experiment: add completed A6000 crossover resultsa6000-x1-results

1 files changed, 2669 insertions, 0 deletions

diff --git a/results/resnet_crossover/r2/resnet-r2-fa-d20-lr0p1.json b/results/resnet_crossover/r2/resnet-r2-fa-d20-lr0p1.json
new file mode 100644
index 0000000..1096d53
--- /dev/null
+++ b/results/resnet_crossover/r2/resnet-r2-fa-d20-lr0p1.json
@@ -0,0 +1,2669 @@
+{
+  "architecture": {
+    "adaptive_apical_parameters": 0,
+    "adaptive_feedback_parameters": 0,
+    "base_width": 16,
+    "blocks_per_stage": 3,
+    "bn_eps": 1e-05,
+    "bn_momentum": 0.1,
+    "depth": 20,
+    "family": "CIFAR 6n+2 ResNet, option-A shortcuts",
+    "fixed_feedback_parameters": 267904,
+    "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": 32,
+    "apical_calibration_mode": "unit_targets",
+    "apical_seed": null,
+    "augment_train": 1,
+    "batch_size": 128,
+    "bn_eps": 1e-05,
+    "bn_momentum": 0.1,
+    "data_dir": "/orion/u/oscarwan/scratch/sdil-data",
+    "depth": 20,
+    "device": "cuda",
+    "epochs": 200,
+    "eta_A": 0.01,
+    "eta_P": 0.01,
+    "eval_every": 1,
+    "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,
+    "mirror_batch_size": 1,
+    "mirror_eta": 0.1,
+    "mirror_every": 16,
+    "mirror_noise_std": 1.0,
+    "mirror_seed": 3000,
+    "mirror_warmup_steps": 0,
+    "mode": "hfa",
+    "momentum": 0.9,
+    "neutral_projection": 0,
+    "normalization": "batchnorm",
+    "nuisance_scale": 0.0,
+    "out": "/orion/u/oscarwan/scratch/sdil-work/sdil-a6000-full-rerun/results/resnet_crossover/r2/resnet-r2-fa-d20-lr0p1.json",
+    "output_lr": 0.1,
+    "pert_directions": 1,
+    "pert_every": 4,
+    "pert_sigma": 0.01,
+    "perturb_seed": 1000,
+    "predictor_every": 0,
+    "predictor_mode": "nlms",
+    "predictor_warmup_steps": 0,
+    "residual_scale": null,
+    "seed": 0,
+    "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": 5000,
+    "vectorizer_mode": "spatial_template",
+    "warmup_epochs": 5,
+    "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": 9000000,
+    "per_example_loss_terms": 0,
+    "perturbation_events": 0,
+    "perturbation_forward_examples": 0,
+    "predictor_update_examples": 0,
+    "predictor_warmup_examples": 0,
+    "traffic_audit_examples": 0,
+    "traffic_calibration_examples": 0
+  },
+  "diagnostics": {
+    "early_third_mean": 0.011225155787542462,
+    "feedback_forward_cosine": [
+      0.06530784070491791,
+      0.23644882440567017,
+      0.14832958579063416,
+      0.27313339710235596,
+      0.11167817562818527,
+      0.22827664017677307,
+      0.11369006335735321,
+      0.18833361566066742,
+      0.08932708203792572,
+      0.2524581551551819,
+      0.10242670029401779,
+      0.24816101789474487,
+      0.08819255232810974,
+      0.19018340110778809,
+      0.08682424575090408,
+      0.2735845446586609,
+      0.09671223163604736,
+      0.26411014795303345,
+      0.7856658697128296
+    ],
+    "feedback_forward_norm_ratio": [
+      1.5932362079620361,
+      1.5769914388656616,
+      1.9141095876693726,
+      1.9809224605560303,
+      2.1602015495300293,
+      2.258635997772217,
+      2.286642551422119,
+      1.7141746282577515,
+      2.1372995376586914,
+      1.9748387336730957,
+      2.4078481197357178,
+      2.3249430656433105,
+      2.3330516815185547,
+      1.5594871044158936,
+      1.7516019344329834,
+      1.3729456663131714,
+      1.6120082139968872,
+      1.1900783777236938,
+      0.22846350073814392
+    ],
+    "feedback_forward_relative_error": [
+      1.8249109983444214,
+      1.6556408405303955,
+      2.0238521099090576,
+      1.9600871801376343,
+      2.276834726333618,
+      2.2517213821411133,
+      2.3893086910247803,
+      1.8145856857299805,
+      2.2773256301879883,
+      1.9755656719207764,
+      2.5108718872070312,
+      2.2916018962860107,
+      2.455934762954712,
+      1.6848807334899902,
+      1.940089225769043,
+      1.4607349634170532,
+      1.812944769859314,
+      1.3370351791381836,
+      0.8325884938240051
+    ],
+    "innovation_negative_gradient_cosine": [
+      0.0030611385591328144,
+      0.006984669715166092,
+      0.015605964697897434,
+      0.010937267914414406,
+      0.02078421413898468,
+      0.009977679699659348,
+      0.03010556846857071,
+      0.02181582897901535,
+      0.06369832903146744,
+      0.040703386068344116,
+      0.09983091056346893,
+      0.06209838390350342,
+      0.13944962620735168,
+      0.11991359293460846,
+      0.28113487362861633,
+      0.28801897168159485,
+      0.4426411986351013,
+      0.47109293937683105,
+      0.8467425107955933
+    ],
+    "mean_feedback_forward_cosine": 0.2022549521766211,
+    "mean_feedback_forward_relative_error": 1.9356060435897426,
+    "normalization_state": "training_batch_stats_without_running_update",
+    "raw_negative_gradient_cosine": [
+      0.0030611385591328144,
+      0.006984669715166092,
+      0.015605964697897434,
+      0.010937267914414406,
+      0.02078421413898468,
+      0.009977679699659348,
+      0.03010556846857071,
+      0.02181582897901535,
+      0.06369832903146744,
+      0.040703386068344116,
+      0.09983091056346893,
+      0.06209838390350342,
+      0.13944962620735168,
+      0.11991359293460846,
+      0.28113487362861633,
+      0.28801897168159485,
+      0.4426411986351013,
+      0.47109293937683105,
+      0.8467425107955933
+    ],
+    "teaching_negative_gradient_cosine": [
+      0.0030611385591328144,
+      0.006984669715166092,
+      0.015605964697897434,
+      0.010937267914414406,
+      0.02078421413898468,
+      0.009977679699659348,
+      0.03010556846857071,
+      0.02181582897901535,
+      0.06369832903146744,
+      0.040703386068344116,
+      0.09983091056346893,
+      0.06209838390350342,
+      0.13944962620735168,
+      0.11991359293460846,
+      0.28113487362861633,
+      0.28801897168159485,
+      0.4426411986351013,
+      0.47109293937683105,
+      0.8467425107955933
+    ],
+    "wall_s": 0.10387682914733887
+  },
+  "epochs": [
+    {
+      "epoch": 1,
+      "eval_accuracy": 0.3056,
+      "eval_loss": 1.9553183349609375,
+      "eval_split": "validation",
+      "lr": 0.02,
+      "output_lr": 0.02,
+      "step": 352,
+      "train_examples": 45000,
+      "train_loss": 2.0639842291302153
+    },
+    {
+      "epoch": 2,
+      "eval_accuracy": 0.3414,
+      "eval_loss": 1.8812863037109375,
+      "eval_split": "validation",
+      "lr": 0.04,
+      "output_lr": 0.04,
+      "step": 704,
+      "train_examples": 45000,
+      "train_loss": 1.91367295850118
+    },
+    {
+      "epoch": 3,
+      "eval_accuracy": 0.3298,
+      "eval_loss": 1.8941572509765625,
+      "eval_split": "validation",
+      "lr": 0.06000000000000001,
+      "output_lr": 0.06000000000000001,
+      "step": 1056,
+      "train_examples": 45000,
+      "train_loss": 2.0059402748743693
+    },
+    {
+      "epoch": 4,
+      "eval_accuracy": 0.3868,
+      "eval_loss": 2.012137060546875,
+      "eval_split": "validation",
+      "lr": 0.08,
+      "output_lr": 0.08,
+      "step": 1408,
+      "train_examples": 45000,
+      "train_loss": 2.1139173667060005
+    },
+    {
+      "epoch": 5,
+      "eval_accuracy": 0.3448,
+      "eval_loss": 2.873762353515625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 1760,
+      "train_examples": 45000,
+      "train_loss": 2.4378785884433323
+    },
+    {
+      "epoch": 6,
+      "eval_accuracy": 0.4028,
+      "eval_loss": 2.19697216796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 2112,
+      "train_examples": 45000,
+      "train_loss": 2.382185635460748
+    },
+    {
+      "epoch": 7,
+      "eval_accuracy": 0.3742,
+      "eval_loss": 2.617165966796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 2464,
+      "train_examples": 45000,
+      "train_loss": 2.240234103329976
+    },
+    {
+      "epoch": 8,
+      "eval_accuracy": 0.3884,
+      "eval_loss": 2.572106884765625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 2816,
+      "train_examples": 45000,
+      "train_loss": 2.071946952946981
+    },
+    {
+      "epoch": 9,
+      "eval_accuracy": 0.4522,
+      "eval_loss": 1.9747636474609376,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 3168,
+      "train_examples": 45000,
+      "train_loss": 2.058614935133192
+    },
+    {
+      "epoch": 10,
+      "eval_accuracy": 0.432,
+      "eval_loss": 2.391416796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 3520,
+      "train_examples": 45000,
+      "train_loss": 1.9395815348519219
+    },
+    {
+      "epoch": 11,
+      "eval_accuracy": 0.3852,
+      "eval_loss": 2.6288126953125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 3872,
+      "train_examples": 45000,
+      "train_loss": 1.9402141478644477
+    },
+    {
+      "epoch": 12,
+      "eval_accuracy": 0.419,
+      "eval_loss": 2.22531181640625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 4224,
+      "train_examples": 45000,
+      "train_loss": 1.9182814904106988
+    },
+    {
+      "epoch": 13,
+      "eval_accuracy": 0.4612,
+      "eval_loss": 1.892072021484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 4576,
+      "train_examples": 45000,
+      "train_loss": 1.7968032083511352
+    },
+    {
+      "epoch": 14,
+      "eval_accuracy": 0.417,
+      "eval_loss": 2.081826513671875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 4928,
+      "train_examples": 45000,
+      "train_loss": 1.8566136685053507
+    },
+    {
+      "epoch": 15,
+      "eval_accuracy": 0.438,
+      "eval_loss": 1.9989041015625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 5280,
+      "train_examples": 45000,
+      "train_loss": 1.7629460686153835
+    },
+    {
+      "epoch": 16,
+      "eval_accuracy": 0.458,
+      "eval_loss": 2.0938974365234375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 5632,
+      "train_examples": 45000,
+      "train_loss": 1.7258217601776122
+    },
+    {
+      "epoch": 17,
+      "eval_accuracy": 0.4368,
+      "eval_loss": 2.362848486328125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 5984,
+      "train_examples": 45000,
+      "train_loss": 1.7666962187872992
+    },
+    {
+      "epoch": 18,
+      "eval_accuracy": 0.4846,
+      "eval_loss": 1.92572080078125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 6336,
+      "train_examples": 45000,
+      "train_loss": 1.7164832483503554
+    },
+    {
+      "epoch": 19,
+      "eval_accuracy": 0.494,
+      "eval_loss": 1.8626768310546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 6688,
+      "train_examples": 45000,
+      "train_loss": 1.6503403984917535
+    },
+    {
+      "epoch": 20,
+      "eval_accuracy": 0.3992,
+      "eval_loss": 3.14110546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 7040,
+      "train_examples": 45000,
+      "train_loss": 1.7138124549018012
+    },
+    {
+      "epoch": 21,
+      "eval_accuracy": 0.4812,
+      "eval_loss": 1.944050341796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 7392,
+      "train_examples": 45000,
+      "train_loss": 1.6447507121192084
+    },
+    {
+      "epoch": 22,
+      "eval_accuracy": 0.4892,
+      "eval_loss": 1.911722021484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 7744,
+      "train_examples": 45000,
+      "train_loss": 1.6547758380042181
+    },
+    {
+      "epoch": 23,
+      "eval_accuracy": 0.485,
+      "eval_loss": 1.815741455078125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 8096,
+      "train_examples": 45000,
+      "train_loss": 1.5776512994130452
+    },
+    {
+      "epoch": 24,
+      "eval_accuracy": 0.5166,
+      "eval_loss": 1.5387634521484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 8448,
+      "train_examples": 45000,
+      "train_loss": 1.61900724834866
+    },
+    {
+      "epoch": 25,
+      "eval_accuracy": 0.4492,
+      "eval_loss": 2.1026521484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 8800,
+      "train_examples": 45000,
+      "train_loss": 1.6446258316040039
+    },
+    {
+      "epoch": 26,
+      "eval_accuracy": 0.461,
+      "eval_loss": 1.787435693359375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 9152,
+      "train_examples": 45000,
+      "train_loss": 1.635316059621175
+    },
+    {
+      "epoch": 27,
+      "eval_accuracy": 0.4612,
+      "eval_loss": 2.1551716796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 9504,
+      "train_examples": 45000,
+      "train_loss": 1.624604059685601
+    },
+    {
+      "epoch": 28,
+      "eval_accuracy": 0.4852,
+      "eval_loss": 1.75483876953125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 9856,
+      "train_examples": 45000,
+      "train_loss": 1.6604180307176377
+    },
+    {
+      "epoch": 29,
+      "eval_accuracy": 0.5064,
+      "eval_loss": 1.5234459716796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 10208,
+      "train_examples": 45000,
+      "train_loss": 1.585603558646308
+    },
+    {
+      "epoch": 30,
+      "eval_accuracy": 0.4848,
+      "eval_loss": 1.8498150634765624,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 10560,
+      "train_examples": 45000,
+      "train_loss": 1.5916083756128947
+    },
+    {
+      "epoch": 31,
+      "eval_accuracy": 0.4718,
+      "eval_loss": 1.84035732421875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 10912,
+      "train_examples": 45000,
+      "train_loss": 1.6653513647927178
+    },
+    {
+      "epoch": 32,
+      "eval_accuracy": 0.5348,
+      "eval_loss": 1.6685882568359376,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 11264,
+      "train_examples": 45000,
+      "train_loss": 1.6522833022647434
+    },
+    {
+      "epoch": 33,
+      "eval_accuracy": 0.463,
+      "eval_loss": 2.010286328125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 11616,
+      "train_examples": 45000,
+      "train_loss": 1.581467076174418
+    },
+    {
+      "epoch": 34,
+      "eval_accuracy": 0.4372,
+      "eval_loss": 2.1312464111328127,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 11968,
+      "train_examples": 45000,
+      "train_loss": 1.6292456374486288
+    },
+    {
+      "epoch": 35,
+      "eval_accuracy": 0.5374,
+      "eval_loss": 1.577523291015625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 12320,
+      "train_examples": 45000,
+      "train_loss": 1.5334896649254692
+    },
+    {
+      "epoch": 36,
+      "eval_accuracy": 0.4628,
+      "eval_loss": 1.90355205078125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 12672,
+      "train_examples": 45000,
+      "train_loss": 1.504615032196045
+    },
+    {
+      "epoch": 37,
+      "eval_accuracy": 0.5238,
+      "eval_loss": 1.5645536376953124,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 13024,
+      "train_examples": 45000,
+      "train_loss": 1.596086120054457
+    },
+    {
+      "epoch": 38,
+      "eval_accuracy": 0.5198,
+      "eval_loss": 1.5882181884765625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 13376,
+      "train_examples": 45000,
+      "train_loss": 1.5680132107628717
+    },
+    {
+      "epoch": 39,
+      "eval_accuracy": 0.5282,
+      "eval_loss": 1.6558640380859375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 13728,
+      "train_examples": 45000,
+      "train_loss": 1.5771280631595188
+    },
+    {
+      "epoch": 40,
+      "eval_accuracy": 0.5684,
+      "eval_loss": 1.37480546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 14080,
+      "train_examples": 45000,
+      "train_loss": 1.6025711276796129
+    },
+    {
+      "epoch": 41,
+      "eval_accuracy": 0.5168,
+      "eval_loss": 1.548530712890625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 14432,
+      "train_examples": 45000,
+      "train_loss": 1.5799489162021214
+    },
+    {
+      "epoch": 42,
+      "eval_accuracy": 0.4688,
+      "eval_loss": 2.0833614990234377,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 14784,
+      "train_examples": 45000,
+      "train_loss": 1.5346269664764405
+    },
+    {
+      "epoch": 43,
+      "eval_accuracy": 0.5302,
+      "eval_loss": 1.5688693359375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 15136,
+      "train_examples": 45000,
+      "train_loss": 1.5281175550248889
+    },
+    {
+      "epoch": 44,
+      "eval_accuracy": 0.483,
+      "eval_loss": 1.8670911376953125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 15488,
+      "train_examples": 45000,
+      "train_loss": 1.4591134276707967
+    },
+    {
+      "epoch": 45,
+      "eval_accuracy": 0.5122,
+      "eval_loss": 1.7008560546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 15840,
+      "train_examples": 45000,
+      "train_loss": 1.4886928064134386
+    },
+    {
+      "epoch": 46,
+      "eval_accuracy": 0.5714,
+      "eval_loss": 1.379425,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 16192,
+      "train_examples": 45000,
+      "train_loss": 1.5013230222066243
+    },
+    {
+      "epoch": 47,
+      "eval_accuracy": 0.5548,
+      "eval_loss": 1.540955224609375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 16544,
+      "train_examples": 45000,
+      "train_loss": 1.423521467505561
+    },
+    {
+      "epoch": 48,
+      "eval_accuracy": 0.5066,
+      "eval_loss": 1.8178400634765626,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 16896,
+      "train_examples": 45000,
+      "train_loss": 1.473559805340237
+    },
+    {
+      "epoch": 49,
+      "eval_accuracy": 0.5778,
+      "eval_loss": 1.350400732421875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 17248,
+      "train_examples": 45000,
+      "train_loss": 1.512079693031311
+    },
+    {
+      "epoch": 50,
+      "eval_accuracy": 0.5722,
+      "eval_loss": 1.516371484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 17600,
+      "train_examples": 45000,
+      "train_loss": 1.461774546453688
+    },
+    {
+      "epoch": 51,
+      "eval_accuracy": 0.507,
+      "eval_loss": 1.701910693359375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 17952,
+      "train_examples": 45000,
+      "train_loss": 1.464529783630371
+    },
+    {
+      "epoch": 52,
+      "eval_accuracy": 0.5038,
+      "eval_loss": 1.78495341796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 18304,
+      "train_examples": 45000,
+      "train_loss": 1.5189528352101644
+    },
+    {
+      "epoch": 53,
+      "eval_accuracy": 0.5186,
+      "eval_loss": 1.6093507080078124,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 18656,
+      "train_examples": 45000,
+      "train_loss": 1.4640200068155924
+    },
+    {
+      "epoch": 54,
+      "eval_accuracy": 0.52,
+      "eval_loss": 1.6729033447265624,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 19008,
+      "train_examples": 45000,
+      "train_loss": 1.4083326490402222
+    },
+    {
+      "epoch": 55,
+      "eval_accuracy": 0.5778,
+      "eval_loss": 1.3512170654296876,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 19360,
+      "train_examples": 45000,
+      "train_loss": 1.416956041378445
+    },
+    {
+      "epoch": 56,
+      "eval_accuracy": 0.5506,
+      "eval_loss": 1.6091921142578125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 19712,
+      "train_examples": 45000,
+      "train_loss": 1.4798751659181384
+    },
+    {
+      "epoch": 57,
+      "eval_accuracy": 0.512,
+      "eval_loss": 1.6767433837890624,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 20064,
+      "train_examples": 45000,
+      "train_loss": 1.4291151117112901
+    },
+    {
+      "epoch": 58,
+      "eval_accuracy": 0.545,
+      "eval_loss": 1.5081556396484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 20416,
+      "train_examples": 45000,
+      "train_loss": 1.3918506565093993
+    },
+    {
+      "epoch": 59,
+      "eval_accuracy": 0.539,
+      "eval_loss": 1.5954337890625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 20768,
+      "train_examples": 45000,
+      "train_loss": 1.5130316936280992
+    },
+    {
+      "epoch": 60,
+      "eval_accuracy": 0.5556,
+      "eval_loss": 1.492790185546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 21120,
+      "train_examples": 45000,
+      "train_loss": 1.47905190410614
+    },
+    {
+      "epoch": 61,
+      "eval_accuracy": 0.521,
+      "eval_loss": 1.7738572021484376,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 21472,
+      "train_examples": 45000,
+      "train_loss": 1.3852702054129706
+    },
+    {
+      "epoch": 62,
+      "eval_accuracy": 0.5432,
+      "eval_loss": 1.5759424560546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 21824,
+      "train_examples": 45000,
+      "train_loss": 1.4184113253699409
+    },
+    {
+      "epoch": 63,
+      "eval_accuracy": 0.5626,
+      "eval_loss": 1.61301494140625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 22176,
+      "train_examples": 45000,
+      "train_loss": 1.5066941794501412
+    },
+    {
+      "epoch": 64,
+      "eval_accuracy": 0.5646,
+      "eval_loss": 1.675884912109375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 22528,
+      "train_examples": 45000,
+      "train_loss": 1.4340424686219957
+    },
+    {
+      "epoch": 65,
+      "eval_accuracy": 0.5372,
+      "eval_loss": 1.64473779296875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 22880,
+      "train_examples": 45000,
+      "train_loss": 1.4120538077672322
+    },
+    {
+      "epoch": 66,
+      "eval_accuracy": 0.5084,
+      "eval_loss": 1.7685004150390624,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 23232,
+      "train_examples": 45000,
+      "train_loss": 1.4140199373457167
+    },
+    {
+      "epoch": 67,
+      "eval_accuracy": 0.5182,
+      "eval_loss": 1.8401543212890625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 23584,
+      "train_examples": 45000,
+      "train_loss": 1.4402784141964382
+    },
+    {
+      "epoch": 68,
+      "eval_accuracy": 0.5674,
+      "eval_loss": 1.4143515380859375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 23936,
+      "train_examples": 45000,
+      "train_loss": 1.4850667947769165
+    },
+    {
+      "epoch": 69,
+      "eval_accuracy": 0.581,
+      "eval_loss": 1.3920767578125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 24288,
+      "train_examples": 45000,
+      "train_loss": 1.3742551008436414
+    },
+    {
+      "epoch": 70,
+      "eval_accuracy": 0.526,
+      "eval_loss": 1.6470415283203126,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 24640,
+      "train_examples": 45000,
+      "train_loss": 1.439255021180047
+    },
+    {
+      "epoch": 71,
+      "eval_accuracy": 0.5476,
+      "eval_loss": 1.4234358154296876,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 24992,
+      "train_examples": 45000,
+      "train_loss": 1.402162924469842
+    },
+    {
+      "epoch": 72,
+      "eval_accuracy": 0.527,
+      "eval_loss": 1.6882696044921874,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 25344,
+      "train_examples": 45000,
+      "train_loss": 1.3987669376797147
+    },
+    {
+      "epoch": 73,
+      "eval_accuracy": 0.5564,
+      "eval_loss": 1.527014453125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 25696,
+      "train_examples": 45000,
+      "train_loss": 1.3347418962902493
+    },
+    {
+      "epoch": 74,
+      "eval_accuracy": 0.497,
+      "eval_loss": 1.6632053955078125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 26048,
+      "train_examples": 45000,
+      "train_loss": 1.3455509513643054
+    },
+    {
+      "epoch": 75,
+      "eval_accuracy": 0.5554,
+      "eval_loss": 1.44293671875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 26400,
+      "train_examples": 45000,
+      "train_loss": 1.3510505442301433
+    },
+    {
+      "epoch": 76,
+      "eval_accuracy": 0.562,
+      "eval_loss": 1.5632307861328125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 26752,
+      "train_examples": 45000,
+      "train_loss": 1.4097723316192627
+    },
+    {
+      "epoch": 77,
+      "eval_accuracy": 0.4878,
+      "eval_loss": 1.86476484375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 27104,
+      "train_examples": 45000,
+      "train_loss": 1.3837750133726332
+    },
+    {
+      "epoch": 78,
+      "eval_accuracy": 0.5332,
+      "eval_loss": 1.5847595947265625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 27456,
+      "train_examples": 45000,
+      "train_loss": 1.396502939499749
+    },
+    {
+      "epoch": 79,
+      "eval_accuracy": 0.5112,
+      "eval_loss": 1.5564174560546875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 27808,
+      "train_examples": 45000,
+      "train_loss": 1.3709289339065551
+    },
+    {
+      "epoch": 80,
+      "eval_accuracy": 0.5394,
+      "eval_loss": 1.9737873046875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 28160,
+      "train_examples": 45000,
+      "train_loss": 1.4022963900883993
+    },
+    {
+      "epoch": 81,
+      "eval_accuracy": 0.5584,
+      "eval_loss": 1.5171572021484374,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 28512,
+      "train_examples": 45000,
+      "train_loss": 1.370726917775472
+    },
+    {
+      "epoch": 82,
+      "eval_accuracy": 0.5358,
+      "eval_loss": 1.4977206787109374,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 28864,
+      "train_examples": 45000,
+      "train_loss": 1.4500145688374837
+    },
+    {
+      "epoch": 83,
+      "eval_accuracy": 0.5152,
+      "eval_loss": 1.8416988037109374,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 29216,
+      "train_examples": 45000,
+      "train_loss": 1.4227493050893147
+    },
+    {
+      "epoch": 84,
+      "eval_accuracy": 0.5836,
+      "eval_loss": 1.352899951171875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 29568,
+      "train_examples": 45000,
+      "train_loss": 1.3736794021606444
+    },
+    {
+      "epoch": 85,
+      "eval_accuracy": 0.5584,
+      "eval_loss": 1.6942635498046874,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 29920,
+      "train_examples": 45000,
+      "train_loss": 1.389241146045261
+    },
+    {
+      "epoch": 86,
+      "eval_accuracy": 0.5596,
+      "eval_loss": 1.4475465087890624,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 30272,
+      "train_examples": 45000,
+      "train_loss": 1.270623020532396
+    },
+    {
+      "epoch": 87,
+      "eval_accuracy": 0.5794,
+      "eval_loss": 1.3361758056640625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 30624,
+      "train_examples": 45000,
+      "train_loss": 1.3661915603425767
+    },
+    {
+      "epoch": 88,
+      "eval_accuracy": 0.5566,
+      "eval_loss": 1.4134527587890624,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 30976,
+      "train_examples": 45000,
+      "train_loss": 1.3660098644044665
+    },
+    {
+      "epoch": 89,
+      "eval_accuracy": 0.53,
+      "eval_loss": 1.5531100341796875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 31328,
+      "train_examples": 45000,
+      "train_loss": 1.2967509341133965
+    },
+    {
+      "epoch": 90,
+      "eval_accuracy": 0.51,
+      "eval_loss": 1.7390384521484374,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 31680,
+      "train_examples": 45000,
+      "train_loss": 1.3260442124684653
+    },
+    {
+      "epoch": 91,
+      "eval_accuracy": 0.5874,
+      "eval_loss": 1.3916345458984376,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 32032,
+      "train_examples": 45000,
+      "train_loss": 1.3153838235431248
+    },
+    {
+      "epoch": 92,
+      "eval_accuracy": 0.5636,
+      "eval_loss": 1.4337053955078125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 32384,
+      "train_examples": 45000,
+      "train_loss": 1.2993551434834798
+    },
+    {
+      "epoch": 93,
+      "eval_accuracy": 0.533,
+      "eval_loss": 1.6667333984375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 32736,
+      "train_examples": 45000,
+      "train_loss": 1.3019694290796915
+    },
+    {
+      "epoch": 94,
+      "eval_accuracy": 0.5652,
+      "eval_loss": 1.4489898681640625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 33088,
+      "train_examples": 45000,
+      "train_loss": 1.2822739417394002
+    },
+    {
+      "epoch": 95,
+      "eval_accuracy": 0.5668,
+      "eval_loss": 1.419910986328125,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 33440,
+      "train_examples": 45000,
+      "train_loss": 1.3006300278769598
+    },
+    {
+      "epoch": 96,
+      "eval_accuracy": 0.5394,
+      "eval_loss": 1.69527265625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 33792,
+      "train_examples": 45000,
+      "train_loss": 1.238620239914788
+    },
+    {
+      "epoch": 97,
+      "eval_accuracy": 0.513,
+      "eval_loss": 1.7494765869140625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 34144,
+      "train_examples": 45000,
+      "train_loss": 1.2671517435497708
+    },
+    {
+      "epoch": 98,
+      "eval_accuracy": 0.5838,
+      "eval_loss": 1.2681653076171875,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 34496,
+      "train_examples": 45000,
+      "train_loss": 1.3089203299416436
+    },
+    {
+      "epoch": 99,
+      "eval_accuracy": 0.5436,
+      "eval_loss": 1.52252724609375,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 34848,
+      "train_examples": 45000,
+      "train_loss": 1.3226820044411554
+    },
+    {
+      "epoch": 100,
+      "eval_accuracy": 0.6308,
+      "eval_loss": 1.1525076416015625,
+      "eval_split": "validation",
+      "lr": 0.1,
+      "output_lr": 0.1,
+      "step": 35200,
+      "train_examples": 45000,
+      "train_loss": 1.2924140570958456
+    },
+    {
+      "epoch": 101,
+      "eval_accuracy": 0.682,
+      "eval_loss": 0.922958056640625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 35552,
+      "train_examples": 45000,
+      "train_loss": 0.9534723613103231
+    },
+    {
+      "epoch": 102,
+      "eval_accuracy": 0.6796,
+      "eval_loss": 0.9390528076171875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 35904,
+      "train_examples": 45000,
+      "train_loss": 0.9059994698312548
+    },
+    {
+      "epoch": 103,
+      "eval_accuracy": 0.686,
+      "eval_loss": 0.9127562133789062,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 36256,
+      "train_examples": 45000,
+      "train_loss": 0.8874733817100525
+    },
+    {
+      "epoch": 104,
+      "eval_accuracy": 0.6914,
+      "eval_loss": 0.89916376953125,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 36608,
+      "train_examples": 45000,
+      "train_loss": 0.8901613323529561
+    },
+    {
+      "epoch": 105,
+      "eval_accuracy": 0.6866,
+      "eval_loss": 0.9040415405273438,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 36960,
+      "train_examples": 45000,
+      "train_loss": 0.8719133009168837
+    },
+    {
+      "epoch": 106,
+      "eval_accuracy": 0.6934,
+      "eval_loss": 0.91411796875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 37312,
+      "train_examples": 45000,
+      "train_loss": 0.8583287696414523
+    },
+    {
+      "epoch": 107,
+      "eval_accuracy": 0.6924,
+      "eval_loss": 0.89068076171875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 37664,
+      "train_examples": 45000,
+      "train_loss": 0.8588273622512818
+    },
+    {
+      "epoch": 108,
+      "eval_accuracy": 0.6922,
+      "eval_loss": 0.9007458984375,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 38016,
+      "train_examples": 45000,
+      "train_loss": 0.8597999311447143
+    },
+    {
+      "epoch": 109,
+      "eval_accuracy": 0.7014,
+      "eval_loss": 0.8806451171875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 38368,
+      "train_examples": 45000,
+      "train_loss": 0.8486230295923021
+    },
+    {
+      "epoch": 110,
+      "eval_accuracy": 0.689,
+      "eval_loss": 0.9200911865234375,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 38720,
+      "train_examples": 45000,
+      "train_loss": 0.8488019719229805
+    },
+    {
+      "epoch": 111,
+      "eval_accuracy": 0.7036,
+      "eval_loss": 0.8662437866210938,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 39072,
+      "train_examples": 45000,
+      "train_loss": 0.8433535183800591
+    },
+    {
+      "epoch": 112,
+      "eval_accuracy": 0.6954,
+      "eval_loss": 0.8857200561523437,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 39424,
+      "train_examples": 45000,
+      "train_loss": 0.8497770226160685
+    },
+    {
+      "epoch": 113,
+      "eval_accuracy": 0.696,
+      "eval_loss": 0.8835077026367187,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 39776,
+      "train_examples": 45000,
+      "train_loss": 0.8537612689124213
+    },
+    {
+      "epoch": 114,
+      "eval_accuracy": 0.6884,
+      "eval_loss": 0.8951915893554687,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 40128,
+      "train_examples": 45000,
+      "train_loss": 0.8630593092494541
+    },
+    {
+      "epoch": 115,
+      "eval_accuracy": 0.692,
+      "eval_loss": 0.89095966796875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 40480,
+      "train_examples": 45000,
+      "train_loss": 0.8638030516730415
+    },
+    {
+      "epoch": 116,
+      "eval_accuracy": 0.7048,
+      "eval_loss": 0.877529052734375,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 40832,
+      "train_examples": 45000,
+      "train_loss": 0.8609290717230903
+    },
+    {
+      "epoch": 117,
+      "eval_accuracy": 0.7066,
+      "eval_loss": 0.8617789794921875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 41184,
+      "train_examples": 45000,
+      "train_loss": 0.8482380716853671
+    },
+    {
+      "epoch": 118,
+      "eval_accuracy": 0.6902,
+      "eval_loss": 0.901466943359375,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 41536,
+      "train_examples": 45000,
+      "train_loss": 0.8500684111277262
+    },
+    {
+      "epoch": 119,
+      "eval_accuracy": 0.6872,
+      "eval_loss": 0.9289511840820313,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 41888,
+      "train_examples": 45000,
+      "train_loss": 0.8877813004069858
+    },
+    {
+      "epoch": 120,
+      "eval_accuracy": 0.6874,
+      "eval_loss": 0.937879931640625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 42240,
+      "train_examples": 45000,
+      "train_loss": 0.9034233251783583
+    },
+    {
+      "epoch": 121,
+      "eval_accuracy": 0.6836,
+      "eval_loss": 0.9336000122070313,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 42592,
+      "train_examples": 45000,
+      "train_loss": 0.9166675170262655
+    },
+    {
+      "epoch": 122,
+      "eval_accuracy": 0.6874,
+      "eval_loss": 0.919823388671875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 42944,
+      "train_examples": 45000,
+      "train_loss": 0.8952297897338867
+    },
+    {
+      "epoch": 123,
+      "eval_accuracy": 0.6752,
+      "eval_loss": 0.9460063232421875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 43296,
+      "train_examples": 45000,
+      "train_loss": 0.9035753014034695
+    },
+    {
+      "epoch": 124,
+      "eval_accuracy": 0.681,
+      "eval_loss": 0.939022802734375,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 43648,
+      "train_examples": 45000,
+      "train_loss": 0.908453873623742
+    },
+    {
+      "epoch": 125,
+      "eval_accuracy": 0.6914,
+      "eval_loss": 0.9120480102539062,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 44000,
+      "train_examples": 45000,
+      "train_loss": 0.9134819952646891
+    },
+    {
+      "epoch": 126,
+      "eval_accuracy": 0.684,
+      "eval_loss": 0.9374011962890625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 44352,
+      "train_examples": 45000,
+      "train_loss": 0.906762312274509
+    },
+    {
+      "epoch": 127,
+      "eval_accuracy": 0.6816,
+      "eval_loss": 0.9479945434570313,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 44704,
+      "train_examples": 45000,
+      "train_loss": 0.900509298239814
+    },
+    {
+      "epoch": 128,
+      "eval_accuracy": 0.6754,
+      "eval_loss": 0.96343857421875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 45056,
+      "train_examples": 45000,
+      "train_loss": 0.8953627472347684
+    },
+    {
+      "epoch": 129,
+      "eval_accuracy": 0.6742,
+      "eval_loss": 0.968261376953125,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 45408,
+      "train_examples": 45000,
+      "train_loss": 0.9279563033845689
+    },
+    {
+      "epoch": 130,
+      "eval_accuracy": 0.6846,
+      "eval_loss": 0.9416985961914063,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 45760,
+      "train_examples": 45000,
+      "train_loss": 0.9155617316775851
+    },
+    {
+      "epoch": 131,
+      "eval_accuracy": 0.6666,
+      "eval_loss": 0.9798123779296875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 46112,
+      "train_examples": 45000,
+      "train_loss": 0.9026238541709052
+    },
+    {
+      "epoch": 132,
+      "eval_accuracy": 0.6714,
+      "eval_loss": 0.96294951171875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 46464,
+      "train_examples": 45000,
+      "train_loss": 0.9165382932874891
+    },
+    {
+      "epoch": 133,
+      "eval_accuracy": 0.678,
+      "eval_loss": 0.9479088256835938,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 46816,
+      "train_examples": 45000,
+      "train_loss": 0.8981040594206916
+    },
+    {
+      "epoch": 134,
+      "eval_accuracy": 0.6714,
+      "eval_loss": 0.9769468994140625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 47168,
+      "train_examples": 45000,
+      "train_loss": 0.9077048551559448
+    },
+    {
+      "epoch": 135,
+      "eval_accuracy": 0.6674,
+      "eval_loss": 0.966569775390625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 47520,
+      "train_examples": 45000,
+      "train_loss": 0.9450435054779053
+    },
+    {
+      "epoch": 136,
+      "eval_accuracy": 0.6658,
+      "eval_loss": 0.9645234619140625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 47872,
+      "train_examples": 45000,
+      "train_loss": 0.9295638001971774
+    },
+    {
+      "epoch": 137,
+      "eval_accuracy": 0.6238,
+      "eval_loss": 1.1051807861328125,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 48224,
+      "train_examples": 45000,
+      "train_loss": 0.95655686923133
+    },
+    {
+      "epoch": 138,
+      "eval_accuracy": 0.642,
+      "eval_loss": 1.03306103515625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 48576,
+      "train_examples": 45000,
+      "train_loss": 1.0022516270955404
+    },
+    {
+      "epoch": 139,
+      "eval_accuracy": 0.6416,
+      "eval_loss": 1.0455357788085937,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 48928,
+      "train_examples": 45000,
+      "train_loss": 0.9911704215155708
+    },
+    {
+      "epoch": 140,
+      "eval_accuracy": 0.6474,
+      "eval_loss": 1.0605947875976562,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 49280,
+      "train_examples": 45000,
+      "train_loss": 0.9922565679762099
+    },
+    {
+      "epoch": 141,
+      "eval_accuracy": 0.6374,
+      "eval_loss": 1.07503681640625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 49632,
+      "train_examples": 45000,
+      "train_loss": 1.0034182527012294
+    },
+    {
+      "epoch": 142,
+      "eval_accuracy": 0.641,
+      "eval_loss": 1.061389990234375,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 49984,
+      "train_examples": 45000,
+      "train_loss": 1.0350198850949606
+    },
+    {
+      "epoch": 143,
+      "eval_accuracy": 0.6272,
+      "eval_loss": 1.091273046875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 50336,
+      "train_examples": 45000,
+      "train_loss": 1.015117134115431
+    },
+    {
+      "epoch": 144,
+      "eval_accuracy": 0.634,
+      "eval_loss": 1.0583884765625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 50688,
+      "train_examples": 45000,
+      "train_loss": 1.0244670013639663
+    },
+    {
+      "epoch": 145,
+      "eval_accuracy": 0.653,
+      "eval_loss": 1.0368972412109374,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 51040,
+      "train_examples": 45000,
+      "train_loss": 1.0196523662779067
+    },
+    {
+      "epoch": 146,
+      "eval_accuracy": 0.6564,
+      "eval_loss": 1.007610009765625,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 51392,
+      "train_examples": 45000,
+      "train_loss": 0.981625235833062
+    },
+    {
+      "epoch": 147,
+      "eval_accuracy": 0.6548,
+      "eval_loss": 1.0109105224609376,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 51744,
+      "train_examples": 45000,
+      "train_loss": 0.9948846951166789
+    },
+    {
+      "epoch": 148,
+      "eval_accuracy": 0.6306,
+      "eval_loss": 1.0539881225585936,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 52096,
+      "train_examples": 45000,
+      "train_loss": 1.052033565235138
+    },
+    {
+      "epoch": 149,
+      "eval_accuracy": 0.61,
+      "eval_loss": 1.15952216796875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 52448,
+      "train_examples": 45000,
+      "train_loss": 1.0489593990325927
+    },
+    {
+      "epoch": 150,
+      "eval_accuracy": 0.6218,
+      "eval_loss": 1.085763623046875,
+      "eval_split": "validation",
+      "lr": 0.010000000000000002,
+      "output_lr": 0.010000000000000002,
+      "step": 52800,
+      "train_examples": 45000,
+      "train_loss": 1.1033223127788967
+    },
+    {
+      "epoch": 151,
+      "eval_accuracy": 0.654,
+      "eval_loss": 1.0092583862304687,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 53152,
+      "train_examples": 45000,
+      "train_loss": 1.0319644333945381
+    },
+    {
+      "epoch": 152,
+      "eval_accuracy": 0.676,
+      "eval_loss": 0.9723513427734375,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 53504,
+      "train_examples": 45000,
+      "train_loss": 0.994410558075375
+    },
+    {
+      "epoch": 153,
+      "eval_accuracy": 0.6766,
+      "eval_loss": 0.9589015747070313,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 53856,
+      "train_examples": 45000,
+      "train_loss": 0.9668529237747192
+    },
+    {
+      "epoch": 154,
+      "eval_accuracy": 0.6738,
+      "eval_loss": 0.95261396484375,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 54208,
+      "train_examples": 45000,
+      "train_loss": 0.9483775505489773
+    },
+    {
+      "epoch": 155,
+      "eval_accuracy": 0.6798,
+      "eval_loss": 0.9446010498046875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 54560,
+      "train_examples": 45000,
+      "train_loss": 0.9422066597832573
+    },
+    {
+      "epoch": 156,
+      "eval_accuracy": 0.6814,
+      "eval_loss": 0.9320489990234375,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 54912,
+      "train_examples": 45000,
+      "train_loss": 0.9290729566150241
+    },
+    {
+      "epoch": 157,
+      "eval_accuracy": 0.6882,
+      "eval_loss": 0.9252060791015625,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 55264,
+      "train_examples": 45000,
+      "train_loss": 0.9260442474683126
+    },
+    {
+      "epoch": 158,
+      "eval_accuracy": 0.6916,
+      "eval_loss": 0.9229399536132813,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 55616,
+      "train_examples": 45000,
+      "train_loss": 0.9205480414920383
+    },
+    {
+      "epoch": 159,
+      "eval_accuracy": 0.6904,
+      "eval_loss": 0.9217549682617188,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 55968,
+      "train_examples": 45000,
+      "train_loss": 0.9238852113935683
+    },
+    {
+      "epoch": 160,
+      "eval_accuracy": 0.6848,
+      "eval_loss": 0.9292841796875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 56320,
+      "train_examples": 45000,
+      "train_loss": 0.9197275682343377
+    },
+    {
+      "epoch": 161,
+      "eval_accuracy": 0.6886,
+      "eval_loss": 0.9162351684570312,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 56672,
+      "train_examples": 45000,
+      "train_loss": 0.9102767890082465
+    },
+    {
+      "epoch": 162,
+      "eval_accuracy": 0.6914,
+      "eval_loss": 0.9114646118164063,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 57024,
+      "train_examples": 45000,
+      "train_loss": 0.9038821694480048
+    },
+    {
+      "epoch": 163,
+      "eval_accuracy": 0.697,
+      "eval_loss": 0.9089584228515625,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 57376,
+      "train_examples": 45000,
+      "train_loss": 0.9043405807918973
+    },
+    {
+      "epoch": 164,
+      "eval_accuracy": 0.6988,
+      "eval_loss": 0.9067180297851563,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 57728,
+      "train_examples": 45000,
+      "train_loss": 0.8994147229936388
+    },
+    {
+      "epoch": 165,
+      "eval_accuracy": 0.6932,
+      "eval_loss": 0.9099502319335937,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 58080,
+      "train_examples": 45000,
+      "train_loss": 0.9026298767301771
+    },
+    {
+      "epoch": 166,
+      "eval_accuracy": 0.6942,
+      "eval_loss": 0.9057412231445312,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 58432,
+      "train_examples": 45000,
+      "train_loss": 0.9046581612269083
+    },
+    {
+      "epoch": 167,
+      "eval_accuracy": 0.6932,
+      "eval_loss": 0.906397509765625,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 58784,
+      "train_examples": 45000,
+      "train_loss": 0.9021518977906969
+    },
+    {
+      "epoch": 168,
+      "eval_accuracy": 0.6908,
+      "eval_loss": 0.9119505981445313,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 59136,
+      "train_examples": 45000,
+      "train_loss": 0.9030990373081631
+    },
+    {
+      "epoch": 169,
+      "eval_accuracy": 0.6868,
+      "eval_loss": 0.9118242309570312,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 59488,
+      "train_examples": 45000,
+      "train_loss": 0.9032874836497836
+    },
+    {
+      "epoch": 170,
+      "eval_accuracy": 0.6822,
+      "eval_loss": 0.92456123046875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 59840,
+      "train_examples": 45000,
+      "train_loss": 0.9015067362467448
+    },
+    {
+      "epoch": 171,
+      "eval_accuracy": 0.686,
+      "eval_loss": 0.9098457397460937,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 60192,
+      "train_examples": 45000,
+      "train_loss": 0.9019311337153116
+    },
+    {
+      "epoch": 172,
+      "eval_accuracy": 0.6856,
+      "eval_loss": 0.9247203979492188,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 60544,
+      "train_examples": 45000,
+      "train_loss": 0.9169406125810411
+    },
+    {
+      "epoch": 173,
+      "eval_accuracy": 0.6864,
+      "eval_loss": 0.9248978637695312,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 60896,
+      "train_examples": 45000,
+      "train_loss": 0.9140570906321208
+    },
+    {
+      "epoch": 174,
+      "eval_accuracy": 0.686,
+      "eval_loss": 0.916908447265625,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 61248,
+      "train_examples": 45000,
+      "train_loss": 0.9104670696258544
+    },
+    {
+      "epoch": 175,
+      "eval_accuracy": 0.686,
+      "eval_loss": 0.92089794921875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 61600,
+      "train_examples": 45000,
+      "train_loss": 0.9013468429777357
+    },
+    {
+      "epoch": 176,
+      "eval_accuracy": 0.6868,
+      "eval_loss": 0.9163492797851562,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 61952,
+      "train_examples": 45000,
+      "train_loss": 0.9033487060334947
+    },
+    {
+      "epoch": 177,
+      "eval_accuracy": 0.6832,
+      "eval_loss": 0.9202944091796875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 62304,
+      "train_examples": 45000,
+      "train_loss": 0.9047088833385044
+    },
+    {
+      "epoch": 178,
+      "eval_accuracy": 0.6766,
+      "eval_loss": 0.9400230834960938,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 62656,
+      "train_examples": 45000,
+      "train_loss": 0.915128051270379
+    },
+    {
+      "epoch": 179,
+      "eval_accuracy": 0.68,
+      "eval_loss": 0.9347542114257813,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 63008,
+      "train_examples": 45000,
+      "train_loss": 0.9253638660642836
+    },
+    {
+      "epoch": 180,
+      "eval_accuracy": 0.6784,
+      "eval_loss": 0.9496093017578126,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 63360,
+      "train_examples": 45000,
+      "train_loss": 0.9585485564867655
+    },
+    {
+      "epoch": 181,
+      "eval_accuracy": 0.6754,
+      "eval_loss": 0.9441274047851562,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 63712,
+      "train_examples": 45000,
+      "train_loss": 0.9477499859068129
+    },
+    {
+      "epoch": 182,
+      "eval_accuracy": 0.6798,
+      "eval_loss": 0.93713427734375,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 64064,
+      "train_examples": 45000,
+      "train_loss": 0.9405778836568197
+    },
+    {
+      "epoch": 183,
+      "eval_accuracy": 0.6762,
+      "eval_loss": 0.9379406616210938,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 64416,
+      "train_examples": 45000,
+      "train_loss": 0.9201891781171163
+    },
+    {
+      "epoch": 184,
+      "eval_accuracy": 0.6798,
+      "eval_loss": 0.9241371215820312,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 64768,
+      "train_examples": 45000,
+      "train_loss": 0.9184683513959249
+    },
+    {
+      "epoch": 185,
+      "eval_accuracy": 0.681,
+      "eval_loss": 0.9289290649414063,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 65120,
+      "train_examples": 45000,
+      "train_loss": 0.9075623473697239
+    },
+    {
+      "epoch": 186,
+      "eval_accuracy": 0.682,
+      "eval_loss": 0.9222794555664062,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 65472,
+      "train_examples": 45000,
+      "train_loss": 0.9138578772544861
+    },
+    {
+      "epoch": 187,
+      "eval_accuracy": 0.6778,
+      "eval_loss": 0.9411468017578125,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 65824,
+      "train_examples": 45000,
+      "train_loss": 0.917494329823388
+    },
+    {
+      "epoch": 188,
+      "eval_accuracy": 0.6734,
+      "eval_loss": 0.952200244140625,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 66176,
+      "train_examples": 45000,
+      "train_loss": 0.9213011764632331
+    },
+    {
+      "epoch": 189,
+      "eval_accuracy": 0.6812,
+      "eval_loss": 0.9310021484375,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 66528,
+      "train_examples": 45000,
+      "train_loss": 0.9217213383992513
+    },
+    {
+      "epoch": 190,
+      "eval_accuracy": 0.682,
+      "eval_loss": 0.9291856201171875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 66880,
+      "train_examples": 45000,
+      "train_loss": 0.9160469234360589
+    },
+    {
+      "epoch": 191,
+      "eval_accuracy": 0.6774,
+      "eval_loss": 0.93428427734375,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 67232,
+      "train_examples": 45000,
+      "train_loss": 0.9122076215638055
+    },
+    {
+      "epoch": 192,
+      "eval_accuracy": 0.6754,
+      "eval_loss": 0.944618798828125,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 67584,
+      "train_examples": 45000,
+      "train_loss": 0.9217920794275072
+    },
+    {
+      "epoch": 193,
+      "eval_accuracy": 0.6768,
+      "eval_loss": 0.9403953857421875,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 67936,
+      "train_examples": 45000,
+      "train_loss": 0.9300884993659125
+    },
+    {
+      "epoch": 194,
+      "eval_accuracy": 0.6846,
+      "eval_loss": 0.9310226684570313,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 68288,
+      "train_examples": 45000,
+      "train_loss": 0.9326077370961507
+    },
+    {
+      "epoch": 195,
+      "eval_accuracy": 0.6812,
+      "eval_loss": 0.9314572143554688,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 68640,
+      "train_examples": 45000,
+      "train_loss": 0.9251852588123746
+    },
+    {
+      "epoch": 196,
+      "eval_accuracy": 0.6792,
+      "eval_loss": 0.9355191040039063,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 68992,
+      "train_examples": 45000,
+      "train_loss": 0.924093881840176
+    },
+    {
+      "epoch": 197,
+      "eval_accuracy": 0.6796,
+      "eval_loss": 0.9381912963867187,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 69344,
+      "train_examples": 45000,
+      "train_loss": 0.9298513369878133
+    },
+    {
+      "epoch": 198,
+      "eval_accuracy": 0.6796,
+      "eval_loss": 0.9357055297851562,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 69696,
+      "train_examples": 45000,
+      "train_loss": 0.9319831230163574
+    },
+    {
+      "epoch": 199,
+      "eval_accuracy": 0.6888,
+      "eval_loss": 0.9251195556640625,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 70048,
+      "train_examples": 45000,
+      "train_loss": 0.9235278500239055
+    },
+    {
+      "epoch": 200,
+      "eval_accuracy": 0.6854,
+      "eval_loss": 0.9467566040039063,
+      "eval_split": "validation",
+      "lr": 0.0010000000000000002,
+      "output_lr": 0.0010000000000000002,
+      "step": 70400,
+      "train_examples": 45000,
+      "train_loss": 0.9284468433168199
+    }
+  ],
+  "evaluation_protocol": {
+    "test_evaluations": 0,
+    "test_used_for_selection": false,
+    "validation_evaluations": 201
+  },
+  "final": {
+    "accuracy": 0.6854,
+    "epoch": 200,
+    "evaluation_split": "validation",
+    "finite": true,
+    "loss": 0.9467566040039063,
+    "step": 70400
+  },
+  "hardware": {
+    "cuda_device_name": "NVIDIA RTX A6000",
+    "cuda_visible_devices": "0",
+    "device": "cuda",
+    "device_total_memory_bytes": 50896568320,
+    "peak_memory_allocated_bytes": 2009673728,
+    "peak_memory_reserved_bytes": 2550136832,
+    "torch_version": "2.3.1+cu118"
+  },
+  "protocol_family": "oral_a_cifar_local_resnet_development",
+  "provenance": {
+    "git_commit": "f55872beaee658ff76ac15fd421b73017820c320",
+    "git_tracked_dirty": false
+  },
+  "schema_version": 1,
+  "split": {
+    "cifar_source_files": [
+      {
+        "bytes": 31035704,
+        "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_1",
+        "sha256": "54636561a3ce25bd3e19253c6b0d8538147b0ae398331ac4a2d86c6d987368cd"
+      },
+      {
+        "bytes": 31035320,
+        "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_2",
+        "sha256": "766b2cef9fbc745cf056b3152224f7cf77163b330ea9a15f9392beb8b89bc5a8"
+      },
+      {
+        "bytes": 31035999,
+        "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_3",
+        "sha256": "0f00d98ebfb30b3ec0ad19f9756dc2630b89003e10525f5e148445e82aa6a1f9"
+      },
+      {
+        "bytes": 31035696,
+        "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_4",
+        "sha256": "3f7bb240661948b8f4d53e36ec720d8306f5668bd0071dcb4e6c947f78e9682b"
+      },
+      {
+        "bytes": 31035623,
+        "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_5",
+        "sha256": "d91802434d8376bbaeeadf58a737e3a1b12ac839077e931237e0dcd43adcb154"
+      },
+      {
+        "bytes": 31035526,
+        "path": "/orion/u/oscarwan/scratch/sdil-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": 21.07404398918152,
+    "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": 1361.8625671863556,
+    "train_wall_s": 1340.767660856247
+  },
+  "work": {
+    "apical_macs_per_example": 40108672,
+    "causal_scalar_observations": 0,
+    "components": {
+      "apical_projection_macs": 360978048000000,
+      "apical_regression_macs": 0,
+      "bp_reverse_macs_estimate": 0,
+      "kp_reciprocal_correlation_macs": 0,
+      "local_weight_correlation_macs": 364959360000000,
+      "mirror_feedback_prediction_macs": 0,
+      "mirror_local_correlation_macs": 0,
+      "mirror_response_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; mixed-traffic/predictor elementwise arithmetic is reported as a conservative operation estimate separately and is not folded into MACs",
+    "elementwise_operations_estimate": 0,
+    "forward_macs_per_example": 40551040,
+    "logical_batch_loss_queries": 0,
+    "mirror_probe_examples": 0,
+    "mirror_readout_probe_examples": 0,
+    "neutral_projection_observations": 0,
+    "per_example_cross_entropy_terms": 0,
+    "total_clean_forward_examples": 9000000,
+    "total_forward_equivalent_examples": 9000000,
+    "total_macs_estimate": 1090896768000000
+  }
+}