1
|
{"args": {"mode": "dfa", "dataset": "cifar10", "depth": 5, "width": 64, "act": "tanh", "residual": 1, "epochs": 5, "batch_size": 128, "train_examples": 0, "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": 0.05, "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, "predictor_mode": "diagonal", "normalize_delta": 0, "settle_steps": 0, "kappa": 0.0, "feedback": "error", "seed": 2, "device": "cuda", "log_every": 500, "max_steps": 0, "probe_bs": 256, "outdir": "results", "tag": "scale_v3_cifar10_dfa_w64_d5_s2", "n_in": 3072, "n_out": 10}, "provenance": {"git_commit": "d4aef1a571c21e6e4e4125c90aa58da7f14e62e3", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.5040488243103027, "cos_r_negg": [0.021998973563313484, 0.04936715215444565, -0.03212035819888115, 0.020308537408709526, -0.019786003977060318], "cos_innovation_negg": [0.021998973563313484, 0.04936715215444565, -0.03212035819888115, 0.020308537408709526, -0.019786003977060318], "cos_apical_negg": [0.021998973563313484, 0.04936715215444565, -0.03212035819888115, 0.020308537408709526, -0.019786003977060318], "cos_Ac_negg": [0.021998973563313484, 0.04936715215444565, -0.03212035819888115, 0.020308537408709526, -0.019786003977060318], "r_norm": [2.4656858444213867, 2.6163930892944336, 2.554706573486328, 2.7663156986236572, 2.5314741134643555]}, {"epoch_end": 0, "step": 391, "test_acc": 0.3463, "test_loss": 1.8788817749023436}, {"step": 500, "epoch": 1, "train_loss": 1.830228567123413, "cos_r_negg": [0.5033991932868958, -0.00035106902942061424, -0.03273634612560272, -0.020603477954864502, 0.03820233419537544], "cos_innovation_negg": [0.5033991932868958, -0.00035106902942061424, -0.03273634612560272, -0.020603477954864502, 0.03820233419537544], "cos_apical_negg": [0.5033991932868958, -0.00035106902942061424, -0.03273634612560272, -0.020603477954864502, 0.03820233419537544], "cos_Ac_negg": [0.5033991932868958, -0.00035106902942061424, -0.03273634612560272, -0.020603477954864502, 0.03820233419537544], "r_norm": [2.1249561309814453, 2.261157989501953, 2.240158796310425, 2.406583786010742, 2.167628049850464]}, {"epoch_end": 1, "step": 782, "test_acc": 0.3479, "test_loss": 1.8727974487304688}, {"step": 1000, "epoch": 2, "train_loss": 1.855167269706726, "cos_r_negg": [0.5128034353256226, 0.1351730227470398, 0.10625163465738297, 0.0443875715136528, 0.0961105078458786], "cos_innovation_negg": [0.5128034353256226, 0.1351730227470398, 0.10625163465738297, 0.0443875715136528, 0.0961105078458786], "cos_apical_negg": [0.5128034353256226, 0.1351730227470398, 0.10625163465738297, 0.0443875715136528, 0.0961105078458786], "cos_Ac_negg": [0.5128034353256226, 0.1351730227470398, 0.10625163465738297, 0.0443875715136528, 0.0961105078458786], "r_norm": [2.0971968173980713, 2.2543911933898926, 2.231079339981079, 2.381404399871826, 2.154534339904785]}, {"epoch_end": 2, "step": 1173, "test_acc": 0.3395, "test_loss": 1.8641392211914063}, {"step": 1500, "epoch": 3, "train_loss": 1.7685799598693848, "cos_r_negg": [0.45721063017845154, 0.1838071197271347, 0.13023564219474792, 0.15230269730091095, 0.20305392146110535], "cos_innovation_negg": [0.45721063017845154, 0.1838071197271347, 0.13023564219474792, 0.15230269730091095, 0.20305392146110535], "cos_apical_negg": [0.45721063017845154, 0.1838071197271347, 0.13023564219474792, 0.15230269730091095, 0.20305392146110535], "cos_Ac_negg": [0.45721063017845154, 0.1838071197271347, 0.13023564219474792, 0.15230269730091095, 0.20305392146110535], "r_norm": [2.0680956840515137, 2.215471029281616, 2.1905603408813477, 2.327026128768921, 2.1160974502563477]}, {"epoch_end": 3, "step": 1564, "test_acc": 0.3494, "test_loss": 1.8513060791015625}, {"epoch_end": 4, "step": 1955, "test_acc": 0.3695, "test_loss": 1.7951519775390625}], "final": {"test_acc": 0.3695, "test_loss": 1.7951519775390625, "wall_s": 5.2740418910980225, "cos_r_negg": [0.5253897905349731, 0.20880332589149475, 0.16356134414672852, 0.21846379339694977, 0.19952347874641418], "cos_innovation_negg": [0.5253897905349731, 0.20880332589149475, 0.16356134414672852, 0.21846379339694977, 0.19952347874641418], "cos_apical_negg": [0.5253897905349731, 0.20880332589149475, 0.16356134414672852, 0.21846379339694977, 0.19952347874641418], "cos_Ac_negg": [0.5253897905349731, 0.20880332589149475, 0.16356134414672852, 0.21846379339694977, 0.19952347874641418]}}
|