summaryrefslogtreecommitdiff
path: root/results/oral_a_dynamic_scaling_v2/dfa_d20_s11.json
diff options
context:
space:
mode:
Diffstat (limited to 'results/oral_a_dynamic_scaling_v2/dfa_d20_s11.json')
-rw-r--r--results/oral_a_dynamic_scaling_v2/dfa_d20_s11.json1993
1 files changed, 1993 insertions, 0 deletions
diff --git a/results/oral_a_dynamic_scaling_v2/dfa_d20_s11.json b/results/oral_a_dynamic_scaling_v2/dfa_d20_s11.json
new file mode 100644
index 0000000..398e9b6
--- /dev/null
+++ b/results/oral_a_dynamic_scaling_v2/dfa_d20_s11.json
@@ -0,0 +1,1993 @@
+{
+ "architecture": {
+ "adaptive_apical_parameters": 2260992,
+ "adaptive_feedback_parameters": 0,
+ "base_width": 16,
+ "blocks_per_stage": 3,
+ "bn_eps": 1e-05,
+ "bn_momentum": 0.1,
+ "depth": 20,
+ "family": "CIFAR 6n+2 ResNet, option-A shortcuts",
+ "fixed_feedback_parameters": 0,
+ "fixed_traffic_coefficients": 188416,
+ "forward_parameters": 269722,
+ "hidden_shapes": [
+ [
+ 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
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ],
+ [
+ 64,
+ 8,
+ 8
+ ]
+ ],
+ "normalization": "batchnorm",
+ "predictor_parameters": 376832,
+ "residual_scale": 1.0,
+ "vectorizer_mode": "spatial_template",
+ "vectorizer_parameters": 1884160
+ },
+ "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": 20,
+ "device": "cuda",
+ "epochs": 200,
+ "eta_A": 0.01,
+ "eta_P": 0.01,
+ "eval_every": 0,
+ "eval_split": "test",
+ "learn_P": 0,
+ "loader_seed": 11,
+ "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_d20_s11.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": 11,
+ "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": 0.0003854467940982431,
+ "innovation_negative_gradient_cosine": [
+ -0.0010154233314096928,
+ 0.0011164090828970075,
+ 0.0018438097322359681,
+ 0.001890377956442535,
+ -0.0010229444596916437,
+ -0.0004995482158847153,
+ 0.0019841650500893593,
+ 0.0077583095990121365,
+ 0.0021534061525017023,
+ 0.0015168748795986176,
+ 0.005302063189446926,
+ 0.0065992907620966434,
+ 0.004412676207721233,
+ 0.02175547368824482,
+ 0.022844579070806503,
+ 0.022070160135626793,
+ 0.029950225725769997,
+ 0.02953408658504486,
+ 0.039382897317409515
+ ],
+ "normalization_state": "training_batch_stats_without_running_update",
+ "raw_negative_gradient_cosine": [
+ -0.0010154233314096928,
+ 0.0011164090828970075,
+ 0.0018438097322359681,
+ 0.001890377956442535,
+ -0.0010229444596916437,
+ -0.0004995482158847153,
+ 0.0019841650500893593,
+ 0.0077583095990121365,
+ 0.0021534061525017023,
+ 0.0015168748795986176,
+ 0.005302063189446926,
+ 0.0065992907620966434,
+ 0.004412676207721233,
+ 0.02175547368824482,
+ 0.022844579070806503,
+ 0.022070160135626793,
+ 0.029950225725769997,
+ 0.02953408658504486,
+ 0.039382897317409515
+ ],
+ "teaching_negative_gradient_cosine": [
+ -0.0010154233314096928,
+ 0.0011164090828970075,
+ 0.0018438097322359681,
+ 0.001890377956442535,
+ -0.0010229444596916437,
+ -0.0004995482158847153,
+ 0.0019841650500893593,
+ 0.0077583095990121365,
+ 0.0021534061525017023,
+ 0.0015168748795986176,
+ 0.005302063189446926,
+ 0.0065992907620966434,
+ 0.004412676207721233,
+ 0.02175547368824482,
+ 0.022844579070806503,
+ 0.022070160135626793,
+ 0.029950225725769997,
+ 0.02953408658504486,
+ 0.039382897317409515
+ ],
+ "wall_s": 0.10302209854125977
+ },
+ "epochs": [
+ {
+ "epoch": 1,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 391,
+ "train_examples": 50000,
+ "train_loss": 2.5068225940704347
+ },
+ {
+ "epoch": 2,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 782,
+ "train_examples": 50000,
+ "train_loss": 3.515857922210693
+ },
+ {
+ "epoch": 3,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1173,
+ "train_examples": 50000,
+ "train_loss": 5.684407122192383
+ },
+ {
+ "epoch": 4,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1564,
+ "train_examples": 50000,
+ "train_loss": 8.086542698059082
+ },
+ {
+ "epoch": 5,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1955,
+ "train_examples": 50000,
+ "train_loss": 12.036429146728516
+ },
+ {
+ "epoch": 6,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 2346,
+ "train_examples": 50000,
+ "train_loss": 16.34205907470703
+ },
+ {
+ "epoch": 7,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 2737,
+ "train_examples": 50000,
+ "train_loss": 21.01639781677246
+ },
+ {
+ "epoch": 8,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3128,
+ "train_examples": 50000,
+ "train_loss": 27.02610977294922
+ },
+ {
+ "epoch": 9,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3519,
+ "train_examples": 50000,
+ "train_loss": 35.431440301513675
+ },
+ {
+ "epoch": 10,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3910,
+ "train_examples": 50000,
+ "train_loss": 41.92359442016602
+ },
+ {
+ "epoch": 11,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 4301,
+ "train_examples": 50000,
+ "train_loss": 50.71551621459961
+ },
+ {
+ "epoch": 12,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 4692,
+ "train_examples": 50000,
+ "train_loss": 59.637366508789064
+ },
+ {
+ "epoch": 13,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5083,
+ "train_examples": 50000,
+ "train_loss": 74.75878812622071
+ },
+ {
+ "epoch": 14,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5474,
+ "train_examples": 50000,
+ "train_loss": 85.2104129663086
+ },
+ {
+ "epoch": 15,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5865,
+ "train_examples": 50000,
+ "train_loss": 89.33222989990234
+ },
+ {
+ "epoch": 16,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 6256,
+ "train_examples": 50000,
+ "train_loss": 114.89599761596679
+ },
+ {
+ "epoch": 17,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 6647,
+ "train_examples": 50000,
+ "train_loss": 122.97920129638672
+ },
+ {
+ "epoch": 18,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7038,
+ "train_examples": 50000,
+ "train_loss": 137.3023895336914
+ },
+ {
+ "epoch": 19,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7429,
+ "train_examples": 50000,
+ "train_loss": 163.0888639453125
+ },
+ {
+ "epoch": 20,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7820,
+ "train_examples": 50000,
+ "train_loss": 180.36466465332032
+ },
+ {
+ "epoch": 21,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8211,
+ "train_examples": 50000,
+ "train_loss": 204.6404056591797
+ },
+ {
+ "epoch": 22,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8602,
+ "train_examples": 50000,
+ "train_loss": 218.40419555664062
+ },
+ {
+ "epoch": 23,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8993,
+ "train_examples": 50000,
+ "train_loss": 243.00024526367187
+ },
+ {
+ "epoch": 24,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 9384,
+ "train_examples": 50000,
+ "train_loss": 254.7752802685547
+ },
+ {
+ "epoch": 25,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 9775,
+ "train_examples": 50000,
+ "train_loss": 292.07824676757815
+ },
+ {
+ "epoch": 26,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10166,
+ "train_examples": 50000,
+ "train_loss": 329.8202194238281
+ },
+ {
+ "epoch": 27,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10557,
+ "train_examples": 50000,
+ "train_loss": 339.64070388671877
+ },
+ {
+ "epoch": 28,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10948,
+ "train_examples": 50000,
+ "train_loss": 399.37932737304686
+ },
+ {
+ "epoch": 29,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 11339,
+ "train_examples": 50000,
+ "train_loss": 428.4542909765625
+ },
+ {
+ "epoch": 30,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 11730,
+ "train_examples": 50000,
+ "train_loss": 410.7046250390625
+ },
+ {
+ "epoch": 31,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12121,
+ "train_examples": 50000,
+ "train_loss": 491.46051373046873
+ },
+ {
+ "epoch": 32,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12512,
+ "train_examples": 50000,
+ "train_loss": 485.348603984375
+ },
+ {
+ "epoch": 33,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12903,
+ "train_examples": 50000,
+ "train_loss": 511.99018515625
+ },
+ {
+ "epoch": 34,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 13294,
+ "train_examples": 50000,
+ "train_loss": 570.5548893359374
+ },
+ {
+ "epoch": 35,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 13685,
+ "train_examples": 50000,
+ "train_loss": 599.0122534863282
+ },
+ {
+ "epoch": 36,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14076,
+ "train_examples": 50000,
+ "train_loss": 638.6346241601562
+ },
+ {
+ "epoch": 37,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14467,
+ "train_examples": 50000,
+ "train_loss": 655.8679623828125
+ },
+ {
+ "epoch": 38,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14858,
+ "train_examples": 50000,
+ "train_loss": 732.1800444726563
+ },
+ {
+ "epoch": 39,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 15249,
+ "train_examples": 50000,
+ "train_loss": 793.89914203125
+ },
+ {
+ "epoch": 40,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 15640,
+ "train_examples": 50000,
+ "train_loss": 770.4782406445313
+ },
+ {
+ "epoch": 41,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16031,
+ "train_examples": 50000,
+ "train_loss": 838.25508484375
+ },
+ {
+ "epoch": 42,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16422,
+ "train_examples": 50000,
+ "train_loss": 868.5379801171875
+ },
+ {
+ "epoch": 43,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16813,
+ "train_examples": 50000,
+ "train_loss": 977.8729441796875
+ },
+ {
+ "epoch": 44,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17204,
+ "train_examples": 50000,
+ "train_loss": 943.8888183007813
+ },
+ {
+ "epoch": 45,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17595,
+ "train_examples": 50000,
+ "train_loss": 1085.8748257226562
+ },
+ {
+ "epoch": 46,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17986,
+ "train_examples": 50000,
+ "train_loss": 1121.91362421875
+ },
+ {
+ "epoch": 47,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 18377,
+ "train_examples": 50000,
+ "train_loss": 1149.345110078125
+ },
+ {
+ "epoch": 48,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 18768,
+ "train_examples": 50000,
+ "train_loss": 1268.7161190429688
+ },
+ {
+ "epoch": 49,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19159,
+ "train_examples": 50000,
+ "train_loss": 1291.9986146484375
+ },
+ {
+ "epoch": 50,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19550,
+ "train_examples": 50000,
+ "train_loss": 1284.3497418359375
+ },
+ {
+ "epoch": 51,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19941,
+ "train_examples": 50000,
+ "train_loss": 1399.3627420703126
+ },
+ {
+ "epoch": 52,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 20332,
+ "train_examples": 50000,
+ "train_loss": 1612.1165806640624
+ },
+ {
+ "epoch": 53,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 20723,
+ "train_examples": 50000,
+ "train_loss": 1648.3374513671874
+ },
+ {
+ "epoch": 54,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21114,
+ "train_examples": 50000,
+ "train_loss": 1674.036724765625
+ },
+ {
+ "epoch": 55,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21505,
+ "train_examples": 50000,
+ "train_loss": 1560.2893973828125
+ },
+ {
+ "epoch": 56,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21896,
+ "train_examples": 50000,
+ "train_loss": 1711.9624044921875
+ },
+ {
+ "epoch": 57,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 22287,
+ "train_examples": 50000,
+ "train_loss": 1712.132676171875
+ },
+ {
+ "epoch": 58,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 22678,
+ "train_examples": 50000,
+ "train_loss": 2055.0074187890623
+ },
+ {
+ "epoch": 59,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23069,
+ "train_examples": 50000,
+ "train_loss": 1829.655229375
+ },
+ {
+ "epoch": 60,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23460,
+ "train_examples": 50000,
+ "train_loss": 2050.5391868359375
+ },
+ {
+ "epoch": 61,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23851,
+ "train_examples": 50000,
+ "train_loss": 2019.0894387890626
+ },
+ {
+ "epoch": 62,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 24242,
+ "train_examples": 50000,
+ "train_loss": 1942.0169362109375
+ },
+ {
+ "epoch": 63,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 24633,
+ "train_examples": 50000,
+ "train_loss": 2274.977286015625
+ },
+ {
+ "epoch": 64,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25024,
+ "train_examples": 50000,
+ "train_loss": 2283.78664203125
+ },
+ {
+ "epoch": 65,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25415,
+ "train_examples": 50000,
+ "train_loss": 2264.264823359375
+ },
+ {
+ "epoch": 66,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25806,
+ "train_examples": 50000,
+ "train_loss": 2488.3162587109373
+ },
+ {
+ "epoch": 67,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26197,
+ "train_examples": 50000,
+ "train_loss": 2646.7587425
+ },
+ {
+ "epoch": 68,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26588,
+ "train_examples": 50000,
+ "train_loss": 2517.2651884375
+ },
+ {
+ "epoch": 69,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26979,
+ "train_examples": 50000,
+ "train_loss": 2730.358718671875
+ },
+ {
+ "epoch": 70,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 27370,
+ "train_examples": 50000,
+ "train_loss": 2989.476571875
+ },
+ {
+ "epoch": 71,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 27761,
+ "train_examples": 50000,
+ "train_loss": 2719.420979453125
+ },
+ {
+ "epoch": 72,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28152,
+ "train_examples": 50000,
+ "train_loss": 2725.74516234375
+ },
+ {
+ "epoch": 73,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28543,
+ "train_examples": 50000,
+ "train_loss": 3095.27936921875
+ },
+ {
+ "epoch": 74,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28934,
+ "train_examples": 50000,
+ "train_loss": 3115.5268525
+ },
+ {
+ "epoch": 75,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 29325,
+ "train_examples": 50000,
+ "train_loss": 3076.67187625
+ },
+ {
+ "epoch": 76,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 29716,
+ "train_examples": 50000,
+ "train_loss": 2864.8534446875
+ },
+ {
+ "epoch": 77,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30107,
+ "train_examples": 50000,
+ "train_loss": 3407.74620171875
+ },
+ {
+ "epoch": 78,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30498,
+ "train_examples": 50000,
+ "train_loss": 3591.5484866796874
+ },
+ {
+ "epoch": 79,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30889,
+ "train_examples": 50000,
+ "train_loss": 3505.355758515625
+ },
+ {
+ "epoch": 80,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 31280,
+ "train_examples": 50000,
+ "train_loss": 3667.21988140625
+ },
+ {
+ "epoch": 81,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 31671,
+ "train_examples": 50000,
+ "train_loss": 3818.790492109375
+ },
+ {
+ "epoch": 82,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32062,
+ "train_examples": 50000,
+ "train_loss": 3990.26935640625
+ },
+ {
+ "epoch": 83,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32453,
+ "train_examples": 50000,
+ "train_loss": 3560.0794140625
+ },
+ {
+ "epoch": 84,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32844,
+ "train_examples": 50000,
+ "train_loss": 4143.06030046875
+ },
+ {
+ "epoch": 85,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 33235,
+ "train_examples": 50000,
+ "train_loss": 4049.86179421875
+ },
+ {
+ "epoch": 86,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 33626,
+ "train_examples": 50000,
+ "train_loss": 4096.51183484375
+ },
+ {
+ "epoch": 87,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34017,
+ "train_examples": 50000,
+ "train_loss": 4365.442724296875
+ },
+ {
+ "epoch": 88,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34408,
+ "train_examples": 50000,
+ "train_loss": 4307.64304
+ },
+ {
+ "epoch": 89,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34799,
+ "train_examples": 50000,
+ "train_loss": 4608.90662203125
+ },
+ {
+ "epoch": 90,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35190,
+ "train_examples": 50000,
+ "train_loss": 4627.5656590625
+ },
+ {
+ "epoch": 91,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35581,
+ "train_examples": 50000,
+ "train_loss": 5346.3826034375
+ },
+ {
+ "epoch": 92,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35972,
+ "train_examples": 50000,
+ "train_loss": 4928.42728703125
+ },
+ {
+ "epoch": 93,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 36363,
+ "train_examples": 50000,
+ "train_loss": 4834.4236396875
+ },
+ {
+ "epoch": 94,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 36754,
+ "train_examples": 50000,
+ "train_loss": 4986.335101875
+ },
+ {
+ "epoch": 95,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37145,
+ "train_examples": 50000,
+ "train_loss": 5177.383709609375
+ },
+ {
+ "epoch": 96,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37536,
+ "train_examples": 50000,
+ "train_loss": 5380.46290078125
+ },
+ {
+ "epoch": 97,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37927,
+ "train_examples": 50000,
+ "train_loss": 5570.4661796875
+ },
+ {
+ "epoch": 98,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 38318,
+ "train_examples": 50000,
+ "train_loss": 5715.67653625
+ },
+ {
+ "epoch": 99,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 38709,
+ "train_examples": 50000,
+ "train_loss": 5869.27433859375
+ },
+ {
+ "epoch": 100,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 39100,
+ "train_examples": 50000,
+ "train_loss": 5563.03640296875
+ },
+ {
+ "epoch": 101,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 39491,
+ "train_examples": 50000,
+ "train_loss": 1193.3901241796875
+ },
+ {
+ "epoch": 102,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 39882,
+ "train_examples": 50000,
+ "train_loss": 901.1035458203125
+ },
+ {
+ "epoch": 103,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 40273,
+ "train_examples": 50000,
+ "train_loss": 804.7639102734375
+ },
+ {
+ "epoch": 104,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 40664,
+ "train_examples": 50000,
+ "train_loss": 716.0620738085937
+ },
+ {
+ "epoch": 105,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41055,
+ "train_examples": 50000,
+ "train_loss": 647.3206634179687
+ },
+ {
+ "epoch": 106,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41446,
+ "train_examples": 50000,
+ "train_loss": 600.5981295703125
+ },
+ {
+ "epoch": 107,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41837,
+ "train_examples": 50000,
+ "train_loss": 555.0965403808593
+ },
+ {
+ "epoch": 108,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 42228,
+ "train_examples": 50000,
+ "train_loss": 531.3540760351563
+ },
+ {
+ "epoch": 109,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 42619,
+ "train_examples": 50000,
+ "train_loss": 524.3530202246094
+ },
+ {
+ "epoch": 110,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43010,
+ "train_examples": 50000,
+ "train_loss": 471.2825883300781
+ },
+ {
+ "epoch": 111,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43401,
+ "train_examples": 50000,
+ "train_loss": 452.90929244140625
+ },
+ {
+ "epoch": 112,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43792,
+ "train_examples": 50000,
+ "train_loss": 438.8323709179688
+ },
+ {
+ "epoch": 113,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44183,
+ "train_examples": 50000,
+ "train_loss": 423.5165996386719
+ },
+ {
+ "epoch": 114,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44574,
+ "train_examples": 50000,
+ "train_loss": 426.702331640625
+ },
+ {
+ "epoch": 115,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44965,
+ "train_examples": 50000,
+ "train_loss": 423.29922956054685
+ },
+ {
+ "epoch": 116,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 45356,
+ "train_examples": 50000,
+ "train_loss": 404.40631328125
+ },
+ {
+ "epoch": 117,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 45747,
+ "train_examples": 50000,
+ "train_loss": 428.2123980078125
+ },
+ {
+ "epoch": 118,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46138,
+ "train_examples": 50000,
+ "train_loss": 400.1490480859375
+ },
+ {
+ "epoch": 119,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46529,
+ "train_examples": 50000,
+ "train_loss": 434.12222455078125
+ },
+ {
+ "epoch": 120,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46920,
+ "train_examples": 50000,
+ "train_loss": 412.13692862304686
+ },
+ {
+ "epoch": 121,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 47311,
+ "train_examples": 50000,
+ "train_loss": 424.259929921875
+ },
+ {
+ "epoch": 122,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 47702,
+ "train_examples": 50000,
+ "train_loss": 459.1871238867187
+ },
+ {
+ "epoch": 123,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48093,
+ "train_examples": 50000,
+ "train_loss": 402.66018516601565
+ },
+ {
+ "epoch": 124,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48484,
+ "train_examples": 50000,
+ "train_loss": 438.0382576464844
+ },
+ {
+ "epoch": 125,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48875,
+ "train_examples": 50000,
+ "train_loss": 437.2812224804687
+ },
+ {
+ "epoch": 126,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 49266,
+ "train_examples": 50000,
+ "train_loss": 451.2520609863281
+ },
+ {
+ "epoch": 127,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 49657,
+ "train_examples": 50000,
+ "train_loss": 421.0700023046875
+ },
+ {
+ "epoch": 128,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50048,
+ "train_examples": 50000,
+ "train_loss": 458.96626716796874
+ },
+ {
+ "epoch": 129,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50439,
+ "train_examples": 50000,
+ "train_loss": 412.06519220703126
+ },
+ {
+ "epoch": 130,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50830,
+ "train_examples": 50000,
+ "train_loss": 474.48770896484376
+ },
+ {
+ "epoch": 131,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 51221,
+ "train_examples": 50000,
+ "train_loss": 411.5285243847656
+ },
+ {
+ "epoch": 132,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 51612,
+ "train_examples": 50000,
+ "train_loss": 475.40324076171873
+ },
+ {
+ "epoch": 133,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52003,
+ "train_examples": 50000,
+ "train_loss": 433.22330268066406
+ },
+ {
+ "epoch": 134,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52394,
+ "train_examples": 50000,
+ "train_loss": 463.81109161132815
+ },
+ {
+ "epoch": 135,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52785,
+ "train_examples": 50000,
+ "train_loss": 449.84764244140626
+ },
+ {
+ "epoch": 136,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53176,
+ "train_examples": 50000,
+ "train_loss": 425.2446093261719
+ },
+ {
+ "epoch": 137,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53567,
+ "train_examples": 50000,
+ "train_loss": 507.0734819042969
+ },
+ {
+ "epoch": 138,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53958,
+ "train_examples": 50000,
+ "train_loss": 400.82680666015625
+ },
+ {
+ "epoch": 139,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 54349,
+ "train_examples": 50000,
+ "train_loss": 497.38916314453127
+ },
+ {
+ "epoch": 140,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 54740,
+ "train_examples": 50000,
+ "train_loss": 468.95010198242187
+ },
+ {
+ "epoch": 141,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55131,
+ "train_examples": 50000,
+ "train_loss": 465.1990760351562
+ },
+ {
+ "epoch": 142,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55522,
+ "train_examples": 50000,
+ "train_loss": 456.9041644140625
+ },
+ {
+ "epoch": 143,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55913,
+ "train_examples": 50000,
+ "train_loss": 470.43679358398435
+ },
+ {
+ "epoch": 144,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 56304,
+ "train_examples": 50000,
+ "train_loss": 462.1703258789062
+ },
+ {
+ "epoch": 145,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 56695,
+ "train_examples": 50000,
+ "train_loss": 444.6902553173828
+ },
+ {
+ "epoch": 146,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57086,
+ "train_examples": 50000,
+ "train_loss": 486.78014923828124
+ },
+ {
+ "epoch": 147,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57477,
+ "train_examples": 50000,
+ "train_loss": 454.7092383691406
+ },
+ {
+ "epoch": 148,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57868,
+ "train_examples": 50000,
+ "train_loss": 451.41220109375
+ },
+ {
+ "epoch": 149,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 58259,
+ "train_examples": 50000,
+ "train_loss": 450.31422676757813
+ },
+ {
+ "epoch": 150,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 58650,
+ "train_examples": 50000,
+ "train_loss": 467.62194791992187
+ },
+ {
+ "epoch": 151,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59041,
+ "train_examples": 50000,
+ "train_loss": 140.5188960180664
+ },
+ {
+ "epoch": 152,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59432,
+ "train_examples": 50000,
+ "train_loss": 122.81540368164063
+ },
+ {
+ "epoch": 153,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59823,
+ "train_examples": 50000,
+ "train_loss": 116.09231616699219
+ },
+ {
+ "epoch": 154,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60214,
+ "train_examples": 50000,
+ "train_loss": 107.69791808105468
+ },
+ {
+ "epoch": 155,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60605,
+ "train_examples": 50000,
+ "train_loss": 102.92118276855469
+ },
+ {
+ "epoch": 156,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60996,
+ "train_examples": 50000,
+ "train_loss": 97.91107972167968
+ },
+ {
+ "epoch": 157,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 61387,
+ "train_examples": 50000,
+ "train_loss": 94.57610254394531
+ },
+ {
+ "epoch": 158,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 61778,
+ "train_examples": 50000,
+ "train_loss": 91.08143123291016
+ },
+ {
+ "epoch": 159,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62169,
+ "train_examples": 50000,
+ "train_loss": 87.3799490209961
+ },
+ {
+ "epoch": 160,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62560,
+ "train_examples": 50000,
+ "train_loss": 83.78194811523437
+ },
+ {
+ "epoch": 161,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62951,
+ "train_examples": 50000,
+ "train_loss": 81.5281500390625
+ },
+ {
+ "epoch": 162,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 63342,
+ "train_examples": 50000,
+ "train_loss": 76.96467525634766
+ },
+ {
+ "epoch": 163,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 63733,
+ "train_examples": 50000,
+ "train_loss": 74.48964189941407
+ },
+ {
+ "epoch": 164,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64124,
+ "train_examples": 50000,
+ "train_loss": 74.4926614038086
+ },
+ {
+ "epoch": 165,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64515,
+ "train_examples": 50000,
+ "train_loss": 72.4107512121582
+ },
+ {
+ "epoch": 166,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64906,
+ "train_examples": 50000,
+ "train_loss": 69.0355158581543
+ },
+ {
+ "epoch": 167,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 65297,
+ "train_examples": 50000,
+ "train_loss": 69.6121717578125
+ },
+ {
+ "epoch": 168,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 65688,
+ "train_examples": 50000,
+ "train_loss": 66.27636252441407
+ },
+ {
+ "epoch": 169,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66079,
+ "train_examples": 50000,
+ "train_loss": 66.68412861816407
+ },
+ {
+ "epoch": 170,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66470,
+ "train_examples": 50000,
+ "train_loss": 66.87026070922852
+ },
+ {
+ "epoch": 171,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66861,
+ "train_examples": 50000,
+ "train_loss": 63.41542741210937
+ },
+ {
+ "epoch": 172,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 67252,
+ "train_examples": 50000,
+ "train_loss": 62.05930561645508
+ },
+ {
+ "epoch": 173,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 67643,
+ "train_examples": 50000,
+ "train_loss": 60.717706286621095
+ },
+ {
+ "epoch": 174,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68034,
+ "train_examples": 50000,
+ "train_loss": 60.02907170166016
+ },
+ {
+ "epoch": 175,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68425,
+ "train_examples": 50000,
+ "train_loss": 59.69362048095703
+ },
+ {
+ "epoch": 176,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68816,
+ "train_examples": 50000,
+ "train_loss": 62.324915122070315
+ },
+ {
+ "epoch": 177,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69207,
+ "train_examples": 50000,
+ "train_loss": 58.233677922363285
+ },
+ {
+ "epoch": 178,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69598,
+ "train_examples": 50000,
+ "train_loss": 55.237775908203126
+ },
+ {
+ "epoch": 179,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69989,
+ "train_examples": 50000,
+ "train_loss": 57.21391374755859
+ },
+ {
+ "epoch": 180,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 70380,
+ "train_examples": 50000,
+ "train_loss": 59.198869759521486
+ },
+ {
+ "epoch": 181,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 70771,
+ "train_examples": 50000,
+ "train_loss": 55.32839068359375
+ },
+ {
+ "epoch": 182,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71162,
+ "train_examples": 50000,
+ "train_loss": 54.57799655761719
+ },
+ {
+ "epoch": 183,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71553,
+ "train_examples": 50000,
+ "train_loss": 57.88844884033203
+ },
+ {
+ "epoch": 184,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71944,
+ "train_examples": 50000,
+ "train_loss": 56.9704599609375
+ },
+ {
+ "epoch": 185,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 72335,
+ "train_examples": 50000,
+ "train_loss": 53.511298244628904
+ },
+ {
+ "epoch": 186,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 72726,
+ "train_examples": 50000,
+ "train_loss": 52.950861934814455
+ },
+ {
+ "epoch": 187,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73117,
+ "train_examples": 50000,
+ "train_loss": 54.97057058349609
+ },
+ {
+ "epoch": 188,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73508,
+ "train_examples": 50000,
+ "train_loss": 57.945177227783205
+ },
+ {
+ "epoch": 189,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73899,
+ "train_examples": 50000,
+ "train_loss": 52.78738121826172
+ },
+ {
+ "epoch": 190,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 74290,
+ "train_examples": 50000,
+ "train_loss": 53.15324090576172
+ },
+ {
+ "epoch": 191,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 74681,
+ "train_examples": 50000,
+ "train_loss": 52.12365180419922
+ },
+ {
+ "epoch": 192,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75072,
+ "train_examples": 50000,
+ "train_loss": 52.32365452270508
+ },
+ {
+ "epoch": 193,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75463,
+ "train_examples": 50000,
+ "train_loss": 50.616903367919925
+ },
+ {
+ "epoch": 194,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75854,
+ "train_examples": 50000,
+ "train_loss": 53.57607133056641
+ },
+ {
+ "epoch": 195,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 76245,
+ "train_examples": 50000,
+ "train_loss": 53.6800094543457
+ },
+ {
+ "epoch": 196,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 76636,
+ "train_examples": 50000,
+ "train_loss": 51.921997236328124
+ },
+ {
+ "epoch": 197,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77027,
+ "train_examples": 50000,
+ "train_loss": 53.47778112060547
+ },
+ {
+ "epoch": 198,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77418,
+ "train_examples": 50000,
+ "train_loss": 51.79992364501953
+ },
+ {
+ "epoch": 199,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77809,
+ "train_examples": 50000,
+ "train_loss": 54.57040306396485
+ },
+ {
+ "epoch": 200,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 78200,
+ "train_examples": 50000,
+ "train_loss": 51.30855579345703
+ }
+ ],
+ "evaluation_protocol": {
+ "test_evaluations": 1,
+ "test_used_for_selection": false,
+ "validation_evaluations": 0
+ },
+ "final": {
+ "accuracy": 0.3374,
+ "epoch": 200,
+ "evaluation_split": "test",
+ "finite": true,
+ "loss": 37.194462109375,
+ "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": 2169915904,
+ "peak_memory_reserved_bytes": 2810183680,
+ "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": 11,
+ "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": 0.7038471698760986,
+ "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": 3643.292373418808,
+ "train_wall_s": 3642.5641095638275
+ },
+ "work": {
+ "apical_macs_per_example": 1884160,
+ "causal_scalar_observations": 0,
+ "components": {
+ "apical_projection_macs": 18841600000000,
+ "apical_regression_macs": 0,
+ "bp_reverse_macs_estimate": 0,
+ "kp_reciprocal_correlation_macs": 0,
+ "local_weight_correlation_macs": 405510400000000,
+ "mirror_feedback_prediction_macs": 0,
+ "mirror_local_correlation_macs": 0,
+ "mirror_response_macs": 0,
+ "ordinary_forward_macs": 405510400000000,
+ "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": 40551040,
+ "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": 829862400000000
+ }
+}
generated by cgit v1.2.3 (git 2.25.1) at 2026-08-01 00:53:03 +0000