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": 3, "device": "cuda", "log_every": 500, "max_steps": 0, "probe_bs": 256, "outdir": "results", "tag": "scale_v3_cifar10_sdil_w64_d5_s3", "n_in": 3072, "n_out": 10}, "provenance": {"git_commit": "d4aef1a571c21e6e4e4125c90aa58da7f14e62e3", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.5312678813934326, "cos_r_negg": [0.015487607568502426, 0.023219961673021317, 0.04400324821472168, 0.019105009734630585, 0.014839014038443565], "cos_innovation_negg": [0.015487607568502426, 0.023219961673021317, 0.04400324821472168, 0.019105009734630585, 0.014839014038443565], "cos_apical_negg": [0.015487607568502426, 0.023219961673021317, 0.04400324821472168, 0.019105009734630585, 0.014839014038443565], "cos_Ac_negg": [0.015487607568502426, 0.023219961673021317, 0.04400324821472168, 0.019105009734630585, 0.014839014038443565], "r_norm": [2.511673927307129, 2.36773681640625, 2.4613699913024902, 2.4606597423553467, 2.576725959777832], "ldr": {"dL_sdil": -0.07907724380493164, "dL_bp": -0.497316837310791, "ratio": 0.15900777506857958}}, {"epoch_end": 0, "step": 391, "test_acc": 0.3615, "test_loss": 1.8238375122070312}, {"step": 500, "epoch": 1, "train_loss": 1.6975302696228027, "cos_r_negg": [0.5979571342468262, 0.13851620256900787, 0.1208072155714035, 0.047273121774196625, 0.1293143630027771], "cos_innovation_negg": [0.5979571342468262, 0.13851620256900787, 0.1208072155714035, 0.047273121774196625, 0.1293143630027771], "cos_apical_negg": [0.5979571342468262, 0.13851620256900787, 0.1208072155714035, 0.047273121774196625, 0.1293143630027771], "cos_Ac_negg": [0.5979571342468262, 0.13851620256900787, 0.1208072155714035, 0.047273121774196625, 0.1293143630027771], "r_norm": [1.8228027820587158, 1.6550185680389404, 1.719961404800415, 1.7156000137329102, 1.7934050559997559]}, {"epoch_end": 1, "step": 782, "test_acc": 0.3985, "test_loss": 1.7598261108398439}, {"step": 1000, "epoch": 2, "train_loss": 1.7444052696228027, "cos_r_negg": [0.7627414464950562, 0.23568947613239288, 0.20892804861068726, 0.12936270236968994, 0.15663352608680725], "cos_innovation_negg": [0.7627414464950562, 0.23568947613239288, 0.20892804861068726, 0.12936270236968994, 0.15663352608680725], "cos_apical_negg": [0.7627414464950562, 0.23568947613239288, 0.20892804861068726, 0.12936270236968994, 0.15663352608680725], "cos_Ac_negg": [0.7627414464950562, 0.23568947613239288, 0.20892804861068726, 0.12936270236968994, 0.15663352608680725], "r_norm": [1.5514581203460693, 1.3559077978134155, 1.3964686393737793, 1.3915657997131348, 1.4553321599960327]}, {"epoch_end": 2, "step": 1173, "test_acc": 0.3983, "test_loss": 1.7326050659179688}, {"step": 1500, "epoch": 3, "train_loss": 1.6366461515426636, "cos_r_negg": [0.7419991493225098, 0.32193809747695923, 0.33374249935150146, 0.26523470878601074, 0.3017241358757019], "cos_innovation_negg": [0.7419991493225098, 0.32193809747695923, 0.33374249935150146, 0.26523470878601074, 0.3017241358757019], "cos_apical_negg": [0.7419991493225098, 0.32193809747695923, 0.33374249935150146, 0.26523470878601074, 0.3017241358757019], "cos_Ac_negg": [0.7419991493225098, 0.32193809747695923, 0.33374249935150146, 0.26523470878601074, 0.3017241358757019], "r_norm": [1.3477027416229248, 1.127578854560852, 1.1581921577453613, 1.1418871879577637, 1.205965280532837]}, {"epoch_end": 3, "step": 1564, "test_acc": 0.4054, "test_loss": 1.717952001953125}, {"epoch_end": 4, "step": 1955, "test_acc": 0.4182, "test_loss": 1.691331396484375}], "final": {"test_acc": 0.4182, "test_loss": 1.691331396484375, "wall_s": 8.942532777786255, "cos_r_negg": [0.7804144620895386, 0.42172783613204956, 0.4234136939048767, 0.3621608018875122, 0.39389240741729736], "cos_innovation_negg": [0.7804144620895386, 0.42172783613204956, 0.4234136939048767, 0.3621608018875122, 0.39389240741729736], "cos_apical_negg": [0.7804144620895386, 0.42172783613204956, 0.4234136939048767, 0.3621608018875122, 0.39389240741729736], "cos_Ac_negg": [0.7804144620895386, 0.42172783613204956, 0.4234136939048767, 0.3621608018875122, 0.39389240741729736]}}
|