summaryrefslogtreecommitdiff
path: root/results/scale_v3_cifar10_sdil_w64_d5_s2.json
blob: afa7a70bd911be489faeadb4def7abc889e28c10 (plain)
1
{"args": {"mode": "sdil", "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": 0.02, "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": 2, "device": "cuda", "log_every": 500, "max_steps": 0, "probe_bs": 256, "outdir": "results", "tag": "scale_v3_cifar10_sdil_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": [0.01960688829421997, 0.05508289486169815, -0.010979766957461834, 0.026988983154296875, -0.01279059611260891], "cos_innovation_negg": [0.01960688829421997, 0.05508289486169815, -0.010979766957461834, 0.026988983154296875, -0.01279059611260891], "cos_apical_negg": [0.01960688829421997, 0.05508289486169815, -0.010979766957461834, 0.026988983154296875, -0.01279059611260891], "cos_Ac_negg": [0.01960688829421997, 0.05508289486169815, -0.010979766957461834, 0.026988983154296875, -0.01279059611260891], "r_norm": [2.4072906970977783, 2.5697972774505615, 2.5095982551574707, 2.7263479232788086, 2.4590187072753906], "ldr": {"dL_sdil": 0.05171966552734375, "dL_bp": -0.5048151016235352, "ratio": -0.10245269081889231}}, {"epoch_end": 0, "step": 391, "test_acc": 0.3681, "test_loss": 1.8036007690429687}, {"step": 500, "epoch": 1, "train_loss": 1.7465269565582275, "cos_r_negg": [0.6627542972564697, 0.14648012816905975, 0.023487048223614693, 0.00028585968539118767, 0.08604652434587479], "cos_innovation_negg": [0.6627542972564697, 0.14648012816905975, 0.023487048223614693, 0.00028585968539118767, 0.08604652434587479], "cos_apical_negg": [0.6627542972564697, 0.14648012816905975, 0.023487048223614693, 0.00028585968539118767, 0.08604652434587479], "cos_Ac_negg": [0.6627542972564697, 0.14648012816905975, 0.023487048223614693, 0.00028585968539118767, 0.08604652434587479], "r_norm": [1.7211203575134277, 1.7824331521987915, 1.7514915466308594, 1.8600280284881592, 1.6989855766296387]}, {"epoch_end": 1, "step": 782, "test_acc": 0.3948, "test_loss": 1.7500324340820312}, {"step": 1000, "epoch": 2, "train_loss": 1.6957550048828125, "cos_r_negg": [0.7608604431152344, 0.21594570577144623, 0.13476064801216125, 0.13577653467655182, 0.16765320301055908], "cos_innovation_negg": [0.7608604431152344, 0.21594570577144623, 0.13476064801216125, 0.13577653467655182, 0.16765320301055908], "cos_apical_negg": [0.7608604431152344, 0.21594570577144623, 0.13476064801216125, 0.13577653467655182, 0.16765320301055908], "cos_Ac_negg": [0.7608604431152344, 0.21594570577144623, 0.13476064801216125, 0.13577653467655182, 0.16765320301055908], "r_norm": [1.488640546798706, 1.4695091247558594, 1.425648808479309, 1.5152889490127563, 1.39373779296875]}, {"epoch_end": 2, "step": 1173, "test_acc": 0.3937, "test_loss": 1.736881640625}, {"step": 1500, "epoch": 3, "train_loss": 1.602670431137085, "cos_r_negg": [0.7490943670272827, 0.3311230540275574, 0.2521922290325165, 0.23987074196338654, 0.2811569273471832], "cos_innovation_negg": [0.7490943670272827, 0.3311230540275574, 0.2521922290325165, 0.23987074196338654, 0.2811569273471832], "cos_apical_negg": [0.7490943670272827, 0.3311230540275574, 0.2521922290325165, 0.23987074196338654, 0.2811569273471832], "cos_Ac_negg": [0.7490943670272827, 0.3311230540275574, 0.2521922290325165, 0.23987074196338654, 0.2811569273471832], "r_norm": [1.2930324077606201, 1.2165963649749756, 1.1653194427490234, 1.2358779907226562, 1.151439905166626]}, {"epoch_end": 3, "step": 1564, "test_acc": 0.4054, "test_loss": 1.7063751342773437}, {"epoch_end": 4, "step": 1955, "test_acc": 0.4121, "test_loss": 1.6912419311523437}], "final": {"test_acc": 0.4121, "test_loss": 1.6912419311523437, "wall_s": 9.297170162200928, "cos_r_negg": [0.8035550117492676, 0.41306135058403015, 0.3653326630592346, 0.3529694080352783, 0.3765638470649719], "cos_innovation_negg": [0.8035550117492676, 0.41306135058403015, 0.3653326630592346, 0.3529694080352783, 0.3765638470649719], "cos_apical_negg": [0.8035550117492676, 0.41306135058403015, 0.3653326630592346, 0.3529694080352783, 0.3765638470649719], "cos_Ac_negg": [0.8035550117492676, 0.41306135058403015, 0.3653326630592346, 0.3529694080352783, 0.3765638470649719]}}