summaryrefslogtreecommitdiff
path: root/results/oral_a_dynamic_scaling_v2/dfa_d32_s13.json
diff options
context:
space:
mode:
Diffstat (limited to 'results/oral_a_dynamic_scaling_v2/dfa_d32_s13.json')
-rw-r--r--results/oral_a_dynamic_scaling_v2/dfa_d32_s13.json2089
1 files changed, 2089 insertions, 0 deletions
diff --git a/results/oral_a_dynamic_scaling_v2/dfa_d32_s13.json b/results/oral_a_dynamic_scaling_v2/dfa_d32_s13.json
new file mode 100644
index 0000000..08910f4
--- /dev/null
+++ b/results/oral_a_dynamic_scaling_v2/dfa_d32_s13.json
@@ -0,0 +1,2089 @@
+{
+ "architecture": {
+ "adaptive_apical_parameters": 3637248,
+ "adaptive_feedback_parameters": 0,
+ "base_width": 16,
+ "blocks_per_stage": 5,
+ "bn_eps": 1e-05,
+ "bn_momentum": 0.1,
+ "depth": 32,
+ "family": "CIFAR 6n+2 ResNet, option-A shortcuts",
+ "fixed_feedback_parameters": 0,
+ "fixed_traffic_coefficients": 303104,
+ "forward_parameters": 464154,
+ "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
+ ],
+ [
+ 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
+ ]
+ ],
+ "normalization": "batchnorm",
+ "predictor_parameters": 606208,
+ "residual_scale": 1.0,
+ "vectorizer_mode": "spatial_template",
+ "vectorizer_parameters": 3031040
+ },
+ "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": 32,
+ "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_d32_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": 0.0007135697436751798,
+ "innovation_negative_gradient_cosine": [
+ 0.0006930246017873287,
+ 0.00025351723888888955,
+ -0.0003096466534771025,
+ 0.00023615991813130677,
+ 0.0005241597536951303,
+ -0.0010539605282247066,
+ 0.0004080611979588866,
+ 0.0015929663786664605,
+ 0.0022438704036176205,
+ 0.002547545125707984,
+ 0.0020522314589470625,
+ -0.0008863675757311285,
+ -0.0011268185917288065,
+ 0.004612869583070278,
+ 0.0021880711428821087,
+ 0.003208225592970848,
+ 0.004468095954507589,
+ 0.0043890392407774925,
+ 0.0015347551088780165,
+ 0.00540960393846035,
+ 0.0036303233355283737,
+ 0.01276873704046011,
+ 0.0168010201305151,
+ 0.01699742302298546,
+ 0.018613245338201523,
+ 0.01744714565575123,
+ 0.01885955035686493,
+ 0.01784776709973812,
+ 0.0240741278976202,
+ 0.01660817861557007,
+ 0.02493393048644066
+ ],
+ "normalization_state": "training_batch_stats_without_running_update",
+ "raw_negative_gradient_cosine": [
+ 0.0006930246017873287,
+ 0.00025351723888888955,
+ -0.0003096466534771025,
+ 0.00023615991813130677,
+ 0.0005241597536951303,
+ -0.0010539605282247066,
+ 0.0004080611979588866,
+ 0.0015929663786664605,
+ 0.0022438704036176205,
+ 0.002547545125707984,
+ 0.0020522314589470625,
+ -0.0008863675757311285,
+ -0.0011268185917288065,
+ 0.004612869583070278,
+ 0.0021880711428821087,
+ 0.003208225592970848,
+ 0.004468095954507589,
+ 0.0043890392407774925,
+ 0.0015347551088780165,
+ 0.00540960393846035,
+ 0.0036303233355283737,
+ 0.01276873704046011,
+ 0.0168010201305151,
+ 0.01699742302298546,
+ 0.018613245338201523,
+ 0.01744714565575123,
+ 0.01885955035686493,
+ 0.01784776709973812,
+ 0.0240741278976202,
+ 0.01660817861557007,
+ 0.02493393048644066
+ ],
+ "teaching_negative_gradient_cosine": [
+ 0.0006930246017873287,
+ 0.00025351723888888955,
+ -0.0003096466534771025,
+ 0.00023615991813130677,
+ 0.0005241597536951303,
+ -0.0010539605282247066,
+ 0.0004080611979588866,
+ 0.0015929663786664605,
+ 0.0022438704036176205,
+ 0.002547545125707984,
+ 0.0020522314589470625,
+ -0.0008863675757311285,
+ -0.0011268185917288065,
+ 0.004612869583070278,
+ 0.0021880711428821087,
+ 0.003208225592970848,
+ 0.004468095954507589,
+ 0.0043890392407774925,
+ 0.0015347551088780165,
+ 0.00540960393846035,
+ 0.0036303233355283737,
+ 0.01276873704046011,
+ 0.0168010201305151,
+ 0.01699742302298546,
+ 0.018613245338201523,
+ 0.01744714565575123,
+ 0.01885955035686493,
+ 0.01784776709973812,
+ 0.0240741278976202,
+ 0.01660817861557007,
+ 0.02493393048644066
+ ],
+ "wall_s": 0.1259937286376953
+ },
+ "epochs": [
+ {
+ "epoch": 1,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 391,
+ "train_examples": 50000,
+ "train_loss": 8.885662679138184
+ },
+ {
+ "epoch": 2,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 782,
+ "train_examples": 50000,
+ "train_loss": 8.873351764526367
+ },
+ {
+ "epoch": 3,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1173,
+ "train_examples": 50000,
+ "train_loss": 10.523940482177734
+ },
+ {
+ "epoch": 4,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1564,
+ "train_examples": 50000,
+ "train_loss": 14.166257419433594
+ },
+ {
+ "epoch": 5,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 1955,
+ "train_examples": 50000,
+ "train_loss": 17.697643399047852
+ },
+ {
+ "epoch": 6,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 2346,
+ "train_examples": 50000,
+ "train_loss": 22.953530103149415
+ },
+ {
+ "epoch": 7,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 2737,
+ "train_examples": 50000,
+ "train_loss": 33.861193767089844
+ },
+ {
+ "epoch": 8,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3128,
+ "train_examples": 50000,
+ "train_loss": 43.3194745703125
+ },
+ {
+ "epoch": 9,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3519,
+ "train_examples": 50000,
+ "train_loss": 54.50201330810547
+ },
+ {
+ "epoch": 10,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 3910,
+ "train_examples": 50000,
+ "train_loss": 71.80728986572265
+ },
+ {
+ "epoch": 11,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 4301,
+ "train_examples": 50000,
+ "train_loss": 84.83739865234375
+ },
+ {
+ "epoch": 12,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 4692,
+ "train_examples": 50000,
+ "train_loss": 116.18452405517579
+ },
+ {
+ "epoch": 13,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5083,
+ "train_examples": 50000,
+ "train_loss": 124.06628524902344
+ },
+ {
+ "epoch": 14,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5474,
+ "train_examples": 50000,
+ "train_loss": 138.6655530493164
+ },
+ {
+ "epoch": 15,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 5865,
+ "train_examples": 50000,
+ "train_loss": 179.30228564941407
+ },
+ {
+ "epoch": 16,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 6256,
+ "train_examples": 50000,
+ "train_loss": 202.8265821875
+ },
+ {
+ "epoch": 17,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 6647,
+ "train_examples": 50000,
+ "train_loss": 258.8684362011719
+ },
+ {
+ "epoch": 18,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7038,
+ "train_examples": 50000,
+ "train_loss": 246.25785841308593
+ },
+ {
+ "epoch": 19,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7429,
+ "train_examples": 50000,
+ "train_loss": 324.236358984375
+ },
+ {
+ "epoch": 20,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 7820,
+ "train_examples": 50000,
+ "train_loss": 376.75855177734377
+ },
+ {
+ "epoch": 21,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8211,
+ "train_examples": 50000,
+ "train_loss": 447.2875923535156
+ },
+ {
+ "epoch": 22,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8602,
+ "train_examples": 50000,
+ "train_loss": 464.5191718359375
+ },
+ {
+ "epoch": 23,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 8993,
+ "train_examples": 50000,
+ "train_loss": 482.16474328125
+ },
+ {
+ "epoch": 24,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 9384,
+ "train_examples": 50000,
+ "train_loss": 585.5187628710937
+ },
+ {
+ "epoch": 25,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 9775,
+ "train_examples": 50000,
+ "train_loss": 692.00428421875
+ },
+ {
+ "epoch": 26,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10166,
+ "train_examples": 50000,
+ "train_loss": 681.547618046875
+ },
+ {
+ "epoch": 27,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10557,
+ "train_examples": 50000,
+ "train_loss": 825.729226953125
+ },
+ {
+ "epoch": 28,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 10948,
+ "train_examples": 50000,
+ "train_loss": 824.1751372460938
+ },
+ {
+ "epoch": 29,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 11339,
+ "train_examples": 50000,
+ "train_loss": 958.9309803320313
+ },
+ {
+ "epoch": 30,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 11730,
+ "train_examples": 50000,
+ "train_loss": 1065.8346053710939
+ },
+ {
+ "epoch": 31,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12121,
+ "train_examples": 50000,
+ "train_loss": 1085.8403963671874
+ },
+ {
+ "epoch": 32,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12512,
+ "train_examples": 50000,
+ "train_loss": 1181.5517649609376
+ },
+ {
+ "epoch": 33,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 12903,
+ "train_examples": 50000,
+ "train_loss": 1312.19564390625
+ },
+ {
+ "epoch": 34,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 13294,
+ "train_examples": 50000,
+ "train_loss": 1427.726493984375
+ },
+ {
+ "epoch": 35,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 13685,
+ "train_examples": 50000,
+ "train_loss": 1616.18884859375
+ },
+ {
+ "epoch": 36,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14076,
+ "train_examples": 50000,
+ "train_loss": 1858.750594609375
+ },
+ {
+ "epoch": 37,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14467,
+ "train_examples": 50000,
+ "train_loss": 1807.61567953125
+ },
+ {
+ "epoch": 38,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 14858,
+ "train_examples": 50000,
+ "train_loss": 1790.74209265625
+ },
+ {
+ "epoch": 39,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 15249,
+ "train_examples": 50000,
+ "train_loss": 1975.14533515625
+ },
+ {
+ "epoch": 40,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 15640,
+ "train_examples": 50000,
+ "train_loss": 2349.380251484375
+ },
+ {
+ "epoch": 41,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16031,
+ "train_examples": 50000,
+ "train_loss": 2143.0995401171876
+ },
+ {
+ "epoch": 42,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16422,
+ "train_examples": 50000,
+ "train_loss": 2357.918343125
+ },
+ {
+ "epoch": 43,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 16813,
+ "train_examples": 50000,
+ "train_loss": 2556.11945171875
+ },
+ {
+ "epoch": 44,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17204,
+ "train_examples": 50000,
+ "train_loss": 2527.44374125
+ },
+ {
+ "epoch": 45,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17595,
+ "train_examples": 50000,
+ "train_loss": 2824.963735
+ },
+ {
+ "epoch": 46,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 17986,
+ "train_examples": 50000,
+ "train_loss": 3050.7419096875
+ },
+ {
+ "epoch": 47,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 18377,
+ "train_examples": 50000,
+ "train_loss": 2988.132359765625
+ },
+ {
+ "epoch": 48,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 18768,
+ "train_examples": 50000,
+ "train_loss": 3266.4915703125
+ },
+ {
+ "epoch": 49,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19159,
+ "train_examples": 50000,
+ "train_loss": 3583.411778125
+ },
+ {
+ "epoch": 50,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19550,
+ "train_examples": 50000,
+ "train_loss": 3418.094278125
+ },
+ {
+ "epoch": 51,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 19941,
+ "train_examples": 50000,
+ "train_loss": 3615.1833865625
+ },
+ {
+ "epoch": 52,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 20332,
+ "train_examples": 50000,
+ "train_loss": 4074.670904375
+ },
+ {
+ "epoch": 53,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 20723,
+ "train_examples": 50000,
+ "train_loss": 4005.102707578125
+ },
+ {
+ "epoch": 54,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21114,
+ "train_examples": 50000,
+ "train_loss": 4052.14866265625
+ },
+ {
+ "epoch": 55,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21505,
+ "train_examples": 50000,
+ "train_loss": 4113.71191921875
+ },
+ {
+ "epoch": 56,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 21896,
+ "train_examples": 50000,
+ "train_loss": 4656.29436578125
+ },
+ {
+ "epoch": 57,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 22287,
+ "train_examples": 50000,
+ "train_loss": 4656.771618515625
+ },
+ {
+ "epoch": 58,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 22678,
+ "train_examples": 50000,
+ "train_loss": 5151.35133859375
+ },
+ {
+ "epoch": 59,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23069,
+ "train_examples": 50000,
+ "train_loss": 5102.476401171875
+ },
+ {
+ "epoch": 60,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23460,
+ "train_examples": 50000,
+ "train_loss": 5173.2870353125
+ },
+ {
+ "epoch": 61,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 23851,
+ "train_examples": 50000,
+ "train_loss": 5772.5653415625
+ },
+ {
+ "epoch": 62,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 24242,
+ "train_examples": 50000,
+ "train_loss": 5678.9386259375
+ },
+ {
+ "epoch": 63,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 24633,
+ "train_examples": 50000,
+ "train_loss": 6244.67910125
+ },
+ {
+ "epoch": 64,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25024,
+ "train_examples": 50000,
+ "train_loss": 6328.75233984375
+ },
+ {
+ "epoch": 65,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25415,
+ "train_examples": 50000,
+ "train_loss": 6564.571815625
+ },
+ {
+ "epoch": 66,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 25806,
+ "train_examples": 50000,
+ "train_loss": 6927.61532125
+ },
+ {
+ "epoch": 67,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26197,
+ "train_examples": 50000,
+ "train_loss": 7323.84527
+ },
+ {
+ "epoch": 68,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26588,
+ "train_examples": 50000,
+ "train_loss": 6918.079596875
+ },
+ {
+ "epoch": 69,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 26979,
+ "train_examples": 50000,
+ "train_loss": 8083.695476875
+ },
+ {
+ "epoch": 70,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 27370,
+ "train_examples": 50000,
+ "train_loss": 7319.2224225
+ },
+ {
+ "epoch": 71,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 27761,
+ "train_examples": 50000,
+ "train_loss": 8799.09626375
+ },
+ {
+ "epoch": 72,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28152,
+ "train_examples": 50000,
+ "train_loss": 8243.33843375
+ },
+ {
+ "epoch": 73,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28543,
+ "train_examples": 50000,
+ "train_loss": 9431.78561390625
+ },
+ {
+ "epoch": 74,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 28934,
+ "train_examples": 50000,
+ "train_loss": 8274.32926125
+ },
+ {
+ "epoch": 75,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 29325,
+ "train_examples": 50000,
+ "train_loss": 9977.9765209375
+ },
+ {
+ "epoch": 76,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 29716,
+ "train_examples": 50000,
+ "train_loss": 9190.4313346875
+ },
+ {
+ "epoch": 77,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30107,
+ "train_examples": 50000,
+ "train_loss": 9297.16496578125
+ },
+ {
+ "epoch": 78,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30498,
+ "train_examples": 50000,
+ "train_loss": 9501.1392225
+ },
+ {
+ "epoch": 79,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 30889,
+ "train_examples": 50000,
+ "train_loss": 10224.1858175
+ },
+ {
+ "epoch": 80,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 31280,
+ "train_examples": 50000,
+ "train_loss": 9808.345829375
+ },
+ {
+ "epoch": 81,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 31671,
+ "train_examples": 50000,
+ "train_loss": 10950.676543125
+ },
+ {
+ "epoch": 82,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32062,
+ "train_examples": 50000,
+ "train_loss": 10793.331165625
+ },
+ {
+ "epoch": 83,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32453,
+ "train_examples": 50000,
+ "train_loss": 10753.8077559375
+ },
+ {
+ "epoch": 84,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 32844,
+ "train_examples": 50000,
+ "train_loss": 11647.875028125
+ },
+ {
+ "epoch": 85,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 33235,
+ "train_examples": 50000,
+ "train_loss": 10658.6340084375
+ },
+ {
+ "epoch": 86,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 33626,
+ "train_examples": 50000,
+ "train_loss": 11020.818275
+ },
+ {
+ "epoch": 87,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34017,
+ "train_examples": 50000,
+ "train_loss": 11799.7495428125
+ },
+ {
+ "epoch": 88,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34408,
+ "train_examples": 50000,
+ "train_loss": 12713.84546875
+ },
+ {
+ "epoch": 89,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 34799,
+ "train_examples": 50000,
+ "train_loss": 11942.6876278125
+ },
+ {
+ "epoch": 90,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35190,
+ "train_examples": 50000,
+ "train_loss": 12927.948485
+ },
+ {
+ "epoch": 91,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35581,
+ "train_examples": 50000,
+ "train_loss": 12512.52836125
+ },
+ {
+ "epoch": 92,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 35972,
+ "train_examples": 50000,
+ "train_loss": 15423.95205234375
+ },
+ {
+ "epoch": 93,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 36363,
+ "train_examples": 50000,
+ "train_loss": 14134.58731625
+ },
+ {
+ "epoch": 94,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 36754,
+ "train_examples": 50000,
+ "train_loss": 14506.948000625
+ },
+ {
+ "epoch": 95,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37145,
+ "train_examples": 50000,
+ "train_loss": 13866.099086875
+ },
+ {
+ "epoch": 96,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37536,
+ "train_examples": 50000,
+ "train_loss": 13612.6851990625
+ },
+ {
+ "epoch": 97,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 37927,
+ "train_examples": 50000,
+ "train_loss": 15303.8284709375
+ },
+ {
+ "epoch": 98,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 38318,
+ "train_examples": 50000,
+ "train_loss": 15936.346960625
+ },
+ {
+ "epoch": 99,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 38709,
+ "train_examples": 50000,
+ "train_loss": 17292.638946875
+ },
+ {
+ "epoch": 100,
+ "lr": 0.03,
+ "output_lr": 0.1,
+ "step": 39100,
+ "train_examples": 50000,
+ "train_loss": 15361.335605625
+ },
+ {
+ "epoch": 101,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 39491,
+ "train_examples": 50000,
+ "train_loss": 3676.14055984375
+ },
+ {
+ "epoch": 102,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 39882,
+ "train_examples": 50000,
+ "train_loss": 2522.29870546875
+ },
+ {
+ "epoch": 103,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 40273,
+ "train_examples": 50000,
+ "train_loss": 2229.5096124609377
+ },
+ {
+ "epoch": 104,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 40664,
+ "train_examples": 50000,
+ "train_loss": 1996.593728828125
+ },
+ {
+ "epoch": 105,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41055,
+ "train_examples": 50000,
+ "train_loss": 1834.0741719140624
+ },
+ {
+ "epoch": 106,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41446,
+ "train_examples": 50000,
+ "train_loss": 1697.5649187109375
+ },
+ {
+ "epoch": 107,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 41837,
+ "train_examples": 50000,
+ "train_loss": 1596.7197078125
+ },
+ {
+ "epoch": 108,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 42228,
+ "train_examples": 50000,
+ "train_loss": 1525.304521328125
+ },
+ {
+ "epoch": 109,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 42619,
+ "train_examples": 50000,
+ "train_loss": 1435.291812734375
+ },
+ {
+ "epoch": 110,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43010,
+ "train_examples": 50000,
+ "train_loss": 1351.258381328125
+ },
+ {
+ "epoch": 111,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43401,
+ "train_examples": 50000,
+ "train_loss": 1353.9311638671875
+ },
+ {
+ "epoch": 112,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 43792,
+ "train_examples": 50000,
+ "train_loss": 1343.518666953125
+ },
+ {
+ "epoch": 113,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44183,
+ "train_examples": 50000,
+ "train_loss": 1229.7389730859375
+ },
+ {
+ "epoch": 114,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44574,
+ "train_examples": 50000,
+ "train_loss": 1295.62979515625
+ },
+ {
+ "epoch": 115,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 44965,
+ "train_examples": 50000,
+ "train_loss": 1245.0034944140625
+ },
+ {
+ "epoch": 116,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 45356,
+ "train_examples": 50000,
+ "train_loss": 1209.270209140625
+ },
+ {
+ "epoch": 117,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 45747,
+ "train_examples": 50000,
+ "train_loss": 1227.6047748828125
+ },
+ {
+ "epoch": 118,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46138,
+ "train_examples": 50000,
+ "train_loss": 1185.1334834765626
+ },
+ {
+ "epoch": 119,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46529,
+ "train_examples": 50000,
+ "train_loss": 1164.7048128515626
+ },
+ {
+ "epoch": 120,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 46920,
+ "train_examples": 50000,
+ "train_loss": 1205.078364296875
+ },
+ {
+ "epoch": 121,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 47311,
+ "train_examples": 50000,
+ "train_loss": 1218.0545451171874
+ },
+ {
+ "epoch": 122,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 47702,
+ "train_examples": 50000,
+ "train_loss": 1224.5220325
+ },
+ {
+ "epoch": 123,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48093,
+ "train_examples": 50000,
+ "train_loss": 1147.2890941796875
+ },
+ {
+ "epoch": 124,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48484,
+ "train_examples": 50000,
+ "train_loss": 1238.18367921875
+ },
+ {
+ "epoch": 125,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 48875,
+ "train_examples": 50000,
+ "train_loss": 1259.841901875
+ },
+ {
+ "epoch": 126,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 49266,
+ "train_examples": 50000,
+ "train_loss": 1158.6707849609375
+ },
+ {
+ "epoch": 127,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 49657,
+ "train_examples": 50000,
+ "train_loss": 1214.9299673632813
+ },
+ {
+ "epoch": 128,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50048,
+ "train_examples": 50000,
+ "train_loss": 1216.136162890625
+ },
+ {
+ "epoch": 129,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50439,
+ "train_examples": 50000,
+ "train_loss": 1050.0023735742188
+ },
+ {
+ "epoch": 130,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 50830,
+ "train_examples": 50000,
+ "train_loss": 1107.405067890625
+ },
+ {
+ "epoch": 131,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 51221,
+ "train_examples": 50000,
+ "train_loss": 1290.2015492382814
+ },
+ {
+ "epoch": 132,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 51612,
+ "train_examples": 50000,
+ "train_loss": 1164.56899234375
+ },
+ {
+ "epoch": 133,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52003,
+ "train_examples": 50000,
+ "train_loss": 1114.820466171875
+ },
+ {
+ "epoch": 134,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52394,
+ "train_examples": 50000,
+ "train_loss": 1163.6818910546874
+ },
+ {
+ "epoch": 135,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 52785,
+ "train_examples": 50000,
+ "train_loss": 1163.4179517578125
+ },
+ {
+ "epoch": 136,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53176,
+ "train_examples": 50000,
+ "train_loss": 1300.3529144921874
+ },
+ {
+ "epoch": 137,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53567,
+ "train_examples": 50000,
+ "train_loss": 1228.5953010546875
+ },
+ {
+ "epoch": 138,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 53958,
+ "train_examples": 50000,
+ "train_loss": 1156.040135859375
+ },
+ {
+ "epoch": 139,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 54349,
+ "train_examples": 50000,
+ "train_loss": 1103.5443741796876
+ },
+ {
+ "epoch": 140,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 54740,
+ "train_examples": 50000,
+ "train_loss": 1178.8367403515624
+ },
+ {
+ "epoch": 141,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55131,
+ "train_examples": 50000,
+ "train_loss": 1208.8990166210938
+ },
+ {
+ "epoch": 142,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55522,
+ "train_examples": 50000,
+ "train_loss": 1097.449611171875
+ },
+ {
+ "epoch": 143,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 55913,
+ "train_examples": 50000,
+ "train_loss": 1271.9026541796875
+ },
+ {
+ "epoch": 144,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 56304,
+ "train_examples": 50000,
+ "train_loss": 1110.8728445703125
+ },
+ {
+ "epoch": 145,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 56695,
+ "train_examples": 50000,
+ "train_loss": 1289.5882077734375
+ },
+ {
+ "epoch": 146,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57086,
+ "train_examples": 50000,
+ "train_loss": 1253.7520938671876
+ },
+ {
+ "epoch": 147,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57477,
+ "train_examples": 50000,
+ "train_loss": 1224.7363398046875
+ },
+ {
+ "epoch": 148,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 57868,
+ "train_examples": 50000,
+ "train_loss": 1056.43241515625
+ },
+ {
+ "epoch": 149,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 58259,
+ "train_examples": 50000,
+ "train_loss": 1340.4110383007812
+ },
+ {
+ "epoch": 150,
+ "lr": 0.003,
+ "output_lr": 0.010000000000000002,
+ "step": 58650,
+ "train_examples": 50000,
+ "train_loss": 1105.03471171875
+ },
+ {
+ "epoch": 151,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59041,
+ "train_examples": 50000,
+ "train_loss": 386.68540423828125
+ },
+ {
+ "epoch": 152,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59432,
+ "train_examples": 50000,
+ "train_loss": 334.39591255859375
+ },
+ {
+ "epoch": 153,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 59823,
+ "train_examples": 50000,
+ "train_loss": 313.3769979003906
+ },
+ {
+ "epoch": 154,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60214,
+ "train_examples": 50000,
+ "train_loss": 289.9292345800781
+ },
+ {
+ "epoch": 155,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60605,
+ "train_examples": 50000,
+ "train_loss": 275.50476001953126
+ },
+ {
+ "epoch": 156,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 60996,
+ "train_examples": 50000,
+ "train_loss": 264.6360742675781
+ },
+ {
+ "epoch": 157,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 61387,
+ "train_examples": 50000,
+ "train_loss": 253.27556083007812
+ },
+ {
+ "epoch": 158,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 61778,
+ "train_examples": 50000,
+ "train_loss": 241.3612371484375
+ },
+ {
+ "epoch": 159,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62169,
+ "train_examples": 50000,
+ "train_loss": 228.4257856933594
+ },
+ {
+ "epoch": 160,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62560,
+ "train_examples": 50000,
+ "train_loss": 219.23440580566407
+ },
+ {
+ "epoch": 161,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 62951,
+ "train_examples": 50000,
+ "train_loss": 214.07328321777345
+ },
+ {
+ "epoch": 162,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 63342,
+ "train_examples": 50000,
+ "train_loss": 206.0354009375
+ },
+ {
+ "epoch": 163,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 63733,
+ "train_examples": 50000,
+ "train_loss": 201.9939564453125
+ },
+ {
+ "epoch": 164,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64124,
+ "train_examples": 50000,
+ "train_loss": 194.39837044433594
+ },
+ {
+ "epoch": 165,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64515,
+ "train_examples": 50000,
+ "train_loss": 191.2468030761719
+ },
+ {
+ "epoch": 166,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 64906,
+ "train_examples": 50000,
+ "train_loss": 186.10323663574218
+ },
+ {
+ "epoch": 167,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 65297,
+ "train_examples": 50000,
+ "train_loss": 179.8349450683594
+ },
+ {
+ "epoch": 168,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 65688,
+ "train_examples": 50000,
+ "train_loss": 178.56441596191405
+ },
+ {
+ "epoch": 169,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66079,
+ "train_examples": 50000,
+ "train_loss": 175.0047825366211
+ },
+ {
+ "epoch": 170,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66470,
+ "train_examples": 50000,
+ "train_loss": 169.16444389648439
+ },
+ {
+ "epoch": 171,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 66861,
+ "train_examples": 50000,
+ "train_loss": 174.33617260253905
+ },
+ {
+ "epoch": 172,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 67252,
+ "train_examples": 50000,
+ "train_loss": 164.6393528076172
+ },
+ {
+ "epoch": 173,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 67643,
+ "train_examples": 50000,
+ "train_loss": 162.44749002441407
+ },
+ {
+ "epoch": 174,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68034,
+ "train_examples": 50000,
+ "train_loss": 167.82198079589844
+ },
+ {
+ "epoch": 175,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68425,
+ "train_examples": 50000,
+ "train_loss": 167.80451487304688
+ },
+ {
+ "epoch": 176,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 68816,
+ "train_examples": 50000,
+ "train_loss": 164.29974818847657
+ },
+ {
+ "epoch": 177,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69207,
+ "train_examples": 50000,
+ "train_loss": 164.52811025878907
+ },
+ {
+ "epoch": 178,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69598,
+ "train_examples": 50000,
+ "train_loss": 162.1312367163086
+ },
+ {
+ "epoch": 179,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 69989,
+ "train_examples": 50000,
+ "train_loss": 157.83586200195313
+ },
+ {
+ "epoch": 180,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 70380,
+ "train_examples": 50000,
+ "train_loss": 155.45913883789063
+ },
+ {
+ "epoch": 181,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 70771,
+ "train_examples": 50000,
+ "train_loss": 166.92303909179688
+ },
+ {
+ "epoch": 182,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71162,
+ "train_examples": 50000,
+ "train_loss": 159.72698168457032
+ },
+ {
+ "epoch": 183,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71553,
+ "train_examples": 50000,
+ "train_loss": 148.46945220458986
+ },
+ {
+ "epoch": 184,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 71944,
+ "train_examples": 50000,
+ "train_loss": 155.47847979492187
+ },
+ {
+ "epoch": 185,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 72335,
+ "train_examples": 50000,
+ "train_loss": 154.63424322265624
+ },
+ {
+ "epoch": 186,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 72726,
+ "train_examples": 50000,
+ "train_loss": 150.73987834960937
+ },
+ {
+ "epoch": 187,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73117,
+ "train_examples": 50000,
+ "train_loss": 156.83884181152345
+ },
+ {
+ "epoch": 188,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73508,
+ "train_examples": 50000,
+ "train_loss": 147.17977385742188
+ },
+ {
+ "epoch": 189,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 73899,
+ "train_examples": 50000,
+ "train_loss": 151.08661865722655
+ },
+ {
+ "epoch": 190,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 74290,
+ "train_examples": 50000,
+ "train_loss": 145.43321890625
+ },
+ {
+ "epoch": 191,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 74681,
+ "train_examples": 50000,
+ "train_loss": 156.03836974121094
+ },
+ {
+ "epoch": 192,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75072,
+ "train_examples": 50000,
+ "train_loss": 147.9519510644531
+ },
+ {
+ "epoch": 193,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75463,
+ "train_examples": 50000,
+ "train_loss": 145.4378096826172
+ },
+ {
+ "epoch": 194,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 75854,
+ "train_examples": 50000,
+ "train_loss": 146.60984659179687
+ },
+ {
+ "epoch": 195,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 76245,
+ "train_examples": 50000,
+ "train_loss": 152.19081804931642
+ },
+ {
+ "epoch": 196,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 76636,
+ "train_examples": 50000,
+ "train_loss": 144.10472280761718
+ },
+ {
+ "epoch": 197,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77027,
+ "train_examples": 50000,
+ "train_loss": 151.817136953125
+ },
+ {
+ "epoch": 198,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77418,
+ "train_examples": 50000,
+ "train_loss": 141.34361458496093
+ },
+ {
+ "epoch": 199,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 77809,
+ "train_examples": 50000,
+ "train_loss": 147.52483091308594
+ },
+ {
+ "epoch": 200,
+ "lr": 0.00030000000000000003,
+ "output_lr": 0.0010000000000000002,
+ "step": 78200,
+ "train_examples": 50000,
+ "train_loss": 138.9903475390625
+ }
+ ],
+ "evaluation_protocol": {
+ "test_evaluations": 1,
+ "test_used_for_selection": false,
+ "validation_evaluations": 0
+ },
+ "final": {
+ "accuracy": 0.2863,
+ "epoch": 200,
+ "evaluation_split": "test",
+ "finite": true,
+ "loss": 162.33229375,
+ "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": 2537912832,
+ "peak_memory_reserved_bytes": 3323985920,
+ "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": 1.145874261856079,
+ "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": 5897.761346578598,
+ "train_wall_s": 5896.593955278397
+ },
+ "work": {
+ "apical_macs_per_example": 3031040,
+ "causal_scalar_observations": 0,
+ "components": {
+ "apical_projection_macs": 30310400000000,
+ "apical_regression_macs": 0,
+ "bp_reverse_macs_estimate": 0,
+ "kp_reciprocal_correlation_macs": 0,
+ "local_weight_correlation_macs": 688625920000000,
+ "mirror_feedback_prediction_macs": 0,
+ "mirror_local_correlation_macs": 0,
+ "mirror_response_macs": 0,
+ "ordinary_forward_macs": 688625920000000,
+ "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": 68862592,
+ "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": 1407562240000000
+ }
+}
generated by cgit v1.2.3 (git 2.25.1) at 2026-08-01 00:51:12 +0000