results: confirm standard-depth SDIL scaling

1 files changed, 2106 insertions, 0 deletions

diff --git a/results/oral_a_dynamic_scaling_v2/bp_d56_s14.json b/results/oral_a_dynamic_scaling_v2/bp_d56_s14.json
new file mode 100644
index 0000000..f754d60
--- /dev/null
+++ b/results/oral_a_dynamic_scaling_v2/bp_d56_s14.json
@@ -0,0 +1,2106 @@
+{
+  "architecture": {
+    "adaptive_apical_parameters": 0,
+    "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": 0,
+    "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": 0,
+    "residual_scale": 1.0,
+    "vectorizer_mode": null,
+    "vectorizer_parameters": 0
+  },
+  "args": {
+    "a_scale": 1.0,
+    "a_warmup_steps": 0,
+    "alignment_probe": 0,
+    "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": 14,
+    "lr": 0.1,
+    "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": "bp",
+    "momentum": 0.9,
+    "neutral_projection": 0,
+    "normalization": "batchnorm",
+    "nuisance_scale": 0.0,
+    "out": "results/oral_a_dynamic_scaling_v2/bp_d56_s14.json",
+    "output_lr": null,
+    "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": 14,
+    "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": null,
+  "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": null,
+  "epochs": [
+    {
+      "epoch": 1,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 391,
+      "train_examples": 50000,
+      "train_loss": 2.1491518631744384
+    },
+    {
+      "epoch": 2,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 782,
+      "train_examples": 50000,
+      "train_loss": 1.5415329328155518
+    },
+    {
+      "epoch": 3,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 1173,
+      "train_examples": 50000,
+      "train_loss": 1.1902907227325439
+    },
+    {
+      "epoch": 4,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 1564,
+      "train_examples": 50000,
+      "train_loss": 0.9493342241477967
+    },
+    {
+      "epoch": 5,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 1955,
+      "train_examples": 50000,
+      "train_loss": 0.8143947232627868
+    },
+    {
+      "epoch": 6,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 2346,
+      "train_examples": 50000,
+      "train_loss": 0.733723431892395
+    },
+    {
+      "epoch": 7,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 2737,
+      "train_examples": 50000,
+      "train_loss": 0.6721547922706604
+    },
+    {
+      "epoch": 8,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 3128,
+      "train_examples": 50000,
+      "train_loss": 0.6167750704574585
+    },
+    {
+      "epoch": 9,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 3519,
+      "train_examples": 50000,
+      "train_loss": 0.5720778098678588
+    },
+    {
+      "epoch": 10,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 3910,
+      "train_examples": 50000,
+      "train_loss": 0.5399607049369812
+    },
+    {
+      "epoch": 11,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 4301,
+      "train_examples": 50000,
+      "train_loss": 0.5111437025451661
+    },
+    {
+      "epoch": 12,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 4692,
+      "train_examples": 50000,
+      "train_loss": 0.4847265879821777
+    },
+    {
+      "epoch": 13,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 5083,
+      "train_examples": 50000,
+      "train_loss": 0.4709283294296265
+    },
+    {
+      "epoch": 14,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 5474,
+      "train_examples": 50000,
+      "train_loss": 0.43931803150177
+    },
+    {
+      "epoch": 15,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 5865,
+      "train_examples": 50000,
+      "train_loss": 0.42153508222579955
+    },
+    {
+      "epoch": 16,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 6256,
+      "train_examples": 50000,
+      "train_loss": 0.4142340124320984
+    },
+    {
+      "epoch": 17,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 6647,
+      "train_examples": 50000,
+      "train_loss": 0.40057316608428956
+    },
+    {
+      "epoch": 18,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 7038,
+      "train_examples": 50000,
+      "train_loss": 0.3751462773513794
+    },
+    {
+      "epoch": 19,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 7429,
+      "train_examples": 50000,
+      "train_loss": 0.3673887727355957
+    },
+    {
+      "epoch": 20,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 7820,
+      "train_examples": 50000,
+      "train_loss": 0.35467099767684934
+    },
+    {
+      "epoch": 21,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 8211,
+      "train_examples": 50000,
+      "train_loss": 0.3524684285068512
+    },
+    {
+      "epoch": 22,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 8602,
+      "train_examples": 50000,
+      "train_loss": 0.34037952941894534
+    },
+    {
+      "epoch": 23,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 8993,
+      "train_examples": 50000,
+      "train_loss": 0.3340830976486206
+    },
+    {
+      "epoch": 24,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 9384,
+      "train_examples": 50000,
+      "train_loss": 0.3162274563121796
+    },
+    {
+      "epoch": 25,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 9775,
+      "train_examples": 50000,
+      "train_loss": 0.3063911257553101
+    },
+    {
+      "epoch": 26,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 10166,
+      "train_examples": 50000,
+      "train_loss": 0.3132950550365448
+    },
+    {
+      "epoch": 27,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 10557,
+      "train_examples": 50000,
+      "train_loss": 0.29966708919048307
+    },
+    {
+      "epoch": 28,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 10948,
+      "train_examples": 50000,
+      "train_loss": 0.29376793018817904
+    },
+    {
+      "epoch": 29,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 11339,
+      "train_examples": 50000,
+      "train_loss": 0.28038517884254455
+    },
+    {
+      "epoch": 30,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 11730,
+      "train_examples": 50000,
+      "train_loss": 0.2781383064031601
+    },
+    {
+      "epoch": 31,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 12121,
+      "train_examples": 50000,
+      "train_loss": 0.2694102143526077
+    },
+    {
+      "epoch": 32,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 12512,
+      "train_examples": 50000,
+      "train_loss": 0.2667882744693756
+    },
+    {
+      "epoch": 33,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 12903,
+      "train_examples": 50000,
+      "train_loss": 0.2611733724021912
+    },
+    {
+      "epoch": 34,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 13294,
+      "train_examples": 50000,
+      "train_loss": 0.25603877515792844
+    },
+    {
+      "epoch": 35,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 13685,
+      "train_examples": 50000,
+      "train_loss": 0.25416267645835877
+    },
+    {
+      "epoch": 36,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 14076,
+      "train_examples": 50000,
+      "train_loss": 0.24760894286632537
+    },
+    {
+      "epoch": 37,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 14467,
+      "train_examples": 50000,
+      "train_loss": 0.24410780872821808
+    },
+    {
+      "epoch": 38,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 14858,
+      "train_examples": 50000,
+      "train_loss": 0.24305856939315795
+    },
+    {
+      "epoch": 39,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 15249,
+      "train_examples": 50000,
+      "train_loss": 0.23772317601203918
+    },
+    {
+      "epoch": 40,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 15640,
+      "train_examples": 50000,
+      "train_loss": 0.2337888228416443
+    },
+    {
+      "epoch": 41,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 16031,
+      "train_examples": 50000,
+      "train_loss": 0.22836926259040832
+    },
+    {
+      "epoch": 42,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 16422,
+      "train_examples": 50000,
+      "train_loss": 0.22831247077941894
+    },
+    {
+      "epoch": 43,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 16813,
+      "train_examples": 50000,
+      "train_loss": 0.22156494766235352
+    },
+    {
+      "epoch": 44,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 17204,
+      "train_examples": 50000,
+      "train_loss": 0.215394432888031
+    },
+    {
+      "epoch": 45,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 17595,
+      "train_examples": 50000,
+      "train_loss": 0.21597371877670288
+    },
+    {
+      "epoch": 46,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 17986,
+      "train_examples": 50000,
+      "train_loss": 0.21680292412757873
+    },
+    {
+      "epoch": 47,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 18377,
+      "train_examples": 50000,
+      "train_loss": 0.2124243455505371
+    },
+    {
+      "epoch": 48,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 18768,
+      "train_examples": 50000,
+      "train_loss": 0.21217001245975495
+    },
+    {
+      "epoch": 49,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 19159,
+      "train_examples": 50000,
+      "train_loss": 0.207936761264801
+    },
+    {
+      "epoch": 50,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 19550,
+      "train_examples": 50000,
+      "train_loss": 0.20709203666687012
+    },
+    {
+      "epoch": 51,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 19941,
+      "train_examples": 50000,
+      "train_loss": 0.19351246401786804
+    },
+    {
+      "epoch": 52,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 20332,
+      "train_examples": 50000,
+      "train_loss": 0.20142732095718383
+    },
+    {
+      "epoch": 53,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 20723,
+      "train_examples": 50000,
+      "train_loss": 0.191461335067749
+    },
+    {
+      "epoch": 54,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 21114,
+      "train_examples": 50000,
+      "train_loss": 0.19068031177997588
+    },
+    {
+      "epoch": 55,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 21505,
+      "train_examples": 50000,
+      "train_loss": 0.1948518556499481
+    },
+    {
+      "epoch": 56,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 21896,
+      "train_examples": 50000,
+      "train_loss": 0.19615788020133973
+    },
+    {
+      "epoch": 57,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 22287,
+      "train_examples": 50000,
+      "train_loss": 0.19002282517433167
+    },
+    {
+      "epoch": 58,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 22678,
+      "train_examples": 50000,
+      "train_loss": 0.1905314666724205
+    },
+    {
+      "epoch": 59,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 23069,
+      "train_examples": 50000,
+      "train_loss": 0.18444560183048248
+    },
+    {
+      "epoch": 60,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 23460,
+      "train_examples": 50000,
+      "train_loss": 0.17953283669233322
+    },
+    {
+      "epoch": 61,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 23851,
+      "train_examples": 50000,
+      "train_loss": 0.17936252091407776
+    },
+    {
+      "epoch": 62,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 24242,
+      "train_examples": 50000,
+      "train_loss": 0.17894449977874755
+    },
+    {
+      "epoch": 63,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 24633,
+      "train_examples": 50000,
+      "train_loss": 0.18608696917533873
+    },
+    {
+      "epoch": 64,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 25024,
+      "train_examples": 50000,
+      "train_loss": 0.1867162022972107
+    },
+    {
+      "epoch": 65,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 25415,
+      "train_examples": 50000,
+      "train_loss": 0.17704015726566313
+    },
+    {
+      "epoch": 66,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 25806,
+      "train_examples": 50000,
+      "train_loss": 0.17832065382003784
+    },
+    {
+      "epoch": 67,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 26197,
+      "train_examples": 50000,
+      "train_loss": 0.1795563365650177
+    },
+    {
+      "epoch": 68,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 26588,
+      "train_examples": 50000,
+      "train_loss": 0.17364684770584107
+    },
+    {
+      "epoch": 69,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 26979,
+      "train_examples": 50000,
+      "train_loss": 0.17744720470428466
+    },
+    {
+      "epoch": 70,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 27370,
+      "train_examples": 50000,
+      "train_loss": 0.17062861320495606
+    },
+    {
+      "epoch": 71,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 27761,
+      "train_examples": 50000,
+      "train_loss": 0.17449909383296966
+    },
+    {
+      "epoch": 72,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 28152,
+      "train_examples": 50000,
+      "train_loss": 0.1682535918712616
+    },
+    {
+      "epoch": 73,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 28543,
+      "train_examples": 50000,
+      "train_loss": 0.17165554725170135
+    },
+    {
+      "epoch": 74,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 28934,
+      "train_examples": 50000,
+      "train_loss": 0.15960627580165862
+    },
+    {
+      "epoch": 75,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 29325,
+      "train_examples": 50000,
+      "train_loss": 0.16947615956306458
+    },
+    {
+      "epoch": 76,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 29716,
+      "train_examples": 50000,
+      "train_loss": 0.16465355763435363
+    },
+    {
+      "epoch": 77,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 30107,
+      "train_examples": 50000,
+      "train_loss": 0.15931276038169861
+    },
+    {
+      "epoch": 78,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 30498,
+      "train_examples": 50000,
+      "train_loss": 0.16503633592128752
+    },
+    {
+      "epoch": 79,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 30889,
+      "train_examples": 50000,
+      "train_loss": 0.16429143577575683
+    },
+    {
+      "epoch": 80,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 31280,
+      "train_examples": 50000,
+      "train_loss": 0.16562639446258545
+    },
+    {
+      "epoch": 81,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 31671,
+      "train_examples": 50000,
+      "train_loss": 0.16291448392152785
+    },
+    {
+      "epoch": 82,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 32062,
+      "train_examples": 50000,
+      "train_loss": 0.15886301523447036
+    },
+    {
+      "epoch": 83,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 32453,
+      "train_examples": 50000,
+      "train_loss": 0.160158472032547
+    },
+    {
+      "epoch": 84,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 32844,
+      "train_examples": 50000,
+      "train_loss": 0.1616283896446228
+    },
+    {
+      "epoch": 85,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 33235,
+      "train_examples": 50000,
+      "train_loss": 0.16116115434646605
+    },
+    {
+      "epoch": 86,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 33626,
+      "train_examples": 50000,
+      "train_loss": 0.15407067019462586
+    },
+    {
+      "epoch": 87,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 34017,
+      "train_examples": 50000,
+      "train_loss": 0.15864971215248108
+    },
+    {
+      "epoch": 88,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 34408,
+      "train_examples": 50000,
+      "train_loss": 0.15416404981136322
+    },
+    {
+      "epoch": 89,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 34799,
+      "train_examples": 50000,
+      "train_loss": 0.15308708516597747
+    },
+    {
+      "epoch": 90,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 35190,
+      "train_examples": 50000,
+      "train_loss": 0.1539277541399002
+    },
+    {
+      "epoch": 91,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 35581,
+      "train_examples": 50000,
+      "train_loss": 0.15563526938915254
+    },
+    {
+      "epoch": 92,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 35972,
+      "train_examples": 50000,
+      "train_loss": 0.15266645413398744
+    },
+    {
+      "epoch": 93,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 36363,
+      "train_examples": 50000,
+      "train_loss": 0.15498985040187835
+    },
+    {
+      "epoch": 94,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 36754,
+      "train_examples": 50000,
+      "train_loss": 0.14707652615547182
+    },
+    {
+      "epoch": 95,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 37145,
+      "train_examples": 50000,
+      "train_loss": 0.15589172424316405
+    },
+    {
+      "epoch": 96,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 37536,
+      "train_examples": 50000,
+      "train_loss": 0.1510642097234726
+    },
+    {
+      "epoch": 97,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 37927,
+      "train_examples": 50000,
+      "train_loss": 0.1477131044793129
+    },
+    {
+      "epoch": 98,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 38318,
+      "train_examples": 50000,
+      "train_loss": 0.1512059341239929
+    },
+    {
+      "epoch": 99,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 38709,
+      "train_examples": 50000,
+      "train_loss": 0.1547452706193924
+    },
+    {
+      "epoch": 100,
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 39100,
+      "train_examples": 50000,
+      "train_loss": 0.14908998366594314
+    },
+    {
+      "epoch": 101,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 39491,
+      "train_examples": 50000,
+      "train_loss": 0.08004976162791252
+    },
+    {
+      "epoch": 102,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 39882,
+      "train_examples": 50000,
+      "train_loss": 0.05500232973217964
+    },
+    {
+      "epoch": 103,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 40273,
+      "train_examples": 50000,
+      "train_loss": 0.04631398814439774
+    },
+    {
+      "epoch": 104,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 40664,
+      "train_examples": 50000,
+      "train_loss": 0.04183367156863212
+    },
+    {
+      "epoch": 105,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 41055,
+      "train_examples": 50000,
+      "train_loss": 0.03675910843372345
+    },
+    {
+      "epoch": 106,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 41446,
+      "train_examples": 50000,
+      "train_loss": 0.03236614109992981
+    },
+    {
+      "epoch": 107,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 41837,
+      "train_examples": 50000,
+      "train_loss": 0.029470725601911545
+    },
+    {
+      "epoch": 108,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 42228,
+      "train_examples": 50000,
+      "train_loss": 0.026720388967990876
+    },
+    {
+      "epoch": 109,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 42619,
+      "train_examples": 50000,
+      "train_loss": 0.02745675653517246
+    },
+    {
+      "epoch": 110,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 43010,
+      "train_examples": 50000,
+      "train_loss": 0.024266670229136943
+    },
+    {
+      "epoch": 111,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 43401,
+      "train_examples": 50000,
+      "train_loss": 0.02235375830113888
+    },
+    {
+      "epoch": 112,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 43792,
+      "train_examples": 50000,
+      "train_loss": 0.021399842903018
+    },
+    {
+      "epoch": 113,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 44183,
+      "train_examples": 50000,
+      "train_loss": 0.019970357822179794
+    },
+    {
+      "epoch": 114,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 44574,
+      "train_examples": 50000,
+      "train_loss": 0.018387679604291916
+    },
+    {
+      "epoch": 115,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 44965,
+      "train_examples": 50000,
+      "train_loss": 0.01710965284049511
+    },
+    {
+      "epoch": 116,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 45356,
+      "train_examples": 50000,
+      "train_loss": 0.016227416690587998
+    },
+    {
+      "epoch": 117,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 45747,
+      "train_examples": 50000,
+      "train_loss": 0.01590447195649147
+    },
+    {
+      "epoch": 118,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 46138,
+      "train_examples": 50000,
+      "train_loss": 0.015522427982091904
+    },
+    {
+      "epoch": 119,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 46529,
+      "train_examples": 50000,
+      "train_loss": 0.013863605805635452
+    },
+    {
+      "epoch": 120,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 46920,
+      "train_examples": 50000,
+      "train_loss": 0.014408985558599234
+    },
+    {
+      "epoch": 121,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 47311,
+      "train_examples": 50000,
+      "train_loss": 0.013013962758034467
+    },
+    {
+      "epoch": 122,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 47702,
+      "train_examples": 50000,
+      "train_loss": 0.01262190055757761
+    },
+    {
+      "epoch": 123,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 48093,
+      "train_examples": 50000,
+      "train_loss": 0.011193660405874252
+    },
+    {
+      "epoch": 124,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 48484,
+      "train_examples": 50000,
+      "train_loss": 0.012185472703874111
+    },
+    {
+      "epoch": 125,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 48875,
+      "train_examples": 50000,
+      "train_loss": 0.012042651894539595
+    },
+    {
+      "epoch": 126,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 49266,
+      "train_examples": 50000,
+      "train_loss": 0.01146743201315403
+    },
+    {
+      "epoch": 127,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 49657,
+      "train_examples": 50000,
+      "train_loss": 0.011375230247974396
+    },
+    {
+      "epoch": 128,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 50048,
+      "train_examples": 50000,
+      "train_loss": 0.010315139727853239
+    },
+    {
+      "epoch": 129,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 50439,
+      "train_examples": 50000,
+      "train_loss": 0.009391635978072882
+    },
+    {
+      "epoch": 130,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 50830,
+      "train_examples": 50000,
+      "train_loss": 0.009801293014287948
+    },
+    {
+      "epoch": 131,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 51221,
+      "train_examples": 50000,
+      "train_loss": 0.008947621556669474
+    },
+    {
+      "epoch": 132,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 51612,
+      "train_examples": 50000,
+      "train_loss": 0.008424072011113167
+    },
+    {
+      "epoch": 133,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 52003,
+      "train_examples": 50000,
+      "train_loss": 0.008684965382665395
+    },
+    {
+      "epoch": 134,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 52394,
+      "train_examples": 50000,
+      "train_loss": 0.00922296491727233
+    },
+    {
+      "epoch": 135,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 52785,
+      "train_examples": 50000,
+      "train_loss": 0.008562647834569215
+    },
+    {
+      "epoch": 136,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 53176,
+      "train_examples": 50000,
+      "train_loss": 0.008652239821851253
+    },
+    {
+      "epoch": 137,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 53567,
+      "train_examples": 50000,
+      "train_loss": 0.008561775663793087
+    },
+    {
+      "epoch": 138,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 53958,
+      "train_examples": 50000,
+      "train_loss": 0.00765712965130806
+    },
+    {
+      "epoch": 139,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 54349,
+      "train_examples": 50000,
+      "train_loss": 0.007651031436175108
+    },
+    {
+      "epoch": 140,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 54740,
+      "train_examples": 50000,
+      "train_loss": 0.007763555864691734
+    },
+    {
+      "epoch": 141,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 55131,
+      "train_examples": 50000,
+      "train_loss": 0.007720564398616552
+    },
+    {
+      "epoch": 142,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 55522,
+      "train_examples": 50000,
+      "train_loss": 0.006966621245443821
+    },
+    {
+      "epoch": 143,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 55913,
+      "train_examples": 50000,
+      "train_loss": 0.007150115132853389
+    },
+    {
+      "epoch": 144,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 56304,
+      "train_examples": 50000,
+      "train_loss": 0.007278579885214567
+    },
+    {
+      "epoch": 145,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 56695,
+      "train_examples": 50000,
+      "train_loss": 0.006301577565260232
+    },
+    {
+      "epoch": 146,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 57086,
+      "train_examples": 50000,
+      "train_loss": 0.006381783106327057
+    },
+    {
+      "epoch": 147,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 57477,
+      "train_examples": 50000,
+      "train_loss": 0.00685056243032217
+    },
+    {
+      "epoch": 148,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 57868,
+      "train_examples": 50000,
+      "train_loss": 0.006106687775440514
+    },
+    {
+      "epoch": 149,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 58259,
+      "train_examples": 50000,
+      "train_loss": 0.006147566365897655
+    },
+    {
+      "epoch": 150,
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 58650,
+      "train_examples": 50000,
+      "train_loss": 0.006381408163979649
+    },
+    {
+      "epoch": 151,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 59041,
+      "train_examples": 50000,
+      "train_loss": 0.005402126539442689
+    },
+    {
+      "epoch": 152,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 59432,
+      "train_examples": 50000,
+      "train_loss": 0.005967684311792254
+    },
+    {
+      "epoch": 153,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 59823,
+      "train_examples": 50000,
+      "train_loss": 0.004802315729185939
+    },
+    {
+      "epoch": 154,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 60214,
+      "train_examples": 50000,
+      "train_loss": 0.004821319012865424
+    },
+    {
+      "epoch": 155,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 60605,
+      "train_examples": 50000,
+      "train_loss": 0.005573809953629971
+    },
+    {
+      "epoch": 156,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 60996,
+      "train_examples": 50000,
+      "train_loss": 0.004940722731947899
+    },
+    {
+      "epoch": 157,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 61387,
+      "train_examples": 50000,
+      "train_loss": 0.004846924012377858
+    },
+    {
+      "epoch": 158,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 61778,
+      "train_examples": 50000,
+      "train_loss": 0.005168056910932064
+    },
+    {
+      "epoch": 159,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 62169,
+      "train_examples": 50000,
+      "train_loss": 0.005365388280507177
+    },
+    {
+      "epoch": 160,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 62560,
+      "train_examples": 50000,
+      "train_loss": 0.0044669982519745825
+    },
+    {
+      "epoch": 161,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 62951,
+      "train_examples": 50000,
+      "train_loss": 0.004651750999689102
+    },
+    {
+      "epoch": 162,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 63342,
+      "train_examples": 50000,
+      "train_loss": 0.0047598918889090415
+    },
+    {
+      "epoch": 163,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 63733,
+      "train_examples": 50000,
+      "train_loss": 0.004254822599813342
+    },
+    {
+      "epoch": 164,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 64124,
+      "train_examples": 50000,
+      "train_loss": 0.004679340793509036
+    },
+    {
+      "epoch": 165,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 64515,
+      "train_examples": 50000,
+      "train_loss": 0.0050124835691601035
+    },
+    {
+      "epoch": 166,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 64906,
+      "train_examples": 50000,
+      "train_loss": 0.004335557008348406
+    },
+    {
+      "epoch": 167,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 65297,
+      "train_examples": 50000,
+      "train_loss": 0.004256838730871677
+    },
+    {
+      "epoch": 168,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 65688,
+      "train_examples": 50000,
+      "train_loss": 0.0041528904421255
+    },
+    {
+      "epoch": 169,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 66079,
+      "train_examples": 50000,
+      "train_loss": 0.004626856329888105
+    },
+    {
+      "epoch": 170,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 66470,
+      "train_examples": 50000,
+      "train_loss": 0.004153371350318193
+    },
+    {
+      "epoch": 171,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 66861,
+      "train_examples": 50000,
+      "train_loss": 0.004468166066259146
+    },
+    {
+      "epoch": 172,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 67252,
+      "train_examples": 50000,
+      "train_loss": 0.004028440014198422
+    },
+    {
+      "epoch": 173,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 67643,
+      "train_examples": 50000,
+      "train_loss": 0.004062429394945502
+    },
+    {
+      "epoch": 174,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 68034,
+      "train_examples": 50000,
+      "train_loss": 0.0044060234404355286
+    },
+    {
+      "epoch": 175,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 68425,
+      "train_examples": 50000,
+      "train_loss": 0.004328454611673951
+    },
+    {
+      "epoch": 176,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 68816,
+      "train_examples": 50000,
+      "train_loss": 0.004056132428869605
+    },
+    {
+      "epoch": 177,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 69207,
+      "train_examples": 50000,
+      "train_loss": 0.004445703057423234
+    },
+    {
+      "epoch": 178,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 69598,
+      "train_examples": 50000,
+      "train_loss": 0.003961552263256163
+    },
+    {
+      "epoch": 179,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 69989,
+      "train_examples": 50000,
+      "train_loss": 0.0038441047351807354
+    },
+    {
+      "epoch": 180,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 70380,
+      "train_examples": 50000,
+      "train_loss": 0.00424440163590014
+    },
+    {
+      "epoch": 181,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 70771,
+      "train_examples": 50000,
+      "train_loss": 0.0037770761930942536
+    },
+    {
+      "epoch": 182,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 71162,
+      "train_examples": 50000,
+      "train_loss": 0.003911129658930004
+    },
+    {
+      "epoch": 183,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 71553,
+      "train_examples": 50000,
+      "train_loss": 0.003980180521793664
+    },
+    {
+      "epoch": 184,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 71944,
+      "train_examples": 50000,
+      "train_loss": 0.004038504354581237
+    },
+    {
+      "epoch": 185,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 72335,
+      "train_examples": 50000,
+      "train_loss": 0.004582509907409549
+    },
+    {
+      "epoch": 186,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 72726,
+      "train_examples": 50000,
+      "train_loss": 0.003771644706055522
+    },
+    {
+      "epoch": 187,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 73117,
+      "train_examples": 50000,
+      "train_loss": 0.003987695770338178
+    },
+    {
+      "epoch": 188,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 73508,
+      "train_examples": 50000,
+      "train_loss": 0.0038091223045438527
+    },
+    {
+      "epoch": 189,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 73899,
+      "train_examples": 50000,
+      "train_loss": 0.003995611964538693
+    },
+    {
+      "epoch": 190,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 74290,
+      "train_examples": 50000,
+      "train_loss": 0.004000969804227352
+    },
+    {
+      "epoch": 191,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 74681,
+      "train_examples": 50000,
+      "train_loss": 0.0037076463522017003
+    },
+    {
+      "epoch": 192,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 75072,
+      "train_examples": 50000,
+      "train_loss": 0.0034565704812482
+    },
+    {
+      "epoch": 193,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 75463,
+      "train_examples": 50000,
+      "train_loss": 0.003468065955489874
+    },
+    {
+      "epoch": 194,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 75854,
+      "train_examples": 50000,
+      "train_loss": 0.004178308155275881
+    },
+    {
+      "epoch": 195,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 76245,
+      "train_examples": 50000,
+      "train_loss": 0.0038809054918959736
+    },
+    {
+      "epoch": 196,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 76636,
+      "train_examples": 50000,
+      "train_loss": 0.003907086752802134
+    },
+    {
+      "epoch": 197,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 77027,
+      "train_examples": 50000,
+      "train_loss": 0.004144384138062596
+    },
+    {
+      "epoch": 198,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 77418,
+      "train_examples": 50000,
+      "train_loss": 0.0036071151252090932
+    },
+    {
+      "epoch": 199,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 77809,
+      "train_examples": 50000,
+      "train_loss": 0.0033969543035514654
+    },
+    {
+      "epoch": 200,
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 78200,
+      "train_examples": 50000,
+      "train_loss": 0.003727299030162394
+    }
+  ],
+  "evaluation_protocol": {
+    "test_evaluations": 1,
+    "test_used_for_selection": false,
+    "validation_evaluations": 0
+  },
+  "final": {
+    "accuracy": 0.9251,
+    "epoch": 200,
+    "evaluation_split": "test",
+    "finite": true,
+    "loss": 0.371705908203125,
+    "step": 78200
+  },
+  "hardware": {
+    "cuda_device_name": "NVIDIA GeForce GTX 1080",
+    "cuda_visible_devices": "5",
+    "device": "cuda",
+    "device_total_memory_bytes": 8507949056,
+    "peak_memory_allocated_bytes": 3320109568,
+    "peak_memory_reserved_bytes": 4542431232,
+    "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": 14,
+    "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": 2.065307855606079,
+    "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": 10386.89548611641,
+    "train_wall_s": 10384.79561829567
+  },
+  "work": {
+    "apical_macs_per_example": 0,
+    "causal_scalar_observations": 0,
+    "components": {
+      "apical_projection_macs": 0,
+      "apical_regression_macs": 0,
+      "bp_reverse_macs_estimate": 2509713920000000,
+      "kp_reciprocal_correlation_macs": 0,
+      "local_weight_correlation_macs": 0,
+      "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": 3764570880000000
+  }
+}