1
|
{"args": {"mode": "bp", "dataset": "teacher", "depth": 1, "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_d1_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": 2.5311508178710938}, {"step": 200, "epoch": 0, "train_loss": 0.840945303440094}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.6182, "eval_loss": 0.9919363891601563, "val_acc": 0.6182, "val_loss": 0.9919363891601563}, {"step": 400, "epoch": 1, "train_loss": 0.733540415763855}, {"step": 600, "epoch": 1, "train_loss": 0.6970023512840271}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.6232, "eval_loss": 0.9807604614257812, "val_acc": 0.6232, "val_loss": 0.9807604614257812}, {"step": 800, "epoch": 2, "train_loss": 0.75431889295578}, {"step": 1000, "epoch": 2, "train_loss": 0.7388406991958618}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.6266, "eval_loss": 0.98731865234375, "val_acc": 0.6266, "val_loss": 0.98731865234375}, {"step": 1200, "epoch": 3, "train_loss": 0.9076403379440308}, {"step": 1400, "epoch": 3, "train_loss": 0.903896152973175}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.6198, "eval_loss": 1.0177142333984375, "val_acc": 0.6198, "val_loss": 1.0177142333984375}, {"step": 1600, "epoch": 4, "train_loss": 0.7635146975517273}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.6228, "eval_loss": 1.001775341796875, "val_acc": 0.6228, "val_loss": 1.001775341796875}, {"step": 1800, "epoch": 5, "train_loss": 0.7955314517021179}, {"step": 2000, "epoch": 5, "train_loss": 0.6706668734550476}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.6208, "eval_loss": 1.0371079345703125, "val_acc": 0.6208, "val_loss": 1.0371079345703125}, {"step": 2200, "epoch": 6, "train_loss": 0.9078772664070129}, {"step": 2400, "epoch": 6, "train_loss": 0.6533659100532532}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.6208, "eval_loss": 1.0158447021484376, "val_acc": 0.6208, "val_loss": 1.0158447021484376}, {"step": 2600, "epoch": 7, "train_loss": 0.7211366891860962}, {"step": 2800, "epoch": 7, "train_loss": 0.9030142426490784}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.6134, "eval_loss": 1.0586219482421875, "val_acc": 0.6134, "val_loss": 1.0586219482421875}, {"step": 3000, "epoch": 8, "train_loss": 0.7051634788513184}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.6198, "eval_loss": 1.05035625, "val_acc": 0.6198, "val_loss": 1.05035625}, {"step": 3200, "epoch": 9, "train_loss": 0.762702465057373}, {"step": 3400, "epoch": 9, "train_loss": 0.6346777081489563}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.6146, "eval_loss": 1.06519990234375, "val_acc": 0.6146, "val_loss": 1.06519990234375}, {"step": 3600, "epoch": 10, "train_loss": 0.6924629807472229}, {"step": 3800, "epoch": 10, "train_loss": 0.7348305583000183}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.6242, "eval_loss": 1.0611874755859374, "val_acc": 0.6242, "val_loss": 1.0611874755859374}, {"step": 4000, "epoch": 11, "train_loss": 0.7039048075675964}, {"step": 4200, "epoch": 11, "train_loss": 0.7568634748458862}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.6134, "eval_loss": 1.087744970703125, "val_acc": 0.6134, "val_loss": 1.087744970703125}, {"step": 4400, "epoch": 12, "train_loss": 0.6606151461601257}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.6156, "eval_loss": 1.0904285400390625, "val_acc": 0.6156, "val_loss": 1.0904285400390625}, {"step": 4600, "epoch": 13, "train_loss": 0.778768002986908}, {"step": 4800, "epoch": 13, "train_loss": 0.7155106067657471}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.6082, "eval_loss": 1.10935302734375, "val_acc": 0.6082, "val_loss": 1.10935302734375}, {"step": 5000, "epoch": 14, "train_loss": 0.7357832789421082}, {"step": 5200, "epoch": 14, "train_loss": 0.6756934523582458}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.6104, "eval_loss": 1.1120667236328126, "val_acc": 0.6104, "val_loss": 1.1120667236328126}, {"step": 5400, "epoch": 15, "train_loss": 0.7333395481109619}, {"step": 5600, "epoch": 15, "train_loss": 0.6376401782035828}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.6106, "eval_loss": 1.119081591796875, "val_acc": 0.6106, "val_loss": 1.119081591796875}, {"step": 5800, "epoch": 16, "train_loss": 0.6608338356018066}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.614, "eval_loss": 1.1165087646484375, "val_acc": 0.614, "val_loss": 1.1165087646484375}, {"step": 6000, "epoch": 17, "train_loss": 0.6254558563232422}, {"step": 6200, "epoch": 17, "train_loss": 0.6748548746109009}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.6048, "eval_loss": 1.14111865234375, "val_acc": 0.6048, "val_loss": 1.14111865234375}, {"step": 6400, "epoch": 18, "train_loss": 0.8265315294265747}, {"step": 6600, "epoch": 18, "train_loss": 0.6702396273612976}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.6044, "eval_loss": 1.1536258056640625, "val_acc": 0.6044, "val_loss": 1.1536258056640625}, {"step": 6800, "epoch": 19, "train_loss": 0.7251787781715393}, {"step": 7000, "epoch": 19, "train_loss": 0.5752092599868774}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.6072, "eval_loss": 1.14927880859375, "val_acc": 0.6072, "val_loss": 1.14927880859375}, {"step": 7200, "epoch": 20, "train_loss": 0.659776508808136}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.6094, "eval_loss": 1.1541574462890625, "val_acc": 0.6094, "val_loss": 1.1541574462890625}, {"step": 7400, "epoch": 21, "train_loss": 0.7678269743919373}, {"step": 7600, "epoch": 21, "train_loss": 0.6077974438667297}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.605, "eval_loss": 1.1698632568359375, "val_acc": 0.605, "val_loss": 1.1698632568359375}, {"step": 7800, "epoch": 22, "train_loss": 0.7145333290100098}, {"step": 8000, "epoch": 22, "train_loss": 0.4914296269416809}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.6026, "eval_loss": 1.1691294677734374, "val_acc": 0.6026, "val_loss": 1.1691294677734374}, {"step": 8200, "epoch": 23, "train_loss": 0.5936515927314758}, {"step": 8400, "epoch": 23, "train_loss": 0.6333053112030029}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.6008, "eval_loss": 1.1842491943359374, "val_acc": 0.6008, "val_loss": 1.1842491943359374}, {"step": 8600, "epoch": 24, "train_loss": 0.5887580513954163}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.6006, "eval_loss": 1.206232275390625, "val_acc": 0.6006, "val_loss": 1.206232275390625}], "final": {"eval_split": "validation", "eval_acc": 0.6006, "eval_loss": 1.206232275390625, "wall_s": 12.816138744354248, "val_acc": 0.6006, "val_loss": 1.206232275390625}}
|
