1
|
{"args": {"mode": "sdil", "dataset": "cifar10", "depth": 10, "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": 0.02, "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": 1, "device": "cuda", "log_every": 500, "max_steps": 0, "probe_bs": 256, "outdir": "results", "tag": "scale_v3_cifar10_sdil_w64_d10_s1", "n_in": 3072, "n_out": 10}, "provenance": {"git_commit": "d4aef1a571c21e6e4e4125c90aa58da7f14e62e3", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.5440011024475098, "cos_r_negg": [0.02870703488588333, 0.06334742158651352, 0.03597518429160118, 0.03449501097202301, 0.002455838257446885, -0.04093361645936966, 0.024038055911660194, 0.045385219156742096, -0.002375698182731867, 0.016220565885305405], "cos_innovation_negg": [0.02870703488588333, 0.06334742158651352, 0.03597518429160118, 0.03449501097202301, 0.002455838257446885, -0.04093361645936966, 0.024038055911660194, 0.045385219156742096, -0.002375698182731867, 0.016220565885305405], "cos_apical_negg": [0.02870703488588333, 0.06334742158651352, 0.03597518429160118, 0.03449501097202301, 0.002455838257446885, -0.04093361645936966, 0.024038055911660194, 0.045385219156742096, -0.002375698182731867, 0.016220565885305405], "cos_Ac_negg": [0.02870703488588333, 0.06334742158651352, 0.03597518429160118, 0.03449501097202301, 0.002455838257446885, -0.04093361645936966, 0.024038055911660194, 0.045385219156742096, -0.002375698182731867, 0.016220565885305405], "r_norm": [2.4114789962768555, 2.429633617401123, 2.4504361152648926, 2.4734389781951904, 2.516995429992676, 2.341196060180664, 2.2977757453918457, 2.5341696739196777, 2.4043684005737305, 2.4307403564453125], "ldr": {"dL_sdil": 0.022606611251831055, "dL_bp": -0.36179614067077637, "ratio": -0.06248439026994042}}, {"epoch_end": 0, "step": 391, "test_acc": 0.3679, "test_loss": 1.8025805908203125}, {"step": 500, "epoch": 1, "train_loss": 1.6827623844146729, "cos_r_negg": [0.6476483941078186, 0.08318375051021576, 0.07068692892789841, 0.1337260752916336, 0.05305585637688637, 0.07424556463956833, 0.10099766403436661, 0.07142257690429688, 0.13083171844482422, 0.058422595262527466], "cos_innovation_negg": [0.6476483941078186, 0.08318375051021576, 0.07068692892789841, 0.1337260752916336, 0.05305585637688637, 0.07424556463956833, 0.10099766403436661, 0.07142257690429688, 0.13083171844482422, 0.058422595262527466], "cos_apical_negg": [0.6476483941078186, 0.08318375051021576, 0.07068692892789841, 0.1337260752916336, 0.05305585637688637, 0.07424556463956833, 0.10099766403436661, 0.07142257690429688, 0.13083171844482422, 0.058422595262527466], "cos_Ac_negg": [0.6476483941078186, 0.08318375051021576, 0.07068692892789841, 0.1337260752916336, 0.05305585637688637, 0.07424556463956833, 0.10099766403436661, 0.07142257690429688, 0.13083171844482422, 0.058422595262527466], "r_norm": [1.7465323209762573, 1.7290608882904053, 1.732311725616455, 1.736166000366211, 1.7904763221740723, 1.652977466583252, 1.620896816253662, 1.7622694969177246, 1.697360634803772, 1.7120652198791504]}, {"epoch_end": 1, "step": 782, "test_acc": 0.3852, "test_loss": 1.75346044921875}, {"step": 1000, "epoch": 2, "train_loss": 1.766517162322998, "cos_r_negg": [0.6951239109039307, 0.15096542239189148, 0.15789170563220978, 0.1574188768863678, 0.17342738807201385, 0.17873521149158478, 0.16676995158195496, 0.14437828958034515, 0.1967092901468277, 0.13212063908576965], "cos_innovation_negg": [0.6951239109039307, 0.15096542239189148, 0.15789170563220978, 0.1574188768863678, 0.17342738807201385, 0.17873521149158478, 0.16676995158195496, 0.14437828958034515, 0.1967092901468277, 0.13212063908576965], "cos_apical_negg": [0.6951239109039307, 0.15096542239189148, 0.15789170563220978, 0.1574188768863678, 0.17342738807201385, 0.17873521149158478, 0.16676995158195496, 0.14437828958034515, 0.1967092901468277, 0.13212063908576965], "cos_Ac_negg": [0.6951239109039307, 0.15096542239189148, 0.15789170563220978, 0.1574188768863678, 0.17342738807201385, 0.17873521149158478, 0.16676995158195496, 0.14437828958034515, 0.1967092901468277, 0.13212063908576965], "r_norm": [1.541092872619629, 1.4447054862976074, 1.4482648372650146, 1.4496841430664062, 1.4926507472991943, 1.3815407752990723, 1.3542249202728271, 1.4730658531188965, 1.4242453575134277, 1.4288713932037354]}, {"epoch_end": 2, "step": 1173, "test_acc": 0.4018, "test_loss": 1.7274202758789063}, {"step": 1500, "epoch": 3, "train_loss": 1.6679010391235352, "cos_r_negg": [0.7612735033035278, 0.28905171155929565, 0.25119340419769287, 0.2751164138317108, 0.23994939029216766, 0.2606850266456604, 0.25932133197784424, 0.24340254068374634, 0.2804219722747803, 0.24803128838539124], "cos_innovation_negg": [0.7612735033035278, 0.28905171155929565, 0.25119340419769287, 0.2751164138317108, 0.23994939029216766, 0.2606850266456604, 0.25932133197784424, 0.24340254068374634, 0.2804219722747803, 0.24803128838539124], "cos_apical_negg": [0.7612735033035278, 0.28905171155929565, 0.25119340419769287, 0.2751164138317108, 0.23994939029216766, 0.2606850266456604, 0.25932133197784424, 0.24340254068374634, 0.2804219722747803, 0.24803128838539124], "cos_Ac_negg": [0.7612735033035278, 0.28905171155929565, 0.25119340419769287, 0.2751164138317108, 0.23994939029216766, 0.2606850266456604, 0.25932133197784424, 0.24340254068374634, 0.2804219722747803, 0.24803128838539124], "r_norm": [1.3115556240081787, 1.1616687774658203, 1.1646950244903564, 1.1733033657073975, 1.1924247741699219, 1.111291766166687, 1.0859054327011108, 1.180570125579834, 1.1456128358840942, 1.1439162492752075]}, {"epoch_end": 3, "step": 1564, "test_acc": 0.4073, "test_loss": 1.715226904296875}, {"epoch_end": 4, "step": 1955, "test_acc": 0.4106, "test_loss": 1.7046821655273436}], "final": {"test_acc": 0.4106, "test_loss": 1.7046821655273436, "wall_s": 24.226540088653564, "cos_r_negg": [0.7927378416061401, 0.37684792280197144, 0.3481896221637726, 0.3503452241420746, 0.34545525908470154, 0.37148505449295044, 0.4088486433029175, 0.3222513794898987, 0.39803820848464966, 0.36300128698349], "cos_innovation_negg": [0.7927378416061401, 0.37684792280197144, 0.3481896221637726, 0.3503452241420746, 0.34545525908470154, 0.37148505449295044, 0.4088486433029175, 0.3222513794898987, 0.39803820848464966, 0.36300128698349], "cos_apical_negg": [0.7927378416061401, 0.37684792280197144, 0.3481896221637726, 0.3503452241420746, 0.34545525908470154, 0.37148505449295044, 0.4088486433029175, 0.3222513794898987, 0.39803820848464966, 0.36300128698349], "cos_Ac_negg": [0.7927378416061401, 0.37684792280197144, 0.3481896221637726, 0.3503452241420746, 0.34545525908470154, 0.37148505449295044, 0.4088486433029175, 0.3222513794898987, 0.39803820848464966, 0.36300128698349]}}
|