "val_acc": 0.5816, "val_loss": 1.334586962890625}, {"step": 6800, "epoch": 19, "train_loss": 0.5441330671310425}, {"step": 7000, "epoch": 19, "train_loss": 0.3414742052555084}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.5872, "eval_loss": 1.3782404296875, "val_acc": 0.5872, "val_loss": 1.3782404296875}, {"step": 7200, "epoch": 20, "train_loss": 0.44521811604499817}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.5802, "eval_loss": 1.4049444580078125, "val_acc": 0.5802, "val_loss": 1.4049444580078125}, {"step": 7400, "epoch": 21, "train_loss": 0.5824435353279114}, {"step": 7600, "epoch": 21, "train_loss": 0.37066158652305603}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.5776, "eval_loss": 1.4556610595703126, "val_acc": 0.5776, "val_loss": 1.4556610595703126}, {"step": 7800, "epoch": 22, "train_loss": 0.39698153734207153}, {"step": 8000, "epoch": 22, "train_loss": 0.3983050584793091}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.5746, "eval_loss": 1.49265849609375, "val_acc": 0.5746, "val_loss": 1.49265849609375}, {"step": 8200, "epoch": 23, "train_loss": 0.32650813460350037}, {"step": 8400, "epoch": 23, "train_loss": 0.48107656836509705}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.5676, "eval_loss": 1.5484889892578124, "val_acc": 0.5676, "val_loss": 1.5484889892578124}, {"step": 8600, "epoch": 24, "train_loss": 0.3155350387096405}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.5674, "eval_loss": 1.574935498046875, "val_acc": 0.5674, "val_loss": 1.574935498046875}], "final": {"eval_split": "validation", "eval_acc": 0.5674, "eval_loss": 1.574935498046875, "wall_s": 43.34907793998718, "val_acc": 0.5674, "val_loss": 1.574935498046875}}
|
