{
  "architecture": {
    "adaptive_apical_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_traffic_coefficients": 0,
    "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": 0,
    "residual_scale": 1.0,
    "vectorizer_mode": null,
    "vectorizer_parameters": 0
  },
  "args": {
    "a_scale": 1.0,
    "a_warmup_steps": 0,
    "alignment_probe": 0,
    "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:0",
    "epochs": 200,
    "eta_A": 0.01,
    "eta_P": 0.01,
    "eval_every": 20,
    "eval_split": "validation",
    "learn_P": 0,
    "loader_seed": 0,
    "lr": 0.1,
    "lr_gamma": 0.1,
    "lr_milestones": "100,150",
    "lr_schedule": "step",
    "max_steps": 0,
    "mode": "bp",
    "momentum": 0.9,
    "normalization": "batchnorm",
    "nuisance_scale": 0.0,
    "out": "results/oral_a_dev/bp_reference_primary.json",
    "output_lr": null,
    "pert_directions": 1,
    "pert_every": 4,
    "pert_sigma": 0.01,
    "perturb_seed": 1000,
    "predictor_warmup_steps": 0,
    "residual_scale": null,
    "seed": 0,
    "split_seed": 2027,
    "train_limit": 0,
    "use_residual": 1,
    "val_examples": 5000,
    "vectorizer_mode": "spatial_template",
    "warmup_epochs": 0,
    "weight_decay": 0.0001,
    "weight_scale": 1.0,
    "width": 16
  },
  "counters": {
    "apical_warmup_examples": 0,
    "calibration_event_examples": 0,
    "causal_scalar_observations": 0,
    "logical_batch_loss_queries": 0,
    "ordinary_examples": 9000000,
    "per_example_loss_terms": 0,
    "perturbation_events": 0,
    "perturbation_forward_examples": 0,
    "predictor_warmup_examples": 0
  },
  "diagnostics": null,
  "epochs": [
    {
      "epoch": 1,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 352,
      "train_examples": 45000,
      "train_loss": 1.629477106920878
    },
    {
      "epoch": 2,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 704,
      "train_examples": 45000,
      "train_loss": 1.1536506812625462
    },
    {
      "epoch": 3,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 1056,
      "train_examples": 45000,
      "train_loss": 0.9180113962067498
    },
    {
      "epoch": 4,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 1408,
      "train_examples": 45000,
      "train_loss": 0.778831071917216
    },
    {
      "epoch": 5,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 1760,
      "train_examples": 45000,
      "train_loss": 0.6890841632313198
    },
    {
      "epoch": 6,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 2112,
      "train_examples": 45000,
      "train_loss": 0.6290077259063721
    },
    {
      "epoch": 7,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 2464,
      "train_examples": 45000,
      "train_loss": 0.5832801877975464
    },
    {
      "epoch": 8,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 2816,
      "train_examples": 45000,
      "train_loss": 0.543322308868832
    },
    {
      "epoch": 9,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 3168,
      "train_examples": 45000,
      "train_loss": 0.5136268635855781
    },
    {
      "epoch": 10,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 3520,
      "train_examples": 45000,
      "train_loss": 0.4896943368858761
    },
    {
      "epoch": 11,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 3872,
      "train_examples": 45000,
      "train_loss": 0.472258512157864
    },
    {
      "epoch": 12,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 4224,
      "train_examples": 45000,
      "train_loss": 0.4496769655333625
    },
    {
      "epoch": 13,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 4576,
      "train_examples": 45000,
      "train_loss": 0.4295078576935662
    },
    {
      "epoch": 14,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 4928,
      "train_examples": 45000,
      "train_loss": 0.4181124239179823
    },
    {
      "epoch": 15,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 5280,
      "train_examples": 45000,
      "train_loss": 0.40270601289007396
    },
    {
      "epoch": 16,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 5632,
      "train_examples": 45000,
      "train_loss": 0.39158849102656046
    },
    {
      "epoch": 17,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 5984,
      "train_examples": 45000,
      "train_loss": 0.3856213444179959
    },
    {
      "epoch": 18,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 6336,
      "train_examples": 45000,
      "train_loss": 0.3813650788201226
    },
    {
      "epoch": 19,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 6688,
      "train_examples": 45000,
      "train_loss": 0.3676941867086622
    },
    {
      "epoch": 20,
      "eval_accuracy": 0.8102,
      "eval_loss": 0.62576728515625,
      "eval_split": "validation",
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 7040,
      "train_examples": 45000,
      "train_loss": 0.3598926120493147
    },
    {
      "epoch": 21,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 7392,
      "train_examples": 45000,
      "train_loss": 0.3462830693827735
    },
    {
      "epoch": 22,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 7744,
      "train_examples": 45000,
      "train_loss": 0.34032730323473614
    },
    {
      "epoch": 23,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 8096,
      "train_examples": 45000,
      "train_loss": 0.3366120978249444
    },
    {
      "epoch": 24,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 8448,
      "train_examples": 45000,
      "train_loss": 0.3290186872588264
    },
    {
      "epoch": 25,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 8800,
      "train_examples": 45000,
      "train_loss": 0.3262175612237718
    },
    {
      "epoch": 26,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 9152,
      "train_examples": 45000,
      "train_loss": 0.32130949290593463
    },
    {
      "epoch": 27,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 9504,
      "train_examples": 45000,
      "train_loss": 0.3114652105993695
    },
    {
      "epoch": 28,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 9856,
      "train_examples": 45000,
      "train_loss": 0.30679907630284625
    },
    {
      "epoch": 29,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 10208,
      "train_examples": 45000,
      "train_loss": 0.3038229338857863
    },
    {
      "epoch": 30,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 10560,
      "train_examples": 45000,
      "train_loss": 0.29693536558151246
    },
    {
      "epoch": 31,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 10912,
      "train_examples": 45000,
      "train_loss": 0.2942162125375536
    },
    {
      "epoch": 32,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 11264,
      "train_examples": 45000,
      "train_loss": 0.2900532486173842
    },
    {
      "epoch": 33,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 11616,
      "train_examples": 45000,
      "train_loss": 0.2877517813046773
    },
    {
      "epoch": 34,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 11968,
      "train_examples": 45000,
      "train_loss": 0.2830353302425808
    },
    {
      "epoch": 35,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 12320,
      "train_examples": 45000,
      "train_loss": 0.27711815487013924
    },
    {
      "epoch": 36,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 12672,
      "train_examples": 45000,
      "train_loss": 0.27821736765967475
    },
    {
      "epoch": 37,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 13024,
      "train_examples": 45000,
      "train_loss": 0.27669897685580785
    },
    {
      "epoch": 38,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 13376,
      "train_examples": 45000,
      "train_loss": 0.26504462584919397
    },
    {
      "epoch": 39,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 13728,
      "train_examples": 45000,
      "train_loss": 0.2690969581815931
    },
    {
      "epoch": 40,
      "eval_accuracy": 0.8656,
      "eval_loss": 0.4462982849121094,
      "eval_split": "validation",
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 14080,
      "train_examples": 45000,
      "train_loss": 0.2612774503389994
    },
    {
      "epoch": 41,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 14432,
      "train_examples": 45000,
      "train_loss": 0.2594946438206567
    },
    {
      "epoch": 42,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 14784,
      "train_examples": 45000,
      "train_loss": 0.25380932817988927
    },
    {
      "epoch": 43,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 15136,
      "train_examples": 45000,
      "train_loss": 0.25994675586488514
    },
    {
      "epoch": 44,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 15488,
      "train_examples": 45000,
      "train_loss": 0.257961567179362
    },
    {
      "epoch": 45,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 15840,
      "train_examples": 45000,
      "train_loss": 0.2525651240958108
    },
    {
      "epoch": 46,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 16192,
      "train_examples": 45000,
      "train_loss": 0.254560999584198
    },
    {
      "epoch": 47,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 16544,
      "train_examples": 45000,
      "train_loss": 0.2471615205552843
    },
    {
      "epoch": 48,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 16896,
      "train_examples": 45000,
      "train_loss": 0.24812334311803183
    },
    {
      "epoch": 49,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 17248,
      "train_examples": 45000,
      "train_loss": 0.2416081429057651
    },
    {
      "epoch": 50,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 17600,
      "train_examples": 45000,
      "train_loss": 0.2418114467408922
    },
    {
      "epoch": 51,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 17952,
      "train_examples": 45000,
      "train_loss": 0.2372267860915926
    },
    {
      "epoch": 52,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 18304,
      "train_examples": 45000,
      "train_loss": 0.232856031343672
    },
    {
      "epoch": 53,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 18656,
      "train_examples": 45000,
      "train_loss": 0.2328554102261861
    },
    {
      "epoch": 54,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 19008,
      "train_examples": 45000,
      "train_loss": 0.23843903945816888
    },
    {
      "epoch": 55,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 19360,
      "train_examples": 45000,
      "train_loss": 0.23594626662466262
    },
    {
      "epoch": 56,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 19712,
      "train_examples": 45000,
      "train_loss": 0.23276948260996078
    },
    {
      "epoch": 57,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 20064,
      "train_examples": 45000,
      "train_loss": 0.2300440298875173
    },
    {
      "epoch": 58,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 20416,
      "train_examples": 45000,
      "train_loss": 0.22433708667755126
    },
    {
      "epoch": 59,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 20768,
      "train_examples": 45000,
      "train_loss": 0.23259679017331866
    },
    {
      "epoch": 60,
      "eval_accuracy": 0.8376,
      "eval_loss": 0.5762138427734375,
      "eval_split": "validation",
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 21120,
      "train_examples": 45000,
      "train_loss": 0.221930826971266
    },
    {
      "epoch": 61,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 21472,
      "train_examples": 45000,
      "train_loss": 0.22306284332010481
    },
    {
      "epoch": 62,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 21824,
      "train_examples": 45000,
      "train_loss": 0.21996720467143588
    },
    {
      "epoch": 63,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 22176,
      "train_examples": 45000,
      "train_loss": 0.21993547526200613
    },
    {
      "epoch": 64,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 22528,
      "train_examples": 45000,
      "train_loss": 0.22175351356665293
    },
    {
      "epoch": 65,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 22880,
      "train_examples": 45000,
      "train_loss": 0.21980250878069135
    },
    {
      "epoch": 66,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 23232,
      "train_examples": 45000,
      "train_loss": 0.22193617814646827
    },
    {
      "epoch": 67,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 23584,
      "train_examples": 45000,
      "train_loss": 0.21733659199078878
    },
    {
      "epoch": 68,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 23936,
      "train_examples": 45000,
      "train_loss": 0.21201883059077792
    },
    {
      "epoch": 69,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 24288,
      "train_examples": 45000,
      "train_loss": 0.21097899855507746
    },
    {
      "epoch": 70,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 24640,
      "train_examples": 45000,
      "train_loss": 0.2180972268342972
    },
    {
      "epoch": 71,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 24992,
      "train_examples": 45000,
      "train_loss": 0.20807346503204768
    },
    {
      "epoch": 72,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 25344,
      "train_examples": 45000,
      "train_loss": 0.2113001620133718
    },
    {
      "epoch": 73,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 25696,
      "train_examples": 45000,
      "train_loss": 0.20991453639666238
    },
    {
      "epoch": 74,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 26048,
      "train_examples": 45000,
      "train_loss": 0.2111522406551573
    },
    {
      "epoch": 75,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 26400,
      "train_examples": 45000,
      "train_loss": 0.20713426054053835
    },
    {
      "epoch": 76,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 26752,
      "train_examples": 45000,
      "train_loss": 0.2090695972919464
    },
    {
      "epoch": 77,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 27104,
      "train_examples": 45000,
      "train_loss": 0.21206452231407166
    },
    {
      "epoch": 78,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 27456,
      "train_examples": 45000,
      "train_loss": 0.2080325007915497
    },
    {
      "epoch": 79,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 27808,
      "train_examples": 45000,
      "train_loss": 0.20438791779147253
    },
    {
      "epoch": 80,
      "eval_accuracy": 0.8716,
      "eval_loss": 0.45828941650390626,
      "eval_split": "validation",
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 28160,
      "train_examples": 45000,
      "train_loss": 0.2006779747698042
    },
    {
      "epoch": 81,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 28512,
      "train_examples": 45000,
      "train_loss": 0.2016580670568678
    },
    {
      "epoch": 82,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 28864,
      "train_examples": 45000,
      "train_loss": 0.2003150415526496
    },
    {
      "epoch": 83,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 29216,
      "train_examples": 45000,
      "train_loss": 0.19765809888839722
    },
    {
      "epoch": 84,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 29568,
      "train_examples": 45000,
      "train_loss": 0.20244128450287713
    },
    {
      "epoch": 85,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 29920,
      "train_examples": 45000,
      "train_loss": 0.19974544689920212
    },
    {
      "epoch": 86,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 30272,
      "train_examples": 45000,
      "train_loss": 0.19561397054460314
    },
    {
      "epoch": 87,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 30624,
      "train_examples": 45000,
      "train_loss": 0.20291065661112467
    },
    {
      "epoch": 88,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 30976,
      "train_examples": 45000,
      "train_loss": 0.1980198811451594
    },
    {
      "epoch": 89,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 31328,
      "train_examples": 45000,
      "train_loss": 0.1991081144505077
    },
    {
      "epoch": 90,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 31680,
      "train_examples": 45000,
      "train_loss": 0.19957862480746374
    },
    {
      "epoch": 91,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 32032,
      "train_examples": 45000,
      "train_loss": 0.2043161620643404
    },
    {
      "epoch": 92,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 32384,
      "train_examples": 45000,
      "train_loss": 0.19370153923299577
    },
    {
      "epoch": 93,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 32736,
      "train_examples": 45000,
      "train_loss": 0.19246835175620186
    },
    {
      "epoch": 94,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 33088,
      "train_examples": 45000,
      "train_loss": 0.19274455511305066
    },
    {
      "epoch": 95,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 33440,
      "train_examples": 45000,
      "train_loss": 0.18495470209121703
    },
    {
      "epoch": 96,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 33792,
      "train_examples": 45000,
      "train_loss": 0.1947995986170239
    },
    {
      "epoch": 97,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 34144,
      "train_examples": 45000,
      "train_loss": 0.19470539894368913
    },
    {
      "epoch": 98,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 34496,
      "train_examples": 45000,
      "train_loss": 0.19306121972931756
    },
    {
      "epoch": 99,
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 34848,
      "train_examples": 45000,
      "train_loss": 0.19163403051959144
    },
    {
      "epoch": 100,
      "eval_accuracy": 0.8688,
      "eval_loss": 0.4629053649902344,
      "eval_split": "validation",
      "lr": 0.1,
      "output_lr": 0.1,
      "step": 35200,
      "train_examples": 45000,
      "train_loss": 0.1941945056743092
    },
    {
      "epoch": 101,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 35552,
      "train_examples": 45000,
      "train_loss": 0.12243073383967082
    },
    {
      "epoch": 102,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 35904,
      "train_examples": 45000,
      "train_loss": 0.09238274479839537
    },
    {
      "epoch": 103,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 36256,
      "train_examples": 45000,
      "train_loss": 0.08373940632475747
    },
    {
      "epoch": 104,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 36608,
      "train_examples": 45000,
      "train_loss": 0.07404987008968988
    },
    {
      "epoch": 105,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 36960,
      "train_examples": 45000,
      "train_loss": 0.07065127567052841
    },
    {
      "epoch": 106,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 37312,
      "train_examples": 45000,
      "train_loss": 0.06621434763802422
    },
    {
      "epoch": 107,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 37664,
      "train_examples": 45000,
      "train_loss": 0.06265637842085627
    },
    {
      "epoch": 108,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 38016,
      "train_examples": 45000,
      "train_loss": 0.06095097218089633
    },
    {
      "epoch": 109,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 38368,
      "train_examples": 45000,
      "train_loss": 0.055481788088215724
    },
    {
      "epoch": 110,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 38720,
      "train_examples": 45000,
      "train_loss": 0.05485475784805086
    },
    {
      "epoch": 111,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 39072,
      "train_examples": 45000,
      "train_loss": 0.052292619497246215
    },
    {
      "epoch": 112,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 39424,
      "train_examples": 45000,
      "train_loss": 0.05090429345236884
    },
    {
      "epoch": 113,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 39776,
      "train_examples": 45000,
      "train_loss": 0.04814498667187161
    },
    {
      "epoch": 114,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 40128,
      "train_examples": 45000,
      "train_loss": 0.04657038381099701
    },
    {
      "epoch": 115,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 40480,
      "train_examples": 45000,
      "train_loss": 0.045128318380978374
    },
    {
      "epoch": 116,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 40832,
      "train_examples": 45000,
      "train_loss": 0.04501160361369451
    },
    {
      "epoch": 117,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 41184,
      "train_examples": 45000,
      "train_loss": 0.04166722600724962
    },
    {
      "epoch": 118,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 41536,
      "train_examples": 45000,
      "train_loss": 0.04189523547887802
    },
    {
      "epoch": 119,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 41888,
      "train_examples": 45000,
      "train_loss": 0.0410013985686832
    },
    {
      "epoch": 120,
      "eval_accuracy": 0.9116,
      "eval_loss": 0.36133499145507814,
      "eval_split": "validation",
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 42240,
      "train_examples": 45000,
      "train_loss": 0.039646570123897655
    },
    {
      "epoch": 121,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 42592,
      "train_examples": 45000,
      "train_loss": 0.038702496344513365
    },
    {
      "epoch": 122,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 42944,
      "train_examples": 45000,
      "train_loss": 0.03543033310820659
    },
    {
      "epoch": 123,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 43296,
      "train_examples": 45000,
      "train_loss": 0.037469266428550085
    },
    {
      "epoch": 124,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 43648,
      "train_examples": 45000,
      "train_loss": 0.03570658944944541
    },
    {
      "epoch": 125,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 44000,
      "train_examples": 45000,
      "train_loss": 0.036685267938176794
    },
    {
      "epoch": 126,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 44352,
      "train_examples": 45000,
      "train_loss": 0.034107129124800364
    },
    {
      "epoch": 127,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 44704,
      "train_examples": 45000,
      "train_loss": 0.03400421402851741
    },
    {
      "epoch": 128,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 45056,
      "train_examples": 45000,
      "train_loss": 0.03237723284893566
    },
    {
      "epoch": 129,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 45408,
      "train_examples": 45000,
      "train_loss": 0.032963586113188004
    },
    {
      "epoch": 130,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 45760,
      "train_examples": 45000,
      "train_loss": 0.029909599794447423
    },
    {
      "epoch": 131,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 46112,
      "train_examples": 45000,
      "train_loss": 0.03154643616692888
    },
    {
      "epoch": 132,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 46464,
      "train_examples": 45000,
      "train_loss": 0.030607043651739756
    },
    {
      "epoch": 133,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 46816,
      "train_examples": 45000,
      "train_loss": 0.029234873590535587
    },
    {
      "epoch": 134,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 47168,
      "train_examples": 45000,
      "train_loss": 0.027656678012675708
    },
    {
      "epoch": 135,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 47520,
      "train_examples": 45000,
      "train_loss": 0.029045232111877865
    },
    {
      "epoch": 136,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 47872,
      "train_examples": 45000,
      "train_loss": 0.027592006523079343
    },
    {
      "epoch": 137,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 48224,
      "train_examples": 45000,
      "train_loss": 0.028573222221930823
    },
    {
      "epoch": 138,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 48576,
      "train_examples": 45000,
      "train_loss": 0.026246732861631445
    },
    {
      "epoch": 139,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 48928,
      "train_examples": 45000,
      "train_loss": 0.026507694556978014
    },
    {
      "epoch": 140,
      "eval_accuracy": 0.9104,
      "eval_loss": 0.41732127685546877,
      "eval_split": "validation",
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 49280,
      "train_examples": 45000,
      "train_loss": 0.024990961343050004
    },
    {
      "epoch": 141,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 49632,
      "train_examples": 45000,
      "train_loss": 0.0275127333773507
    },
    {
      "epoch": 142,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 49984,
      "train_examples": 45000,
      "train_loss": 0.024205242278178534
    },
    {
      "epoch": 143,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 50336,
      "train_examples": 45000,
      "train_loss": 0.02479735333588388
    },
    {
      "epoch": 144,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 50688,
      "train_examples": 45000,
      "train_loss": 0.023810774516893757
    },
    {
      "epoch": 145,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 51040,
      "train_examples": 45000,
      "train_loss": 0.025666841644710964
    },
    {
      "epoch": 146,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 51392,
      "train_examples": 45000,
      "train_loss": 0.023461431486739053
    },
    {
      "epoch": 147,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 51744,
      "train_examples": 45000,
      "train_loss": 0.02378552058537801
    },
    {
      "epoch": 148,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 52096,
      "train_examples": 45000,
      "train_loss": 0.023739897603789965
    },
    {
      "epoch": 149,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 52448,
      "train_examples": 45000,
      "train_loss": 0.022505639592144223
    },
    {
      "epoch": 150,
      "lr": 0.010000000000000002,
      "output_lr": 0.010000000000000002,
      "step": 52800,
      "train_examples": 45000,
      "train_loss": 0.020517713482843505
    },
    {
      "epoch": 151,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 53152,
      "train_examples": 45000,
      "train_loss": 0.019582247539361317
    },
    {
      "epoch": 152,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 53504,
      "train_examples": 45000,
      "train_loss": 0.019260346906714968
    },
    {
      "epoch": 153,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 53856,
      "train_examples": 45000,
      "train_loss": 0.017384963521858057
    },
    {
      "epoch": 154,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 54208,
      "train_examples": 45000,
      "train_loss": 0.017249316464033392
    },
    {
      "epoch": 155,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 54560,
      "train_examples": 45000,
      "train_loss": 0.016512117852105036
    },
    {
      "epoch": 156,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 54912,
      "train_examples": 45000,
      "train_loss": 0.01810812204281489
    },
    {
      "epoch": 157,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 55264,
      "train_examples": 45000,
      "train_loss": 0.017985127888123193
    },
    {
      "epoch": 158,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 55616,
      "train_examples": 45000,
      "train_loss": 0.016650764848291873
    },
    {
      "epoch": 159,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 55968,
      "train_examples": 45000,
      "train_loss": 0.016366760524776248
    },
    {
      "epoch": 160,
      "eval_accuracy": 0.916,
      "eval_loss": 0.41492218017578125,
      "eval_split": "validation",
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 56320,
      "train_examples": 45000,
      "train_loss": 0.017086565133598117
    },
    {
      "epoch": 161,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 56672,
      "train_examples": 45000,
      "train_loss": 0.015958600907524427
    },
    {
      "epoch": 162,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 57024,
      "train_examples": 45000,
      "train_loss": 0.016478912470158605
    },
    {
      "epoch": 163,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 57376,
      "train_examples": 45000,
      "train_loss": 0.016271839502453804
    },
    {
      "epoch": 164,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 57728,
      "train_examples": 45000,
      "train_loss": 0.016035092358125582
    },
    {
      "epoch": 165,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 58080,
      "train_examples": 45000,
      "train_loss": 0.016342996538347667
    },
    {
      "epoch": 166,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 58432,
      "train_examples": 45000,
      "train_loss": 0.015872134373295637
    },
    {
      "epoch": 167,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 58784,
      "train_examples": 45000,
      "train_loss": 0.016195324125885962
    },
    {
      "epoch": 168,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 59136,
      "train_examples": 45000,
      "train_loss": 0.015301043385598394
    },
    {
      "epoch": 169,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 59488,
      "train_examples": 45000,
      "train_loss": 0.015656031346321105
    },
    {
      "epoch": 170,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 59840,
      "train_examples": 45000,
      "train_loss": 0.015911012617084715
    },
    {
      "epoch": 171,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 60192,
      "train_examples": 45000,
      "train_loss": 0.016036496425006125
    },
    {
      "epoch": 172,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 60544,
      "train_examples": 45000,
      "train_loss": 0.015414181932475832
    },
    {
      "epoch": 173,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 60896,
      "train_examples": 45000,
      "train_loss": 0.015800860342052248
    },
    {
      "epoch": 174,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 61248,
      "train_examples": 45000,
      "train_loss": 0.015768850159976217
    },
    {
      "epoch": 175,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 61600,
      "train_examples": 45000,
      "train_loss": 0.015496016425887743
    },
    {
      "epoch": 176,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 61952,
      "train_examples": 45000,
      "train_loss": 0.014675625985695256
    },
    {
      "epoch": 177,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 62304,
      "train_examples": 45000,
      "train_loss": 0.014459471281700664
    },
    {
      "epoch": 178,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 62656,
      "train_examples": 45000,
      "train_loss": 0.015453773557477526
    },
    {
      "epoch": 179,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 63008,
      "train_examples": 45000,
      "train_loss": 0.014391206828753153
    },
    {
      "epoch": 180,
      "eval_accuracy": 0.9152,
      "eval_loss": 0.41804723510742187,
      "eval_split": "validation",
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 63360,
      "train_examples": 45000,
      "train_loss": 0.015613929537600942
    },
    {
      "epoch": 181,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 63712,
      "train_examples": 45000,
      "train_loss": 0.015048576422201263
    },
    {
      "epoch": 182,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 64064,
      "train_examples": 45000,
      "train_loss": 0.01569497499167919
    },
    {
      "epoch": 183,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 64416,
      "train_examples": 45000,
      "train_loss": 0.014797748851279418
    },
    {
      "epoch": 184,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 64768,
      "train_examples": 45000,
      "train_loss": 0.014610261275370916
    },
    {
      "epoch": 185,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 65120,
      "train_examples": 45000,
      "train_loss": 0.013999619007110596
    },
    {
      "epoch": 186,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 65472,
      "train_examples": 45000,
      "train_loss": 0.014426271572377946
    },
    {
      "epoch": 187,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 65824,
      "train_examples": 45000,
      "train_loss": 0.013885394373204973
    },
    {
      "epoch": 188,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 66176,
      "train_examples": 45000,
      "train_loss": 0.014582767293105522
    },
    {
      "epoch": 189,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 66528,
      "train_examples": 45000,
      "train_loss": 0.015294219624665048
    },
    {
      "epoch": 190,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 66880,
      "train_examples": 45000,
      "train_loss": 0.014451671667562591
    },
    {
      "epoch": 191,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 67232,
      "train_examples": 45000,
      "train_loss": 0.014434750450319713
    },
    {
      "epoch": 192,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 67584,
      "train_examples": 45000,
      "train_loss": 0.013754335116677814
    },
    {
      "epoch": 193,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 67936,
      "train_examples": 45000,
      "train_loss": 0.014247742917802599
    },
    {
      "epoch": 194,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 68288,
      "train_examples": 45000,
      "train_loss": 0.01404493830733829
    },
    {
      "epoch": 195,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 68640,
      "train_examples": 45000,
      "train_loss": 0.014140160277816984
    },
    {
      "epoch": 196,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 68992,
      "train_examples": 45000,
      "train_loss": 0.013918220730291473
    },
    {
      "epoch": 197,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 69344,
      "train_examples": 45000,
      "train_loss": 0.01464969994003574
    },
    {
      "epoch": 198,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 69696,
      "train_examples": 45000,
      "train_loss": 0.013554245031542247
    },
    {
      "epoch": 199,
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 70048,
      "train_examples": 45000,
      "train_loss": 0.014428260192275047
    },
    {
      "epoch": 200,
      "eval_accuracy": 0.9162,
      "eval_loss": 0.42326114501953127,
      "eval_split": "validation",
      "lr": 0.0010000000000000002,
      "output_lr": 0.0010000000000000002,
      "step": 70400,
      "train_examples": 45000,
      "train_loss": 0.01528975431472063
    }
  ],
  "evaluation_protocol": {
    "test_evaluations": 0,
    "test_used_for_selection": false,
    "validation_evaluations": 11
  },
  "final": {
    "accuracy": 0.9162,
    "epoch": 200,
    "evaluation_split": "validation",
    "finite": true,
    "loss": 0.42326114501953127,
    "step": 70400
  },
  "hardware": {
    "cuda_device_name": "NVIDIA GeForce GTX 1080",
    "cuda_visible_devices": "5",
    "device": "cuda:0",
    "device_total_memory_bytes": 8507949056,
    "peak_memory_allocated_bytes": 2169162752,
    "peak_memory_reserved_bytes": 2755657728,
    "torch_version": "2.3.1+cu118"
  },
  "protocol_family": "oral_a_cifar_local_resnet_development",
  "provenance": {
    "git_commit": "c8ae43fb284cf5be54066b4c0b5b17dcb088227f",
    "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": "validation",
    "input_layout": "NCHW",
    "input_shape": [
      3,
      32,
      32
    ],
    "loader_seed": 0,
    "normalization_mean": [
      0.49140000343322754,
      0.4821999967098236,
      0.4465000033378601
    ],
    "normalization_std": [
      0.24699999392032623,
      0.2434999942779541,
      0.26159998774528503
    ],
    "split_from_training_only": true,
    "split_seed": 2027,
    "test_examples": 10000,
    "train_examples": 45000,
    "training_augmentation": "random_crop_32_padding_4_zero_then_horizontal_flip_p0.5",
    "validation_class_counts": {
      "0": 500,
      "1": 500,
      "2": 500,
      "3": 500,
      "4": 500,
      "5": 500,
      "6": 500,
      "7": 500,
      "8": 500,
      "9": 500
    },
    "validation_examples": 5000,
    "validation_index_sha256": "8328b206a97c420e49e54e3eca4abe3274c4756b084355784ea3fb8059e4515b"
  },
  "timing": {
    "apical_warmup_wall_s": 0.0,
    "evaluation_wall_s": 3.8121464252471924,
    "predictor_warmup_wall_s": 0.0,
    "timing_excludes_data_loading_hashing_and_model_construction": true,
    "total_timed_wall_s": 3405.721784353256,
    "train_wall_s": 3401.876144886017
  },
  "work": {
    "apical_macs_per_example": 0,
    "causal_scalar_observations": 0,
    "components": {
      "apical_projection_macs": 0,
      "apical_regression_macs": 0,
      "bp_reverse_macs_estimate": 729918720000000,
      "local_weight_correlation_macs": 0,
      "ordinary_forward_macs": 364959360000000,
      "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; elementwise nonlinearities and optimizer arithmetic excluded",
    "forward_macs_per_example": 40551040,
    "logical_batch_loss_queries": 0,
    "per_example_cross_entropy_terms": 0,
    "total_clean_forward_examples": 9000000,
    "total_forward_equivalent_examples": 9000000,
    "total_macs_estimate": 1094878080000000
  }
}