1
|
{"args": {"mode": "bp", "dataset": "teacher", "depth": 4, "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_d4_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": 2.4882280826568604}, {"step": 200, "epoch": 0, "train_loss": 0.9054072499275208}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.6162, "eval_loss": 1.0237540161132812, "val_acc": 0.6162, "val_loss": 1.0237540161132812}, {"step": 400, "epoch": 1, "train_loss": 0.7456945180892944}, {"step": 600, "epoch": 1, "train_loss": 0.7214746475219727}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.6178, "eval_loss": 1.0069253051757812, "val_acc": 0.6178, "val_loss": 1.0069253051757812}, {"step": 800, "epoch": 2, "train_loss": 0.7784503698348999}, {"step": 1000, "epoch": 2, "train_loss": 0.7319915890693665}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.6192, "eval_loss": 1.0175387329101562, "val_acc": 0.6192, "val_loss": 1.0175387329101562}, {"step": 1200, "epoch": 3, "train_loss": 0.8478718400001526}, {"step": 1400, "epoch": 3, "train_loss": 0.9292729496955872}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.6124, "eval_loss": 1.0385492553710938, "val_acc": 0.6124, "val_loss": 1.0385492553710938}, {"step": 1600, "epoch": 4, "train_loss": 0.6804534792900085}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.6198, "eval_loss": 1.0237674072265626, "val_acc": 0.6198, "val_loss": 1.0237674072265626}, {"step": 1800, "epoch": 5, "train_loss": 0.7726636528968811}, {"step": 2000, "epoch": 5, "train_loss": 0.6768437027931213}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.6182, "eval_loss": 1.0602895385742188, "val_acc": 0.6182, "val_loss": 1.0602895385742188}, {"step": 2200, "epoch": 6, "train_loss": 0.7809285521507263}, {"step": 2400, "epoch": 6, "train_loss": 0.7038103342056274}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.6186, "eval_loss": 1.0337721313476562, "val_acc": 0.6186, "val_loss": 1.0337721313476562}, {"step": 2600, "epoch": 7, "train_loss": 0.7057100534439087}, {"step": 2800, "epoch": 7, "train_loss": 0.891932487487793}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.6074, "eval_loss": 1.0968569091796876, "val_acc": 0.6074, "val_loss": 1.0968569091796876}, {"step": 3000, "epoch": 8, "train_loss": 0.7323035597801208}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.6156, "eval_loss": 1.0815179443359375, "val_acc": 0.6156, "val_loss": 1.0815179443359375}, {"step": 3200, "epoch": 9, "train_loss": 0.6907006502151489}, {"step": 3400, "epoch": 9, "train_loss": 0.7368860244750977}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.6042, "eval_loss": 1.1099479736328124, "val_acc": 0.6042, "val_loss": 1.1099479736328124}, {"step": 3600, "epoch": 10, "train_loss": 0.6811631321907043}, {"step": 3800, "epoch": 10, "train_loss": 0.6820284724235535}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.6106, "eval_loss": 1.102748095703125, "val_acc": 0.6106, "val_loss": 1.102748095703125}, {"step": 4000, "epoch": 11, "train_loss": 0.6430980563163757}, {"step": 4200, "epoch": 11, "train_loss": 0.7353668212890625}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.5954, "eval_loss": 1.1535440673828126, "val_acc": 0.5954, "val_loss": 1.1535440673828126}, {"step": 4400, "epoch": 12, "train_loss": 0.6156095266342163}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.599, "eval_loss": 1.14924482421875, "val_acc": 0.599, "val_loss": 1.14924482421875}, {"step": 4600, "epoch": 13, "train_loss": 0.696312427520752}, {"step": 4800, "epoch": 13, "train_loss": 0.6119407415390015}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.5936, "eval_loss": 1.1916462890625, "val_acc": 0.5936, "val_loss": 1.1916462890625}, {"step": 5000, "epoch": 14, "train_loss": 0.6664527058601379}, {"step": 5200, "epoch": 14, "train_loss": 0.5013567209243774}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.5924, "eval_loss": 1.209540234375, "val_acc": 0.5924, "val_loss": 1.209540234375}, {"step": 5400, "epoch": 15, "train_loss": 0.653146505355835}, {"step": 5600, "epoch": 15, "train_loss": 0.5721628665924072}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.5892, "eval_loss": 1.232269873046875, "val_acc": 0.5892, "val_loss": 1.232269873046875}, {"step": 5800, "epoch": 16, "train_loss": 0.604249119758606}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.5996, "eval_loss": 1.2246965087890624, "val_acc": 0.5996, "val_loss": 1.2246965087890624}, {"step": 6000, "epoch": 17, "train_loss": 0.6022902131080627}, {"step": 6200, "epoch": 17, "train_loss": 0.6290896534919739}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.5898, "eval_loss": 1.2786100830078124, "val_acc": 0.5898, "val_loss": 1.2786100830078124}, {"step": 6400, "epoch": 18, "train_loss": 0.5739793181419373}, {"step": 6600, "epoch": 18, "train_loss": 0.5909854769706726}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.5836, "eval_loss": 1.2779569091796874, "val_acc": 0.5836, "val_loss": 1.2779569091796874}, {"step": 6800, "epoch": 19, "train_loss": 0.5921121835708618}, {"step": 7000, "epoch": 19, "train_loss": 0.4734802544116974}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.585, "eval_loss": 1.2972416015625, "val_acc": 0.585, "val_loss": 1.2972416015625}, {"step": 7200, "epoch": 20, "train_loss": 0.5883579254150391}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.5826, "eval_loss": 1.3348498779296876, "val_acc": 0.5826, "val_loss": 1.3348498779296876}, {"step": 7400, "epoch": 21, "train_loss": 0.5489111542701721}, {"step": 7600, "epoch": 21, "train_loss": 0.5122550129890442}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.5814, "eval_loss": 1.3466891845703124, "val_acc": 0.5814, "val_loss": 1.3466891845703124}, {"step": 7800, "epoch": 22, "train_loss": 0.5395210981369019}, {"step": 8000, "epoch": 22, "train_loss": 0.5577489137649536}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.5816, "eval_loss": 1.371901025390625, "val_acc": 0.5816, "val_loss": 1.371901025390625}, {"step": 8200, "epoch": 23, "train_loss": 0.4875646233558655}, {"step": 8400, "epoch": 23, "train_loss": 0.6638826727867126}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.5766, "eval_loss": 1.4083632568359374, "val_acc": 0.5766, "val_loss": 1.4083632568359374}, {"step": 8600, "epoch": 24, "train_loss": 0.3843157887458801}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.5738, "eval_loss": 1.431485693359375, "val_acc": 0.5738, "val_loss": 1.431485693359375}], "final": {"eval_split": "validation", "eval_acc": 0.5738, "eval_loss": 1.431485693359375, "wall_s": 26.39419412612915, "val_acc": 0.5738, "val_loss": 1.431485693359375}}
|
