summaryrefslogtreecommitdiff
path: root/results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json
diff options
context:
space:
mode:
Diffstat (limited to 'results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json')
-rw-r--r--results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json2281
1 files changed, 2281 insertions, 0 deletions
diff --git a/results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json b/results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json
new file mode 100644
index 0000000..24ebf20
--- /dev/null
+++ b/results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json
@@ -0,0 +1,2281 @@
+{
+ "architecture": {
+ "adaptive_apical_parameters": 6389760,
+ "adaptive_feedback_parameters": 0,
+ "base_width": 16,
+ "blocks_per_stage": 9,
+ "bn_eps": 1e-05,
+ "bn_momentum": 0.1,
+ "depth": 56,
+ "family": "CIFAR 6n+2 ResNet, option-A shortcuts",
+ "fixed_feedback_parameters": 0,
+ "fixed_traffic_coefficients": 532480,
+ "forward_parameters": 853018,
+ "hidden_shapes": [
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 16,
+ 32,
+ 32
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 32,
+ 16,
+ 16
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ]
+ ],
+ "normalization": "batchnorm",
+ "predictor_parameters": 1064960,
+ "residual_scale": 1.0,
+ "vectorizer_mode": "spatial_template",
+ "vectorizer_parameters": 5324800
+ },
+ "args": {
+ "a_scale": 1.0,
+ "a_warmup_steps": 0,
+ "alignment_probe": 32,
+ "apical_calibration_mode": "unit_targets",
+ "apical_seed": null,
+ "augment_train": 1,
+ "batch_size": 128,
+ "bn_eps": 1e-05,
+ "bn_momentum": 0.1,
+ "data_dir": "/home/yurenh2/sdrn/data",
+ "depth": 56,
+ "device": "cuda",
+ "epochs": 200,
+ "eta_A": 0.01,
+ "eta_P": 0.01,
+ "eval_every": 0,
+ "eval_split": "test",
+ "learn_P": 0,
+ "loader_seed": 13,
+ "lr": 0.03,
+ "lr_gamma": 0.1,
+ "lr_milestones": "100,150",
+ "lr_schedule": "step",
+ "max_steps": 0,
+ "mirror_batch_size": 1,
+ "mirror_eta": 0.1,
+ "mirror_every": 16,
+ "mirror_noise_std": 1.0,
+ "mirror_seed": 3000,
+ "mirror_warmup_steps": 0,
+ "mode": "dfa",
+ "momentum": 0.9,
+ "neutral_projection": 0,
+ "normalization": "batchnorm",
+ "nuisance_scale": 0.0,
+ "out": "results/oral_a_dynamic_scaling_v2/dfa_d56_s13.json",
+ "output_lr": 0.1,
+ "pert_directions": 1,
+ "pert_every": 4,
+ "pert_sigma": 0.01,
+ "perturb_seed": 1000,
+ "predictor_every": 0,
+ "predictor_mode": "nlms",
+ "predictor_warmup_steps": 0,
+ "residual_scale": null,
+ "seed": 13,
+ "split_seed": 2027,
+ "traffic_calibration_examples": 64,
+ "traffic_ratio": 4.0,
+ "traffic_rule": "innovation",
+ "traffic_seed": 4000,
+ "train_limit": 0,
+ "use_residual": 1,
+ "val_examples": 0,
+ "vectorizer_mode": "spatial_template",
+ "warmup_epochs": 0,
+ "weight_decay": 0.0001,
+ "weight_scale": 1.0,
+ "width": 16
+ },
+ "calibration_metric_space": "full_hidden_field",
+ "counters": {
+ "apical_warmup_examples": 0,
+ "calibration_event_examples": 0,
+ "causal_scalar_observations": 0,
+ "logical_batch_loss_queries": 0,
+ "mirror_conv_examples": 0,
+ "mirror_events": 0,
+ "mirror_readout_examples": 0,
+ "neutral_projection_examples": 0,
+ "ordinary_examples": 10000000,
+ "per_example_loss_terms": 0,
+ "perturbation_events": 0,
+ "perturbation_forward_examples": 0,
+ "predictor_update_examples": 0,
+ "predictor_warmup_examples": 0,
+ "traffic_audit_examples": 0,
+ "traffic_calibration_examples": 0
+ },
+ "diagnostics": {
+ "early_third_mean": 0.00043991863236038224,
+ "innovation_negative_gradient_cosine": [
+ 0.0005255909636616707,
+ -0.0010781302116811275,
+ 0.0025069178082048893,
+ 0.0016919649206101894,
+ 0.00010788301005959511,
+ 2.1327403374016285e-05,
+ -0.000367272412404418,
+ 0.0003776442026719451,
+ 0.0016317082336172462,
+ 0.0014963049907237291,
+ -0.0005021097604185343,
+ -0.0002937753451988101,
+ -0.0004791772225871682,
+ 6.390659837052226e-05,
+ -0.001030154642648995,
+ 0.0014018112560734153,
+ 0.0018113252008333802,
+ 3.277038922533393e-05,
+ -4.067507688887417e-05,
+ -0.0006657886551693082,
+ 0.0020185261964797974,
+ -0.0015718279173597693,
+ -0.0024743354879319668,
+ 0.0002824710390996188,
+ 0.003487984649837017,
+ 0.0005018783849664032,
+ -0.0022507067769765854,
+ -0.0005763396038673818,
+ 0.0007501582731492817,
+ 0.0011279437458142638,
+ 0.0008351888973265886,
+ 0.0006766966544091702,
+ 0.0011843631509691477,
+ -3.760406980291009e-05,
+ 0.003942208830267191,
+ 0.00019845661881845444,
+ 0.002232921775430441,
+ 0.004017248749732971,
+ 0.007412542589008808,
+ 0.0004585348069667816,
+ 0.007836693897843361,
+ 0.004988133907318115,
+ 0.002706740517169237,
+ 0.003983347676694393,
+ 0.006801050156354904,
+ 0.0037632600869983435,
+ 0.010236972942948341,
+ 0.00855157058686018,
+ 0.009376580826938152,
+ 0.006161993369460106,
+ 0.008103354834020138,
+ 0.011376159265637398,
+ 0.014972970820963383,
+ 0.012046518735587597,
+ 0.010956445708870888
+ ],
+ "normalization_state": "training_batch_stats_without_running_update",
+ "raw_negative_gradient_cosine": [
+ 0.0005255909636616707,
+ -0.0010781302116811275,
+ 0.0025069178082048893,
+ 0.0016919649206101894,
+ 0.00010788301005959511,
+ 2.1327403374016285e-05,
+ -0.000367272412404418,
+ 0.0003776442026719451,
+ 0.0016317082336172462,
+ 0.0014963049907237291,
+ -0.0005021097604185343,
+ -0.0002937753451988101,
+ -0.0004791772225871682,
+ 6.390659837052226e-05,
+ -0.001030154642648995,
+ 0.0014018112560734153,
+ 0.0018113252008333802,
+ 3.277038922533393e-05,
+ -4.067507688887417e-05,
+ -0.0006657886551693082,
+ 0.0020185261964797974,
+ -0.0015718279173597693,
+ -0.0024743354879319668,
+ 0.0002824710390996188,
+ 0.003487984649837017,
+ 0.0005018783849664032,
+ -0.0022507067769765854,
+ -0.0005763396038673818,
+ 0.0007501582731492817,
+ 0.0011279437458142638,
+ 0.0008351888973265886,
+ 0.0006766966544091702,
+ 0.0011843631509691477,
+ -3.760406980291009e-05,
+ 0.003942208830267191,
+ 0.00019845661881845444,
+ 0.002232921775430441,
+ 0.004017248749732971,
+ 0.007412542589008808,
+ 0.0004585348069667816,
+ 0.007836693897843361,
+ 0.004988133907318115,
+ 0.002706740517169237,
+ 0.003983347676694393,
+ 0.006801050156354904,
+ 0.0037632600869983435,
+ 0.010236972942948341,
+ 0.00855157058686018,
+ 0.009376580826938152,
+ 0.006161993369460106,
+ 0.008103354834020138,
+ 0.011376159265637398,
+ 0.014972970820963383,
+ 0.012046518735587597,
+ 0.010956445708870888
+ ],
+ "teaching_negative_gradient_cosine": [
+ 0.0005255909636616707,
+ -0.0010781302116811275,
+ 0.0025069178082048893,
+ 0.0016919649206101894,
+ 0.00010788301005959511,
+ 2.1327403374016285e-05,
+ -0.000367272412404418,
+ 0.0003776442026719451,
+ 0.0016317082336172462,
+ 0.0014963049907237291,
+ -0.0005021097604185343,
+ -0.0002937753451988101,
+ -0.0004791772225871682,
+ 6.390659837052226e-05,
+ -0.001030154642648995,
+ 0.0014018112560734153,
+ 0.0018113252008333802,
+ 3.277038922533393e-05,
+ -4.067507688887417e-05,
+ -0.0006657886551693082,
+ 0.0020185261964797974,
+ -0.0015718279173597693,
+ -0.0024743354879319668,
+ 0.0002824710390996188,
+ 0.003487984649837017,
+ 0.0005018783849664032,
+ -0.0022507067769765854,
+ -0.0005763396038673818,
+ 0.0007501582731492817,
+ 0.0011279437458142638,
+ 0.0008351888973265886,
+ 0.0006766966544091702,
+ 0.0011843631509691477,
+ -3.760406980291009e-05,
+ 0.003942208830267191,
+ 0.00019845661881845444,
+ 0.002232921775430441,
+ 0.004017248749732971,
+ 0.007412542589008808,
+ 0.0004585348069667816,
+ 0.007836693897843361,
+ 0.004988133907318115,
+ 0.002706740517169237,
+ 0.003983347676694393,
+ 0.006801050156354904,
+ 0.0037632600869983435,
+ 0.010236972942948341,
+ 0.00855157058686018,
+ 0.009376580826938152,
+ 0.006161993369460106,
+ 0.008103354834020138,
+ 0.011376159265637398,
+ 0.014972970820963383,
+ 0.012046518735587597,
+ 0.010956445708870888
+ ],
+ "wall_s": 0.20534896850585938
+ },
+ "epochs": [
+ {
+ "epoch": 1,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 391,
+ "train_examples": 50000,
+ "train_loss": 19.85203016052246
+ },
+ {
+ "epoch": 2,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 782,
+ "train_examples": 50000,
+ "train_loss": 15.277210861816407
+ },
+ {
+ "epoch": 3,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1173,
+ "train_examples": 50000,
+ "train_loss": 22.87974688232422
+ },
+ {
+ "epoch": 4,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1564,
+ "train_examples": 50000,
+ "train_loss": 25.236587622680663
+ },
+ {
+ "epoch": 5,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1955,
+ "train_examples": 50000,
+ "train_loss": 27.989678770751954
+ },
+ {
+ "epoch": 6,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 2346,
+ "train_examples": 50000,
+ "train_loss": 44.29404400390625
+ },
+ {
+ "epoch": 7,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 2737,
+ "train_examples": 50000,
+ "train_loss": 55.89321231445312
+ },
+ {
+ "epoch": 8,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3128,
+ "train_examples": 50000,
+ "train_loss": 86.94409305175782
+ },
+ {
+ "epoch": 9,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3519,
+ "train_examples": 50000,
+ "train_loss": 114.98562063476562
+ },
+ {
+ "epoch": 10,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3910,
+ "train_examples": 50000,
+ "train_loss": 165.99049699707032
+ },
+ {
+ "epoch": 11,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 4301,
+ "train_examples": 50000,
+ "train_loss": 228.40711592285157
+ },
+ {
+ "epoch": 12,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 4692,
+ "train_examples": 50000,
+ "train_loss": 289.34972493164065
+ },
+ {
+ "epoch": 13,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5083,
+ "train_examples": 50000,
+ "train_loss": 364.11808366210937
+ },
+ {
+ "epoch": 14,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5474,
+ "train_examples": 50000,
+ "train_loss": 439.4125665820313
+ },
+ {
+ "epoch": 15,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5865,
+ "train_examples": 50000,
+ "train_loss": 671.6069704296875
+ },
+ {
+ "epoch": 16,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 6256,
+ "train_examples": 50000,
+ "train_loss": 712.7547475976562
+ },
+ {
+ "epoch": 17,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 6647,
+ "train_examples": 50000,
+ "train_loss": 980.140315546875
+ },
+ {
+ "epoch": 18,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7038,
+ "train_examples": 50000,
+ "train_loss": 1103.293306484375
+ },
+ {
+ "epoch": 19,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7429,
+ "train_examples": 50000,
+ "train_loss": 1320.8525594921875
+ },
+ {
+ "epoch": 20,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7820,
+ "train_examples": 50000,
+ "train_loss": 1465.6310321875
+ },
+ {
+ "epoch": 21,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8211,
+ "train_examples": 50000,
+ "train_loss": 1796.9750183007814
+ },
+ {
+ "epoch": 22,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8602,
+ "train_examples": 50000,
+ "train_loss": 1948.1184459765625
+ },
+ {
+ "epoch": 23,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8993,
+ "train_examples": 50000,
+ "train_loss": 2287.01851765625
+ },
+ {
+ "epoch": 24,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 9384,
+ "train_examples": 50000,
+ "train_loss": 2444.085239296875
+ },
+ {
+ "epoch": 25,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 9775,
+ "train_examples": 50000,
+ "train_loss": 2908.4220490625
+ },
+ {
+ "epoch": 26,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10166,
+ "train_examples": 50000,
+ "train_loss": 3158.44132859375
+ },
+ {
+ "epoch": 27,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10557,
+ "train_examples": 50000,
+ "train_loss": 3313.92180625
+ },
+ {
+ "epoch": 28,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10948,
+ "train_examples": 50000,
+ "train_loss": 3877.503080390625
+ },
+ {
+ "epoch": 29,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 11339,
+ "train_examples": 50000,
+ "train_loss": 4540.05339828125
+ },
+ {
+ "epoch": 30,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 11730,
+ "train_examples": 50000,
+ "train_loss": 4695.8629065625
+ },
+ {
+ "epoch": 31,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12121,
+ "train_examples": 50000,
+ "train_loss": 5154.7014296875
+ },
+ {
+ "epoch": 32,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12512,
+ "train_examples": 50000,
+ "train_loss": 5821.560145234375
+ },
+ {
+ "epoch": 33,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12903,
+ "train_examples": 50000,
+ "train_loss": 6480.87629140625
+ },
+ {
+ "epoch": 34,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 13294,
+ "train_examples": 50000,
+ "train_loss": 6341.7844628125
+ },
+ {
+ "epoch": 35,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 13685,
+ "train_examples": 50000,
+ "train_loss": 6923.978560625
+ },
+ {
+ "epoch": 36,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14076,
+ "train_examples": 50000,
+ "train_loss": 7701.088639375
+ },
+ {
+ "epoch": 37,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14467,
+ "train_examples": 50000,
+ "train_loss": 7469.7678303125
+ },
+ {
+ "epoch": 38,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14858,
+ "train_examples": 50000,
+ "train_loss": 8743.23543609375
+ },
+ {
+ "epoch": 39,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 15249,
+ "train_examples": 50000,
+ "train_loss": 8046.617881875
+ },
+ {
+ "epoch": 40,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 15640,
+ "train_examples": 50000,
+ "train_loss": 9276.7875521875
+ },
+ {
+ "epoch": 41,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16031,
+ "train_examples": 50000,
+ "train_loss": 10553.15291
+ },
+ {
+ "epoch": 42,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16422,
+ "train_examples": 50000,
+ "train_loss": 10569.499498125
+ },
+ {
+ "epoch": 43,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16813,
+ "train_examples": 50000,
+ "train_loss": 11961.35611875
+ },
+ {
+ "epoch": 44,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17204,
+ "train_examples": 50000,
+ "train_loss": 12400.59445
+ },
+ {
+ "epoch": 45,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17595,
+ "train_examples": 50000,
+ "train_loss": 12074.1799596875
+ },
+ {
+ "epoch": 46,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17986,
+ "train_examples": 50000,
+ "train_loss": 14217.12017125
+ },
+ {
+ "epoch": 47,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 18377,
+ "train_examples": 50000,
+ "train_loss": 13769.598210625
+ },
+ {
+ "epoch": 48,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 18768,
+ "train_examples": 50000,
+ "train_loss": 13573.48423125
+ },
+ {
+ "epoch": 49,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19159,
+ "train_examples": 50000,
+ "train_loss": 15667.26407375
+ },
+ {
+ "epoch": 50,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19550,
+ "train_examples": 50000,
+ "train_loss": 16657.7972615625
+ },
+ {
+ "epoch": 51,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19941,
+ "train_examples": 50000,
+ "train_loss": 15693.64605375
+ },
+ {
+ "epoch": 52,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 20332,
+ "train_examples": 50000,
+ "train_loss": 18511.232810625
+ },
+ {
+ "epoch": 53,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 20723,
+ "train_examples": 50000,
+ "train_loss": 18231.63009625
+ },
+ {
+ "epoch": 54,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21114,
+ "train_examples": 50000,
+ "train_loss": 17852.489756875
+ },
+ {
+ "epoch": 55,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21505,
+ "train_examples": 50000,
+ "train_loss": 18452.930683125
+ },
+ {
+ "epoch": 56,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21896,
+ "train_examples": 50000,
+ "train_loss": 21131.5533175
+ },
+ {
+ "epoch": 57,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 22287,
+ "train_examples": 50000,
+ "train_loss": 21925.75662875
+ },
+ {
+ "epoch": 58,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 22678,
+ "train_examples": 50000,
+ "train_loss": 23009.8898825
+ },
+ {
+ "epoch": 59,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23069,
+ "train_examples": 50000,
+ "train_loss": 22536.6942965625
+ },
+ {
+ "epoch": 60,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23460,
+ "train_examples": 50000,
+ "train_loss": 23355.11246875
+ },
+ {
+ "epoch": 61,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23851,
+ "train_examples": 50000,
+ "train_loss": 23856.826025
+ },
+ {
+ "epoch": 62,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 24242,
+ "train_examples": 50000,
+ "train_loss": 26181.8635275
+ },
+ {
+ "epoch": 63,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 24633,
+ "train_examples": 50000,
+ "train_loss": 25788.6849225
+ },
+ {
+ "epoch": 64,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25024,
+ "train_examples": 50000,
+ "train_loss": 28495.095256875
+ },
+ {
+ "epoch": 65,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25415,
+ "train_examples": 50000,
+ "train_loss": 28800.2621128125
+ },
+ {
+ "epoch": 66,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25806,
+ "train_examples": 50000,
+ "train_loss": 29564.81689
+ },
+ {
+ "epoch": 67,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26197,
+ "train_examples": 50000,
+ "train_loss": 28106.36562625
+ },
+ {
+ "epoch": 68,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26588,
+ "train_examples": 50000,
+ "train_loss": 30395.82846625
+ },
+ {
+ "epoch": 69,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26979,
+ "train_examples": 50000,
+ "train_loss": 34965.03667
+ },
+ {
+ "epoch": 70,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 27370,
+ "train_examples": 50000,
+ "train_loss": 32706.780568125
+ },
+ {
+ "epoch": 71,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 27761,
+ "train_examples": 50000,
+ "train_loss": 37163.961801875
+ },
+ {
+ "epoch": 72,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28152,
+ "train_examples": 50000,
+ "train_loss": 37854.28123375
+ },
+ {
+ "epoch": 73,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28543,
+ "train_examples": 50000,
+ "train_loss": 33942.487185
+ },
+ {
+ "epoch": 74,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28934,
+ "train_examples": 50000,
+ "train_loss": 38028.11356375
+ },
+ {
+ "epoch": 75,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 29325,
+ "train_examples": 50000,
+ "train_loss": 39465.98959875
+ },
+ {
+ "epoch": 76,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 29716,
+ "train_examples": 50000,
+ "train_loss": 40007.19174
+ },
+ {
+ "epoch": 77,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30107,
+ "train_examples": 50000,
+ "train_loss": 42236.5418225
+ },
+ {
+ "epoch": 78,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30498,
+ "train_examples": 50000,
+ "train_loss": 42935.5670475
+ },
+ {
+ "epoch": 79,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30889,
+ "train_examples": 50000,
+ "train_loss": 44710.2431825
+ },
+ {
+ "epoch": 80,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 31280,
+ "train_examples": 50000,
+ "train_loss": 46254.5428475
+ },
+ {
+ "epoch": 81,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 31671,
+ "train_examples": 50000,
+ "train_loss": 45069.667475
+ },
+ {
+ "epoch": 82,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32062,
+ "train_examples": 50000,
+ "train_loss": 45259.70246625
+ },
+ {
+ "epoch": 83,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32453,
+ "train_examples": 50000,
+ "train_loss": 45629.35346375
+ },
+ {
+ "epoch": 84,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32844,
+ "train_examples": 50000,
+ "train_loss": 45828.0199975
+ },
+ {
+ "epoch": 85,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 33235,
+ "train_examples": 50000,
+ "train_loss": 51119.8721825
+ },
+ {
+ "epoch": 86,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 33626,
+ "train_examples": 50000,
+ "train_loss": 54685.955385
+ },
+ {
+ "epoch": 87,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34017,
+ "train_examples": 50000,
+ "train_loss": 48313.62319375
+ },
+ {
+ "epoch": 88,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34408,
+ "train_examples": 50000,
+ "train_loss": 53704.67348
+ },
+ {
+ "epoch": 89,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34799,
+ "train_examples": 50000,
+ "train_loss": 49807.59677
+ },
+ {
+ "epoch": 90,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35190,
+ "train_examples": 50000,
+ "train_loss": 54847.301465
+ },
+ {
+ "epoch": 91,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35581,
+ "train_examples": 50000,
+ "train_loss": 52811.8402475
+ },
+ {
+ "epoch": 92,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35972,
+ "train_examples": 50000,
+ "train_loss": 53685.5629375
+ },
+ {
+ "epoch": 93,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 36363,
+ "train_examples": 50000,
+ "train_loss": 58321.18193875
+ },
+ {
+ "epoch": 94,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 36754,
+ "train_examples": 50000,
+ "train_loss": 53370.31019
+ },
+ {
+ "epoch": 95,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37145,
+ "train_examples": 50000,
+ "train_loss": 55259.42770875
+ },
+ {
+ "epoch": 96,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37536,
+ "train_examples": 50000,
+ "train_loss": 62743.202595
+ },
+ {
+ "epoch": 97,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37927,
+ "train_examples": 50000,
+ "train_loss": 71849.919185
+ },
+ {
+ "epoch": 98,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 38318,
+ "train_examples": 50000,
+ "train_loss": 62031.9489825
+ },
+ {
+ "epoch": 99,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 38709,
+ "train_examples": 50000,
+ "train_loss": 68600.14554
+ },
+ {
+ "epoch": 100,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 39100,
+ "train_examples": 50000,
+ "train_loss": 68424.4290175
+ },
+ {
+ "epoch": 101,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 39491,
+ "train_examples": 50000,
+ "train_loss": 13175.8239165625
+ },
+ {
+ "epoch": 102,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 39882,
+ "train_examples": 50000,
+ "train_loss": 9052.00531375
+ },
+ {
+ "epoch": 103,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 40273,
+ "train_examples": 50000,
+ "train_loss": 8383.80109984375
+ },
+ {
+ "epoch": 104,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 40664,
+ "train_examples": 50000,
+ "train_loss": 8050.7802765625
+ },
+ {
+ "epoch": 105,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41055,
+ "train_examples": 50000,
+ "train_loss": 7575.8674025
+ },
+ {
+ "epoch": 106,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41446,
+ "train_examples": 50000,
+ "train_loss": 7245.3161803125
+ },
+ {
+ "epoch": 107,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41837,
+ "train_examples": 50000,
+ "train_loss": 6778.7499434375
+ },
+ {
+ "epoch": 108,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 42228,
+ "train_examples": 50000,
+ "train_loss": 6560.2829946875
+ },
+ {
+ "epoch": 109,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 42619,
+ "train_examples": 50000,
+ "train_loss": 6089.2685134375
+ },
+ {
+ "epoch": 110,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43010,
+ "train_examples": 50000,
+ "train_loss": 5857.6013596875
+ },
+ {
+ "epoch": 111,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43401,
+ "train_examples": 50000,
+ "train_loss": 5703.1139596875
+ },
+ {
+ "epoch": 112,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43792,
+ "train_examples": 50000,
+ "train_loss": 5420.600954375
+ },
+ {
+ "epoch": 113,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44183,
+ "train_examples": 50000,
+ "train_loss": 5243.32939
+ },
+ {
+ "epoch": 114,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44574,
+ "train_examples": 50000,
+ "train_loss": 5076.914825
+ },
+ {
+ "epoch": 115,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44965,
+ "train_examples": 50000,
+ "train_loss": 5148.14045984375
+ },
+ {
+ "epoch": 116,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 45356,
+ "train_examples": 50000,
+ "train_loss": 5220.29229140625
+ },
+ {
+ "epoch": 117,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 45747,
+ "train_examples": 50000,
+ "train_loss": 4584.908725703125
+ },
+ {
+ "epoch": 118,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46138,
+ "train_examples": 50000,
+ "train_loss": 5142.28931734375
+ },
+ {
+ "epoch": 119,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46529,
+ "train_examples": 50000,
+ "train_loss": 5081.4689378125
+ },
+ {
+ "epoch": 120,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46920,
+ "train_examples": 50000,
+ "train_loss": 4670.038472578125
+ },
+ {
+ "epoch": 121,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 47311,
+ "train_examples": 50000,
+ "train_loss": 4729.2374884375
+ },
+ {
+ "epoch": 122,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 47702,
+ "train_examples": 50000,
+ "train_loss": 4437.689307421875
+ },
+ {
+ "epoch": 123,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48093,
+ "train_examples": 50000,
+ "train_loss": 4643.063570703125
+ },
+ {
+ "epoch": 124,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48484,
+ "train_examples": 50000,
+ "train_loss": 4598.3138490625
+ },
+ {
+ "epoch": 125,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48875,
+ "train_examples": 50000,
+ "train_loss": 4710.731543203125
+ },
+ {
+ "epoch": 126,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 49266,
+ "train_examples": 50000,
+ "train_loss": 4922.6924784375
+ },
+ {
+ "epoch": 127,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 49657,
+ "train_examples": 50000,
+ "train_loss": 4706.8065034375
+ },
+ {
+ "epoch": 128,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50048,
+ "train_examples": 50000,
+ "train_loss": 4620.84078453125
+ },
+ {
+ "epoch": 129,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50439,
+ "train_examples": 50000,
+ "train_loss": 4983.67323234375
+ },
+ {
+ "epoch": 130,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50830,
+ "train_examples": 50000,
+ "train_loss": 4512.0828665625
+ },
+ {
+ "epoch": 131,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 51221,
+ "train_examples": 50000,
+ "train_loss": 4907.38452609375
+ },
+ {
+ "epoch": 132,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 51612,
+ "train_examples": 50000,
+ "train_loss": 4981.26236390625
+ },
+ {
+ "epoch": 133,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52003,
+ "train_examples": 50000,
+ "train_loss": 4804.82025484375
+ },
+ {
+ "epoch": 134,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52394,
+ "train_examples": 50000,
+ "train_loss": 4912.26073234375
+ },
+ {
+ "epoch": 135,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52785,
+ "train_examples": 50000,
+ "train_loss": 5047.9833865625
+ },
+ {
+ "epoch": 136,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53176,
+ "train_examples": 50000,
+ "train_loss": 4543.504403125
+ },
+ {
+ "epoch": 137,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53567,
+ "train_examples": 50000,
+ "train_loss": 5049.753786875
+ },
+ {
+ "epoch": 138,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53958,
+ "train_examples": 50000,
+ "train_loss": 4841.258673125
+ },
+ {
+ "epoch": 139,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 54349,
+ "train_examples": 50000,
+ "train_loss": 4753.69637796875
+ },
+ {
+ "epoch": 140,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 54740,
+ "train_examples": 50000,
+ "train_loss": 5182.0467875
+ },
+ {
+ "epoch": 141,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55131,
+ "train_examples": 50000,
+ "train_loss": 5353.002407578125
+ },
+ {
+ "epoch": 142,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55522,
+ "train_examples": 50000,
+ "train_loss": 4778.885925625
+ },
+ {
+ "epoch": 143,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55913,
+ "train_examples": 50000,
+ "train_loss": 5801.38689421875
+ },
+ {
+ "epoch": 144,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 56304,
+ "train_examples": 50000,
+ "train_loss": 4708.11906
+ },
+ {
+ "epoch": 145,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 56695,
+ "train_examples": 50000,
+ "train_loss": 5412.79496265625
+ },
+ {
+ "epoch": 146,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57086,
+ "train_examples": 50000,
+ "train_loss": 5151.65164765625
+ },
+ {
+ "epoch": 147,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57477,
+ "train_examples": 50000,
+ "train_loss": 5323.728211171875
+ },
+ {
+ "epoch": 148,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57868,
+ "train_examples": 50000,
+ "train_loss": 4849.5092225
+ },
+ {
+ "epoch": 149,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 58259,
+ "train_examples": 50000,
+ "train_loss": 5417.35654671875
+ },
+ {
+ "epoch": 150,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 58650,
+ "train_examples": 50000,
+ "train_loss": 4851.2205709375
+ },
+ {
+ "epoch": 151,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59041,
+ "train_examples": 50000,
+ "train_loss": 1917.0817833203125
+ },
+ {
+ "epoch": 152,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59432,
+ "train_examples": 50000,
+ "train_loss": 1419.874765390625
+ },
+ {
+ "epoch": 153,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59823,
+ "train_examples": 50000,
+ "train_loss": 1290.1946449609375
+ },
+ {
+ "epoch": 154,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60214,
+ "train_examples": 50000,
+ "train_loss": 1215.0555855859375
+ },
+ {
+ "epoch": 155,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60605,
+ "train_examples": 50000,
+ "train_loss": 1163.6108630859376
+ },
+ {
+ "epoch": 156,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60996,
+ "train_examples": 50000,
+ "train_loss": 1105.2684625390625
+ },
+ {
+ "epoch": 157,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 61387,
+ "train_examples": 50000,
+ "train_loss": 1058.7536784765625
+ },
+ {
+ "epoch": 158,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 61778,
+ "train_examples": 50000,
+ "train_loss": 1024.7534498828124
+ },
+ {
+ "epoch": 159,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62169,
+ "train_examples": 50000,
+ "train_loss": 972.62020796875
+ },
+ {
+ "epoch": 160,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62560,
+ "train_examples": 50000,
+ "train_loss": 953.9492185546875
+ },
+ {
+ "epoch": 161,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62951,
+ "train_examples": 50000,
+ "train_loss": 929.1100657421875
+ },
+ {
+ "epoch": 162,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 63342,
+ "train_examples": 50000,
+ "train_loss": 911.06114390625
+ },
+ {
+ "epoch": 163,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 63733,
+ "train_examples": 50000,
+ "train_loss": 873.1208018945313
+ },
+ {
+ "epoch": 164,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64124,
+ "train_examples": 50000,
+ "train_loss": 873.9455255078125
+ },
+ {
+ "epoch": 165,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64515,
+ "train_examples": 50000,
+ "train_loss": 855.66897359375
+ },
+ {
+ "epoch": 166,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64906,
+ "train_examples": 50000,
+ "train_loss": 832.5822851953125
+ },
+ {
+ "epoch": 167,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 65297,
+ "train_examples": 50000,
+ "train_loss": 820.8964893359375
+ },
+ {
+ "epoch": 168,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 65688,
+ "train_examples": 50000,
+ "train_loss": 809.5474932226563
+ },
+ {
+ "epoch": 169,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66079,
+ "train_examples": 50000,
+ "train_loss": 785.56792296875
+ },
+ {
+ "epoch": 170,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66470,
+ "train_examples": 50000,
+ "train_loss": 760.2091385546875
+ },
+ {
+ "epoch": 171,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66861,
+ "train_examples": 50000,
+ "train_loss": 751.9481428515625
+ },
+ {
+ "epoch": 172,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 67252,
+ "train_examples": 50000,
+ "train_loss": 768.3446512304688
+ },
+ {
+ "epoch": 173,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 67643,
+ "train_examples": 50000,
+ "train_loss": 739.8024948632813
+ },
+ {
+ "epoch": 174,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68034,
+ "train_examples": 50000,
+ "train_loss": 721.0074110742188
+ },
+ {
+ "epoch": 175,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68425,
+ "train_examples": 50000,
+ "train_loss": 748.7323610351563
+ },
+ {
+ "epoch": 176,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68816,
+ "train_examples": 50000,
+ "train_loss": 701.091491328125
+ },
+ {
+ "epoch": 177,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69207,
+ "train_examples": 50000,
+ "train_loss": 703.9146436914062
+ },
+ {
+ "epoch": 178,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69598,
+ "train_examples": 50000,
+ "train_loss": 720.5603126953125
+ },
+ {
+ "epoch": 179,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69989,
+ "train_examples": 50000,
+ "train_loss": 660.6486718359375
+ },
+ {
+ "epoch": 180,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 70380,
+ "train_examples": 50000,
+ "train_loss": 693.0599039257812
+ },
+ {
+ "epoch": 181,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 70771,
+ "train_examples": 50000,
+ "train_loss": 679.6050608203125
+ },
+ {
+ "epoch": 182,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71162,
+ "train_examples": 50000,
+ "train_loss": 692.9751318359375
+ },
+ {
+ "epoch": 183,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71553,
+ "train_examples": 50000,
+ "train_loss": 683.6897475585937
+ },
+ {
+ "epoch": 184,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71944,
+ "train_examples": 50000,
+ "train_loss": 668.1460813085937
+ },
+ {
+ "epoch": 185,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 72335,
+ "train_examples": 50000,
+ "train_loss": 667.1369373828124
+ },
+ {
+ "epoch": 186,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 72726,
+ "train_examples": 50000,
+ "train_loss": 662.4229653515625
+ },
+ {
+ "epoch": 187,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73117,
+ "train_examples": 50000,
+ "train_loss": 699.9143864453125
+ },
+ {
+ "epoch": 188,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73508,
+ "train_examples": 50000,
+ "train_loss": 630.793762109375
+ },
+ {
+ "epoch": 189,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73899,
+ "train_examples": 50000,
+ "train_loss": 675.000643828125
+ },
+ {
+ "epoch": 190,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 74290,
+ "train_examples": 50000,
+ "train_loss": 607.7093490722656
+ },
+ {
+ "epoch": 191,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 74681,
+ "train_examples": 50000,
+ "train_loss": 658.6411365429688
+ },
+ {
+ "epoch": 192,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75072,
+ "train_examples": 50000,
+ "train_loss": 657.6614402148438
+ },
+ {
+ "epoch": 193,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75463,
+ "train_examples": 50000,
+ "train_loss": 666.3172574804687
+ },
+ {
+ "epoch": 194,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75854,
+ "train_examples": 50000,
+ "train_loss": 655.2991161328125
+ },
+ {
+ "epoch": 195,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 76245,
+ "train_examples": 50000,
+ "train_loss": 624.0942075195312
+ },
+ {
+ "epoch": 196,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 76636,
+ "train_examples": 50000,
+ "train_loss": 620.8854083203125
+ },
+ {
+ "epoch": 197,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77027,
+ "train_examples": 50000,
+ "train_loss": 639.8727234960937
+ },
+ {
+ "epoch": 198,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77418,
+ "train_examples": 50000,
+ "train_loss": 638.6959756640625
+ },
+ {
+ "epoch": 199,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77809,
+ "train_examples": 50000,
+ "train_loss": 617.4043270703125
+ },
+ {
+ "epoch": 200,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 78200,
+ "train_examples": 50000,
+ "train_loss": 615.680954453125
+ }
+ ],
+ "evaluation_protocol": {
+ "test_evaluations": 1,
+ "test_used_for_selection": false,
+ "validation_evaluations": 0
+ },
+ "final": {
+ "accuracy": 0.3488,
+ "epoch": 200,
+ "evaluation_split": "test",
+ "finite": true,
+ "loss": 502.805196875,
+ "step": 78200
+ },
+ "hardware": {
+ "cuda_device_name": "NVIDIA GeForce GTX 1080",
+ "cuda_visible_devices": "7",
+ "device": "cuda",
+ "device_total_memory_bytes": 8507949056,
+ "peak_memory_allocated_bytes": 3339442688,
+ "peak_memory_reserved_bytes": 4676648960,
+ "torch_version": "2.3.1+cu118"
+ },
+ "protocol_family": "oral_a_cifar_local_resnet_development",
+ "provenance": {
+ "git_commit": "8f9364e8ee2e2a19eb83b58452fa0e8cb155353c",
+ "git_tracked_dirty": false
+ },
+ "schema_version": 1,
+ "split": {
+ "cifar_source_files": [
+ {
+ "bytes": 31035704,
+ "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_1",
+ "sha256": "54636561a3ce25bd3e19253c6b0d8538147b0ae398331ac4a2d86c6d987368cd"
+ },
+ {
+ "bytes": 31035320,
+ "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_2",
+ "sha256": "766b2cef9fbc745cf056b3152224f7cf77163b330ea9a15f9392beb8b89bc5a8"
+ },
+ {
+ "bytes": 31035999,
+ "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_3",
+ "sha256": "0f00d98ebfb30b3ec0ad19f9756dc2630b89003e10525f5e148445e82aa6a1f9"
+ },
+ {
+ "bytes": 31035696,
+ "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_4",
+ "sha256": "3f7bb240661948b8f4d53e36ec720d8306f5668bd0071dcb4e6c947f78e9682b"
+ },
+ {
+ "bytes": 31035623,
+ "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_5",
+ "sha256": "d91802434d8376bbaeeadf58a737e3a1b12ac839077e931237e0dcd43adcb154"
+ },
+ {
+ "bytes": 31035526,
+ "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/test_batch",
+ "sha256": "f53d8d457504f7cff4ea9e021afcf0e0ad8e24a91f3fc42091b8adef61157831"
+ }
+ ],
+ "dataset": "cifar10",
+ "evaluation_split": "test",
+ "input_layout": "NCHW",
+ "input_shape": [
+ 3,
+ 32,
+ 32
+ ],
+ "loader_seed": 13,
+ "normalization_mean": [
+ 0.49140000343322754,
+ 0.4821999967098236,
+ 0.4465000033378601
+ ],
+ "normalization_std": [
+ 0.24699999392032623,
+ 0.2434999942779541,
+ 0.26159998774528503
+ ],
+ "split_from_training_only": true,
+ "split_seed": null,
+ "test_examples": 10000,
+ "train_examples": 50000,
+ "training_augmentation": "random_crop_32_padding_4_zero_then_horizontal_flip_p0.5",
+ "validation_class_counts": {},
+ "validation_examples": 0,
+ "validation_index_sha256": null
+ },
+ "timing": {
+ "apical_warmup_wall_s": 0.0,
+ "evaluation_wall_s": 2.0348587036132812,
+ "mirror_warmup_wall_s": 0.0,
+ "predictor_warmup_wall_s": 0.0,
+ "timing_excludes_data_loading_hashing_and_model_construction": true,
+ "total_timed_wall_s": 10378.02499294281,
+ "train_wall_s": 10375.962883234024
+ },
+ "work": {
+ "apical_macs_per_example": 5324800,
+ "causal_scalar_observations": 0,
+ "components": {
+ "apical_projection_macs": 53248000000000,
+ "apical_regression_macs": 0,
+ "bp_reverse_macs_estimate": 0,
+ "kp_reciprocal_correlation_macs": 0,
+ "local_weight_correlation_macs": 1254856960000000,
+ "mirror_feedback_prediction_macs": 0,
+ "mirror_local_correlation_macs": 0,
+ "mirror_response_macs": 0,
+ "ordinary_forward_macs": 1254856960000000,
+ "perturbation_forward_macs": 0,
+ "warmup_clean_forward_macs": 0
+ },
+ "definition": "multiply-accumulates in conv/linear maps; one local weight correlation equals one forward-weight MAC count; BP reverse is estimated as one weight-gradient plus one activation-gradient convolution per forward convolution; mixed-traffic/predictor elementwise arithmetic is reported as a conservative operation estimate separately and is not folded into MACs",
+ "elementwise_operations_estimate": 0,
+ "forward_macs_per_example": 125485696,
+ "logical_batch_loss_queries": 0,
+ "mirror_probe_examples": 0,
+ "mirror_readout_probe_examples": 0,
+ "neutral_projection_observations": 0,
+ "per_example_cross_entropy_terms": 0,
+ "total_clean_forward_examples": 10000000,
+ "total_forward_equivalent_examples": 10000000,
+ "total_macs_estimate": 2562961920000000
+ }
+}