summaryrefslogtreecommitdiff
path: root/results/scale_v3_cifar10_dfa_w64_d5_s1.json
blob: b1a92128768ecff9562c6c192194fd995741c5ca (plain)
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": 1, "device": "cuda", "log_every": 500, "max_steps": 0, "probe_bs": 256, "outdir": "results", "tag": "scale_v3_cifar10_dfa_w64_d5_s1", "n_in": 3072, "n_out": 10}, "provenance": {"git_commit": "d4aef1a571c21e6e4e4125c90aa58da7f14e62e3", "git_dirty": false}, "steps": [{"step": 0, "epoch": 0, "train_loss": 2.534536123275757, "cos_r_negg": [0.018351692706346512, -0.005833066999912262, 0.06346645951271057, 0.07211896032094955, 0.04269023239612579], "cos_innovation_negg": [0.018351692706346512, -0.005833066999912262, 0.06346645951271057, 0.07211896032094955, 0.04269023239612579], "cos_apical_negg": [0.018351692706346512, -0.005833066999912262, 0.06346645951271057, 0.07211896032094955, 0.04269023239612579], "cos_Ac_negg": [0.018351692706346512, -0.005833066999912262, 0.06346645951271057, 0.07211896032094955, 0.04269023239612579], "r_norm": [2.4398975372314453, 2.485071897506714, 2.5272884368896484, 2.4926908016204834, 2.4530205726623535]}, {"epoch_end": 0, "step": 391, "test_acc": 0.3325, "test_loss": 1.8760508911132812}, {"step": 500, "epoch": 1, "train_loss": 1.6629706621170044, "cos_r_negg": [0.5893737077713013, 0.04215037822723389, 0.061585791409015656, -0.0012980964966118336, 0.014921776950359344], "cos_innovation_negg": [0.5893737077713013, 0.04215037822723389, 0.061585791409015656, -0.0012980964966118336, 0.014921776950359344], "cos_apical_negg": [0.5893737077713013, 0.04215037822723389, 0.061585791409015656, -0.0012980964966118336, 0.014921776950359344], "cos_Ac_negg": [0.5893737077713013, 0.04215037822723389, 0.061585791409015656, -0.0012980964966118336, 0.014921776950359344], "r_norm": [2.122007131576538, 2.1985862255096436, 2.181349754333496, 2.1871256828308105, 2.1227478981018066]}, {"epoch_end": 1, "step": 782, "test_acc": 0.3661, "test_loss": 1.8418917236328125}, {"step": 1000, "epoch": 2, "train_loss": 1.805628776550293, "cos_r_negg": [0.5251514911651611, 0.12546846270561218, 0.151041179895401, 0.06834886968135834, 0.03871941566467285], "cos_innovation_negg": [0.5251514911651611, 0.12546846270561218, 0.151041179895401, 0.06834886968135834, 0.03871941566467285], "cos_apical_negg": [0.5251514911651611, 0.12546846270561218, 0.151041179895401, 0.06834886968135834, 0.03871941566467285], "cos_Ac_negg": [0.5251514911651611, 0.12546846270561218, 0.151041179895401, 0.06834886968135834, 0.03871941566467285], "r_norm": [2.107701539993286, 2.171658754348755, 2.160916805267334, 2.1639881134033203, 2.1090219020843506]}, {"epoch_end": 2, "step": 1173, "test_acc": 0.3543, "test_loss": 1.8146392822265625}, {"step": 1500, "epoch": 3, "train_loss": 1.8765915632247925, "cos_r_negg": [0.527320146560669, 0.1302276849746704, 0.1787794530391693, 0.13235454261302948, 0.11650253087282181], "cos_innovation_negg": [0.527320146560669, 0.1302276849746704, 0.1787794530391693, 0.13235454261302948, 0.11650253087282181], "cos_apical_negg": [0.527320146560669, 0.1302276849746704, 0.1787794530391693, 0.13235454261302948, 0.11650253087282181], "cos_Ac_negg": [0.527320146560669, 0.1302276849746704, 0.1787794530391693, 0.13235454261302948, 0.11650253087282181], "r_norm": [2.11971378326416, 2.195514678955078, 2.1841204166412354, 2.189694404602051, 2.129333019256592]}, {"epoch_end": 3, "step": 1564, "test_acc": 0.3719, "test_loss": 1.8052692016601561}, {"epoch_end": 4, "step": 1955, "test_acc": 0.3741, "test_loss": 1.7912805053710938}], "final": {"test_acc": 0.3741, "test_loss": 1.7912805053710938, "wall_s": 5.032076835632324, "cos_r_negg": [0.4774821400642395, 0.2160918414592743, 0.19929102063179016, 0.1342172622680664, 0.1082502007484436], "cos_innovation_negg": [0.4774821400642395, 0.2160918414592743, 0.19929102063179016, 0.1342172622680664, 0.1082502007484436], "cos_apical_negg": [0.4774821400642395, 0.2160918414592743, 0.19929102063179016, 0.1342172622680664, 0.1082502007484436], "cos_Ac_negg": [0.4774821400642395, 0.2160918414592743, 0.19929102063179016, 0.1342172622680664, 0.1082502007484436]}}