1
|
{"args": {"mode": "bp", "dataset": "teacher", "depth": 1, "width": 16, "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_w16_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": 2.393876552581787}, {"step": 200, "epoch": 0, "train_loss": 1.0103332996368408}, {"epoch_end": 0, "step": 352, "eval_split": "validation", "eval_acc": 0.5818, "eval_loss": 1.1113587158203124, "val_acc": 0.5818, "val_loss": 1.1113587158203124}, {"step": 400, "epoch": 1, "train_loss": 0.8591128587722778}, {"step": 600, "epoch": 1, "train_loss": 0.90489661693573}, {"epoch_end": 1, "step": 704, "eval_split": "validation", "eval_acc": 0.5902, "eval_loss": 1.09662138671875, "val_acc": 0.5902, "val_loss": 1.09662138671875}, {"step": 800, "epoch": 2, "train_loss": 0.8954740166664124}, {"step": 1000, "epoch": 2, "train_loss": 0.8610763549804688}, {"epoch_end": 2, "step": 1056, "eval_split": "validation", "eval_acc": 0.5886, "eval_loss": 1.1042962646484376, "val_acc": 0.5886, "val_loss": 1.1042962646484376}, {"step": 1200, "epoch": 3, "train_loss": 1.0534071922302246}, {"step": 1400, "epoch": 3, "train_loss": 1.0942528247833252}, {"epoch_end": 3, "step": 1408, "eval_split": "validation", "eval_acc": 0.5898, "eval_loss": 1.121703271484375, "val_acc": 0.5898, "val_loss": 1.121703271484375}, {"step": 1600, "epoch": 4, "train_loss": 0.8067715764045715}, {"epoch_end": 4, "step": 1760, "eval_split": "validation", "eval_acc": 0.5914, "eval_loss": 1.10259443359375, "val_acc": 0.5914, "val_loss": 1.10259443359375}, {"step": 1800, "epoch": 5, "train_loss": 1.033132553100586}, {"step": 2000, "epoch": 5, "train_loss": 0.803702712059021}, {"epoch_end": 5, "step": 2112, "eval_split": "validation", "eval_acc": 0.5864, "eval_loss": 1.117203564453125, "val_acc": 0.5864, "val_loss": 1.117203564453125}, {"step": 2200, "epoch": 6, "train_loss": 1.0454679727554321}, {"step": 2400, "epoch": 6, "train_loss": 0.865119993686676}, {"epoch_end": 6, "step": 2464, "eval_split": "validation", "eval_acc": 0.5908, "eval_loss": 1.096959912109375, "val_acc": 0.5908, "val_loss": 1.096959912109375}, {"step": 2600, "epoch": 7, "train_loss": 0.9102245569229126}, {"step": 2800, "epoch": 7, "train_loss": 0.9988207817077637}, {"epoch_end": 7, "step": 2816, "eval_split": "validation", "eval_acc": 0.5786, "eval_loss": 1.129971337890625, "val_acc": 0.5786, "val_loss": 1.129971337890625}, {"step": 3000, "epoch": 8, "train_loss": 0.9479234218597412}, {"epoch_end": 8, "step": 3168, "eval_split": "validation", "eval_acc": 0.5908, "eval_loss": 1.1039165283203125, "val_acc": 0.5908, "val_loss": 1.1039165283203125}, {"step": 3200, "epoch": 9, "train_loss": 0.8889789581298828}, {"step": 3400, "epoch": 9, "train_loss": 0.8190938234329224}, {"epoch_end": 9, "step": 3520, "eval_split": "validation", "eval_acc": 0.5846, "eval_loss": 1.108134228515625, "val_acc": 0.5846, "val_loss": 1.108134228515625}, {"step": 3600, "epoch": 10, "train_loss": 0.9284778833389282}, {"step": 3800, "epoch": 10, "train_loss": 0.9862697720527649}, {"epoch_end": 10, "step": 3872, "eval_split": "validation", "eval_acc": 0.5986, "eval_loss": 1.0914903686523438, "val_acc": 0.5986, "val_loss": 1.0914903686523438}, {"step": 4000, "epoch": 11, "train_loss": 1.0149346590042114}, {"step": 4200, "epoch": 11, "train_loss": 1.0448325872421265}, {"epoch_end": 11, "step": 4224, "eval_split": "validation", "eval_acc": 0.5942, "eval_loss": 1.102526171875, "val_acc": 0.5942, "val_loss": 1.102526171875}, {"step": 4400, "epoch": 12, "train_loss": 0.9033969044685364}, {"epoch_end": 12, "step": 4576, "eval_split": "validation", "eval_acc": 0.5926, "eval_loss": 1.097378271484375, "val_acc": 0.5926, "val_loss": 1.097378271484375}, {"step": 4600, "epoch": 13, "train_loss": 1.0780540704727173}, {"step": 4800, "epoch": 13, "train_loss": 0.9211264848709106}, {"epoch_end": 13, "step": 4928, "eval_split": "validation", "eval_acc": 0.5846, "eval_loss": 1.1193604248046876, "val_acc": 0.5846, "val_loss": 1.1193604248046876}, {"step": 5000, "epoch": 14, "train_loss": 0.9996097087860107}, {"step": 5200, "epoch": 14, "train_loss": 0.9626591801643372}, {"epoch_end": 14, "step": 5280, "eval_split": "validation", "eval_acc": 0.5884, "eval_loss": 1.099669091796875, "val_acc": 0.5884, "val_loss": 1.099669091796875}, {"step": 5400, "epoch": 15, "train_loss": 1.048741340637207}, {"step": 5600, "epoch": 15, "train_loss": 0.9683579802513123}, {"epoch_end": 15, "step": 5632, "eval_split": "validation", "eval_acc": 0.5924, "eval_loss": 1.10992041015625, "val_acc": 0.5924, "val_loss": 1.10992041015625}, {"step": 5800, "epoch": 16, "train_loss": 0.988154411315918}, {"epoch_end": 16, "step": 5984, "eval_split": "validation", "eval_acc": 0.5918, "eval_loss": 1.091592236328125, "val_acc": 0.5918, "val_loss": 1.091592236328125}, {"step": 6000, "epoch": 17, "train_loss": 0.9127556681632996}, {"step": 6200, "epoch": 17, "train_loss": 0.9000439643859863}, {"epoch_end": 17, "step": 6336, "eval_split": "validation", "eval_acc": 0.5818, "eval_loss": 1.1109916748046875, "val_acc": 0.5818, "val_loss": 1.1109916748046875}, {"step": 6400, "epoch": 18, "train_loss": 1.0868996381759644}, {"step": 6600, "epoch": 18, "train_loss": 0.9564498662948608}, {"epoch_end": 18, "step": 6688, "eval_split": "validation", "eval_acc": 0.5918, "eval_loss": 1.1101337646484375, "val_acc": 0.5918, "val_loss": 1.1101337646484375}, {"step": 6800, "epoch": 19, "train_loss": 1.1510719060897827}, {"step": 7000, "epoch": 19, "train_loss": 0.809857189655304}, {"epoch_end": 19, "step": 7040, "eval_split": "validation", "eval_acc": 0.5962, "eval_loss": 1.0926285400390625, "val_acc": 0.5962, "val_loss": 1.0926285400390625}, {"step": 7200, "epoch": 20, "train_loss": 0.8911164402961731}, {"epoch_end": 20, "step": 7392, "eval_split": "validation", "eval_acc": 0.5942, "eval_loss": 1.094940185546875, "val_acc": 0.5942, "val_loss": 1.094940185546875}, {"step": 7400, "epoch": 21, "train_loss": 1.1491795778274536}, {"step": 7600, "epoch": 21, "train_loss": 0.9027130603790283}, {"epoch_end": 21, "step": 7744, "eval_split": "validation", "eval_acc": 0.5924, "eval_loss": 1.1105342041015624, "val_acc": 0.5924, "val_loss": 1.1105342041015624}, {"step": 7800, "epoch": 22, "train_loss": 0.9469124674797058}, {"step": 8000, "epoch": 22, "train_loss": 0.9074727892875671}, {"epoch_end": 22, "step": 8096, "eval_split": "validation", "eval_acc": 0.5928, "eval_loss": 1.0956957763671875, "val_acc": 0.5928, "val_loss": 1.0956957763671875}, {"step": 8200, "epoch": 23, "train_loss": 0.8933815360069275}, {"step": 8400, "epoch": 23, "train_loss": 0.9923908710479736}, {"epoch_end": 23, "step": 8448, "eval_split": "validation", "eval_acc": 0.5894, "eval_loss": 1.1095331787109375, "val_acc": 0.5894, "val_loss": 1.1095331787109375}, {"step": 8600, "epoch": 24, "train_loss": 0.7992337942123413}, {"epoch_end": 24, "step": 8800, "eval_split": "validation", "eval_acc": 0.5918, "eval_loss": 1.10451015625, "val_acc": 0.5918, "val_loss": 1.10451015625}], "final": {"eval_split": "validation", "eval_acc": 0.5918, "eval_loss": 1.10451015625, "wall_s": 12.920711278915405, "val_acc": 0.5918, "val_loss": 1.10451015625}}
|
