{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 42724,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.6355140186915887e-06,
      "loss": 10.7543,
      "theoretical_loss": 20.81281780154715,
      "tokens_seen": 65536
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.177570093457944e-05,
      "loss": 8.6636,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001635514018691589,
      "loss": 6.6432,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002453271028037383,
      "loss": 5.9475,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003271028037383178,
      "loss": 5.5534,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004088785046728972,
      "loss": 5.3664,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004906542056074766,
      "loss": 5.1899,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005724299065420561,
      "loss": 5.0234,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006542056074766356,
      "loss": 4.9249,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006996358993758275,
      "loss": 4.7711,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006988083979572536,
      "loss": 4.7076,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979808965386797,
      "loss": 4.5841,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000697153395120106,
      "loss": 4.5051,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006963258937015321,
      "loss": 4.4155,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954983922829582,
      "loss": 4.3751,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006946708908643843,
      "loss": 4.2199,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006938433894458105,
      "loss": 4.189,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006930158880272367,
      "loss": 4.1234,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921883866086628,
      "loss": 4.051,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006913608851900889,
      "loss": 3.9171,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690533383771515,
      "loss": 3.8106,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897058823529412,
      "loss": 3.7893,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888783809343674,
      "loss": 3.7301,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006880508795157935,
      "loss": 3.6744,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006872233780972196,
      "loss": 3.5873,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863958766786457,
      "loss": 3.5755,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006855683752600718,
      "loss": 3.5132,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684740873841498,
      "loss": 3.5089,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839133724229242,
      "loss": 3.4237,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830858710043503,
      "loss": 3.4596,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006822583695857764,
      "loss": 3.4046,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006814308681672025,
      "loss": 3.4409,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006806033667486286,
      "loss": 3.4431,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797758653300548,
      "loss": 3.4137,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000678948363911481,
      "loss": 3.3936,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781208624929071,
      "loss": 3.3917,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772933610743332,
      "loss": 3.3676,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006764658596557593,
      "loss": 3.3325,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006756383582371856,
      "loss": 3.345,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748108568186117,
      "loss": 3.371,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739833554000378,
      "loss": 3.2953,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673155853981464,
      "loss": 3.3211,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006723283525628902,
      "loss": 3.2745,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715008511443163,
      "loss": 3.289,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006706733497257424,
      "loss": 3.2753,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006698458483071685,
      "loss": 3.2546,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690183468885946,
      "loss": 3.2145,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681908454700209,
      "loss": 3.2252,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667363344051447,
      "loss": 3.2379,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006665358426328731,
      "loss": 3.2175,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.027871694415807724,
      "objective/train/avg_weight": 0.9700481295585632,
      "objective/train/docs_used": 104000,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.27850341796875,
      "objective/train/std_weight": 0.08482376486063004,
      "objective/train/theoretical_loss": 4.467096263641219,
      "objective/train/tokens_used": 184300000,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657083412142992,
      "loss": 3.2065,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648808397957253,
      "loss": 3.2165,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006640533383771514,
      "loss": 3.168,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632258369585777,
      "loss": 3.118,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623983355400038,
      "loss": 3.1702,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006615708341214299,
      "loss": 3.1235,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660743332702856,
      "loss": 3.1592,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599158312842821,
      "loss": 3.1533,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590883298657083,
      "loss": 3.1398,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006582608284471345,
      "loss": 3.1768,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574333270285606,
      "loss": 3.136,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566058256099867,
      "loss": 3.134,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006557783241914128,
      "loss": 3.0886,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006549508227728391,
      "loss": 3.1173,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006541233213542652,
      "loss": 3.1278,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532958199356913,
      "loss": 3.0931,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006524683185171174,
      "loss": 3.0638,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006516408170985437,
      "loss": 3.1152,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508133156799698,
      "loss": 3.1342,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499858142613959,
      "loss": 3.1163,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649158312842822,
      "loss": 3.152,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006483308114242481,
      "loss": 3.1198,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475033100056744,
      "loss": 3.0873,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006466758085871005,
      "loss": 3.0382,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006458483071685266,
      "loss": 3.0672,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450208057499527,
      "loss": 3.0837,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006441933043313788,
      "loss": 3.1016,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006433658029128049,
      "loss": 3.0698,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425383014942312,
      "loss": 3.0794,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417108000756573,
      "loss": 3.0643,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006408832986570834,
      "loss": 3.0542,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006400557972385095,
      "loss": 2.9911,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392282958199356,
      "loss": 2.9473,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384007944013618,
      "loss": 2.9018,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637573292982788,
      "loss": 2.9774,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006367457915642141,
      "loss": 2.9891,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359182901456402,
      "loss": 2.952,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006350907887270663,
      "loss": 2.9969,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006342632873084925,
      "loss": 3.0117,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006334357858899187,
      "loss": 2.9764,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326082844713448,
      "loss": 2.997,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006317807830527709,
      "loss": 2.9709,
      "theoretical_loss": 4.149370777139286,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309532816341972,
      "loss": 2.968,
      "theoretical_loss": 4.144052829464249,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301257802156233,
      "loss": 2.9237,
      "theoretical_loss": 4.138808360656742,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006292982787970494,
      "loss": 2.9387,
      "theoretical_loss": 4.133635581570836,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006284707773784755,
      "loss": 2.9275,
      "theoretical_loss": 4.128532765070004,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006276432759599016,
      "loss": 2.9465,
      "theoretical_loss": 4.123498243252032,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268157745413279,
      "loss": 2.9295,
      "theoretical_loss": 4.118530404825556,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000625988273122754,
      "loss": 2.9456,
      "theoretical_loss": 4.113627692628464,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006251607717041801,
      "loss": 2.949,
      "theoretical_loss": 4.108788601279149,
      "tokens_seen": 324403200
    },
    {
      "debugging/Self-BLEU-5": 0.5365128506817183,
      "debugging/distinct-1-grams": 0.7612814402327299,
      "debugging/distinct-2-grams": 0.9694583753853511,
      "debugging/entropy-1-grams": 6.003629944255698,
      "debugging/entropy-2-grams": 7.054987089269872,
      "debugging/length": 495.25,
      "debugging/num_segments": 16,
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.04385810345411301,
      "objective/train/avg_weight": 0.9528664350509644,
      "objective/train/docs_used": 197327,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1396584510803223,
      "objective/train/std_weight": 0.1686321198940277,
      "objective/train/theoretical_loss": 4.10401167495222,
      "objective/train/tokens_used": 348140000,
      "theoretical_loss": 4.10401167495222,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243332702856062,
      "loss": 2.9772,
      "theoretical_loss": 4.10401167495222,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235057688670323,
      "loss": 2.9764,
      "theoretical_loss": 4.099295505270921,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006226782674484584,
      "loss": 2.9224,
      "theoretical_loss": 4.094638729309031,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218507660298847,
      "loss": 2.9348,
      "theoretical_loss": 4.090040027695556,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210232646113108,
      "loss": 2.8915,
      "theoretical_loss": 4.085498122815992,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006201957631927369,
      "loss": 2.9147,
      "theoretical_loss": 4.081011777104333,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619368261774163,
      "loss": 2.9209,
      "theoretical_loss": 4.076579791420469,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185407603555891,
      "loss": 2.9303,
      "theoretical_loss": 4.0722010035079155,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177132589370153,
      "loss": 2.9261,
      "theoretical_loss": 4.067874286527197,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006168857575184414,
      "loss": 2.924,
      "theoretical_loss": 4.063598547660519,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160582560998676,
      "loss": 2.9546,
      "theoretical_loss": 4.05937272678363,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152307546812937,
      "loss": 2.9394,
      "theoretical_loss": 4.055195795201069,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144032532627198,
      "loss": 2.8913,
      "theoretical_loss": 4.051066754441235,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613575751844146,
      "loss": 2.8973,
      "theoretical_loss": 4.04698463510794,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127482504255721,
      "loss": 2.901,
      "theoretical_loss": 4.042948495785312,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119207490069983,
      "loss": 2.8647,
      "theoretical_loss": 4.038957421993153,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006110932475884244,
      "loss": 2.8707,
      "theoretical_loss": 4.035010525189982,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102657461698505,
      "loss": 2.904,
      "theoretical_loss": 4.031106941821218,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094382447512768,
      "loss": 2.8677,
      "theoretical_loss": 4.027245832410079,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086107433327029,
      "loss": 2.8421,
      "theoretical_loss": 4.023426380688943,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607783241914129,
      "loss": 2.8849,
      "theoretical_loss": 4.019647792769048,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006069557404955551,
      "loss": 2.9447,
      "theoretical_loss": 4.015909296346521,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061447891053528,
      "loss": 2.9014,
      "theoretical_loss": 4.012210139942894,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053172876867789,
      "loss": 2.9463,
      "theoretical_loss": 4.008549592178291,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604489786268205,
      "loss": 2.8978,
      "theoretical_loss": 4.004926941075674,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036622848496311,
      "loss": 2.9521,
      "theoretical_loss": 4.001341493394558,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028347834310573,
      "loss": 2.9652,
      "theoretical_loss": 3.997792573992726,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020072820124835,
      "loss": 2.9098,
      "theoretical_loss": 3.994279525214554,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011797805939096,
      "loss": 2.9572,
      "theoretical_loss": 3.990801706304647,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003522791753358,
      "loss": 2.9129,
      "theoretical_loss": 3.987358492845532,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995247777567619,
      "loss": 2.9439,
      "theoretical_loss": 3.9839492762182647,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000598697276338188,
      "loss": 2.9134,
      "theoretical_loss": 3.9805734630848306,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978697749196142,
      "loss": 2.8965,
      "theoretical_loss": 3.9772304748913054,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970422735010403,
      "loss": 2.8898,
      "theoretical_loss": 3.973919747390801,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962147720824664,
      "loss": 2.9002,
      "theoretical_loss": 3.9706407301852487,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953872706638926,
      "loss": 2.8783,
      "theoretical_loss": 3.9673928862851655,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945597692453187,
      "loss": 2.8682,
      "theoretical_loss": 3.9641756916865463,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937322678267449,
      "loss": 2.8783,
      "theoretical_loss": 3.960988634964113,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592904766408171,
      "loss": 2.8511,
      "theoretical_loss": 3.9578312168801597,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920772649895971,
      "loss": 2.851,
      "theoretical_loss": 3.954702950008308,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912497635710232,
      "loss": 2.835,
      "theoretical_loss": 3.9516033583714734,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005904222621524494,
      "loss": 2.8607,
      "theoretical_loss": 3.9485319770934355,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005895947607338756,
      "loss": 2.8461,
      "theoretical_loss": 3.945488352063391,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887672593153017,
      "loss": 2.8307,
      "theoretical_loss": 3.942472039612926,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879397578967278,
      "loss": 2.8269,
      "theoretical_loss": 3.939482606204863,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005871122564781539,
      "loss": 2.8407,
      "theoretical_loss": 3.936519628133466,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058628475505958,
      "loss": 2.8754,
      "theoretical_loss": 3.9335826912355114,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854572536410061,
      "loss": 2.9168,
      "theoretical_loss": 3.93067139061177,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005846297522224324,
      "loss": 2.8452,
      "theoretical_loss": 3.927785330358441,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005838022508038585,
      "loss": 2.8359,
      "theoretical_loss": 3.9249241233081333,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.004644907079637051,
      "objective/train/avg_weight": 0.995009183883667,
      "objective/train/docs_used": 287192,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.014026165008545,
      "objective/train/std_weight": 0.02483248896896839,
      "objective/train/theoretical_loss": 3.92208739077998,
      "objective/train/tokens_used": 511980000,
      "theoretical_loss": 3.92208739077998,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829747493852846,
      "loss": 2.8155,
      "theoretical_loss": 3.92208739077998,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821472479667107,
      "loss": 2.8248,
      "theoretical_loss": 3.919274762338519,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000581319746548137,
      "loss": 2.8481,
      "theoretical_loss": 3.9164858755609613,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005804922451295631,
      "loss": 2.8296,
      "theoretical_loss": 3.9137203758125176,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796647437109892,
      "loss": 2.7929,
      "theoretical_loss": 3.910977916029439,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788372422924154,
      "loss": 2.8029,
      "theoretical_loss": 3.908258156509472,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005780097408738415,
      "loss": 2.7949,
      "theoretical_loss": 3.905560764709417,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771822394552677,
      "loss": 2.7728,
      "theoretical_loss": 3.9028854150495143,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005763547380366938,
      "loss": 2.7991,
      "theoretical_loss": 3.9002317887243834,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755272366181199,
      "loss": 2.7956,
      "theoretical_loss": 3.897599573520247,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746997351995461,
      "loss": 2.8261,
      "theoretical_loss": 3.8949884636382106,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738722337809722,
      "loss": 2.8118,
      "theoretical_loss": 3.892398159523345,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730447323623984,
      "loss": 2.8643,
      "theoretical_loss": 3.889828367699349,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005722172309438245,
      "loss": 2.8021,
      "theoretical_loss": 3.8872788006085894,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713897295252506,
      "loss": 2.8285,
      "theoretical_loss": 3.8847491764572926,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705622281066767,
      "loss": 2.8119,
      "theoretical_loss": 3.882239219065708,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697347266881029,
      "loss": 2.8372,
      "theoretical_loss": 3.879748657723039,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005689072252695291,
      "loss": 2.8645,
      "theoretical_loss": 3.8772772270469824,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680797238509552,
      "loss": 2.8276,
      "theoretical_loss": 3.8748246668476827,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672522224323813,
      "loss": 2.8475,
      "theoretical_loss": 3.8723907219959486,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005664247210138074,
      "loss": 2.8644,
      "theoretical_loss": 3.869975142295573,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655972195952335,
      "loss": 2.9145,
      "theoretical_loss": 3.8675776823595998,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647697181766596,
      "loss": 2.8816,
      "theoretical_loss": 3.8651981014904027,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005639422167580859,
      "loss": 2.8532,
      "theoretical_loss": 3.8628361635634265,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005631312653678835,
      "loss": 2.8788,
      "theoretical_loss": 3.8604916369144666,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005623037639493096,
      "loss": 2.8551,
      "theoretical_loss": 3.858164294230354,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614762625307357,
      "loss": 2.8309,
      "theoretical_loss": 3.85585391244293,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606487611121619,
      "loss": 2.8414,
      "theoretical_loss": 3.8535602726261864,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559821259693588,
      "loss": 2.8235,
      "theoretical_loss": 3.851283159896468,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589937582750142,
      "loss": 2.8092,
      "theoretical_loss": 3.8490223633156173,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581662568564403,
      "loss": 2.7915,
      "theoretical_loss": 3.846777675796974,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005573387554378664,
      "loss": 2.7981,
      "theoretical_loss": 3.844548894014116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005565112540192926,
      "loss": 2.8022,
      "theoretical_loss": 3.8423358183122582,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556837526007187,
      "loss": 2.7754,
      "theoretical_loss": 3.840138252622208,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548562511821448,
      "loss": 2.8037,
      "theoretical_loss": 3.837956004376799,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554028749763571,
      "loss": 2.7709,
      "theoretical_loss": 3.8357888844297094,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532012483449971,
      "loss": 2.7927,
      "theoretical_loss": 3.8336367069765958,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523737469264233,
      "loss": 2.7451,
      "theoretical_loss": 3.8314992894784536,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005515462455078494,
      "loss": 2.7896,
      "theoretical_loss": 3.829376452587134,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005507187440892756,
      "loss": 2.8047,
      "theoretical_loss": 3.827268020072948,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498912426707017,
      "loss": 2.8391,
      "theoretical_loss": 3.8251738187542843,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490637412521278,
      "loss": 2.8195,
      "theoretical_loss": 3.8230936784291787,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005482362398335541,
      "loss": 2.8531,
      "theoretical_loss": 3.8210274318087656,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005474087384149802,
      "loss": 2.8001,
      "theoretical_loss": 3.818974914452557,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465812369964063,
      "loss": 2.7996,
      "theoretical_loss": 3.8169359647054835,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005457537355778324,
      "loss": 2.7817,
      "theoretical_loss": 3.8149104236366433,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005449262341592585,
      "loss": 2.7613,
      "theoretical_loss": 3.8128981349797098,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440987327406847,
      "loss": 2.7851,
      "theoretical_loss": 3.8108989450749293,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005432712313221109,
      "loss": 2.7738,
      "theoretical_loss": 3.8089127028126764,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000542443729903537,
      "loss": 2.8114,
      "theoretical_loss": 3.8069392595785083,
      "tokens_seen": 652083200
    },
    {
      "debugging/Self-BLEU-5": 0.5265375629586004,
      "debugging/distinct-1-grams": 0.7435820408094715,
      "debugging/distinct-2-grams": 0.9558103821233092,
      "debugging/entropy-1-grams": 5.931434510687563,
      "debugging/entropy-2-grams": 6.886416755326388,
      "debugging/length": 521.9230769230769,
      "debugging/num_segments": 13,
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.022742915898561478,
      "objective/train/avg_weight": 0.9755392670631409,
      "objective/train/docs_used": 379091,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0632386207580566,
      "objective/train/std_weight": 0.08433963358402252,
      "objective/train/theoretical_loss": 3.804978469199669,
      "objective/train/tokens_used": 675820000,
      "theoretical_loss": 3.804978469199669,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005416162284849631,
      "loss": 2.805,
      "theoretical_loss": 3.804978469199669,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407887270663892,
      "loss": 2.8212,
      "theoretical_loss": 3.803030187893005,
      "tokens_seen": 658636800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399612256478154,
      "loss": 2.82,
      "theoretical_loss": 3.8010942742142415,
      "tokens_seen": 661913600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391337242292415,
      "loss": 2.8474,
      "theoretical_loss": 3.799170589008585,
      "tokens_seen": 665190400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383062228106677,
      "loss": 2.8322,
      "theoretical_loss": 3.7972589953626006,
      "tokens_seen": 668467200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005374787213920938,
      "loss": 2.8432,
      "theoretical_loss": 3.795359358557337,
      "tokens_seen": 671744000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005366512199735199,
      "loss": 2.7784,
      "theoretical_loss": 3.79347154602265,
      "tokens_seen": 675020800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005358237185549461,
      "loss": 2.824,
      "theoretical_loss": 3.7915954272926955,
      "tokens_seen": 678297600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349962171363722,
      "loss": 2.7824,
      "theoretical_loss": 3.789730873962557,
      "tokens_seen": 681574400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005341687157177983,
      "loss": 2.7185,
      "theoretical_loss": 3.787877759645963,
      "tokens_seen": 684851200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005333412142992244,
      "loss": 2.7511,
      "theoretical_loss": 3.7860359599340776,
      "tokens_seen": 688128000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005325137128806506,
      "loss": 2.791,
      "theoretical_loss": 3.784205352355321,
      "tokens_seen": 691404800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316862114620768,
      "loss": 2.8214,
      "theoretical_loss": 3.782385816336189,
      "tokens_seen": 694681600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005308587100435029,
      "loss": 2.7788,
      "theoretical_loss": 3.7805772331630516,
      "tokens_seen": 697958400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005300312086249291,
      "loss": 2.7553,
      "theoretical_loss": 3.7787794859448898,
      "tokens_seen": 701235200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292037072063552,
      "loss": 2.7382,
      "theoretical_loss": 3.7769924595769546,
      "tokens_seen": 704512000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283762057877813,
      "loss": 2.7255,
      "theoretical_loss": 3.7752160407053115,
      "tokens_seen": 707788800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005275487043692075,
      "loss": 2.7101,
      "theoretical_loss": 3.7734501176922493,
      "tokens_seen": 711065600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005267212029506337,
      "loss": 2.7514,
      "theoretical_loss": 3.7716945805825337,
      "tokens_seen": 714342400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258937015320598,
      "loss": 2.7886,
      "theoretical_loss": 3.7699493210704667,
      "tokens_seen": 717619200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250662001134859,
      "loss": 2.7644,
      "theoretical_loss": 3.7682142324677455,
      "tokens_seen": 720896000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000524238698694912,
      "loss": 2.8249,
      "theoretical_loss": 3.7664892096720886,
      "tokens_seen": 724172800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005234111972763382,
      "loss": 2.7615,
      "theoretical_loss": 3.7647741491366067,
      "tokens_seen": 727449600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225836958577644,
      "loss": 2.7764,
      "theoretical_loss": 3.7630689488399027,
      "tokens_seen": 730726400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005217561944391905,
      "loss": 2.7431,
      "theoretical_loss": 3.7613735082568764,
      "tokens_seen": 734003200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209286930206166,
      "loss": 2.7793,
      "theoretical_loss": 3.759687728330217,
      "tokens_seen": 737280000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201011916020427,
      "loss": 2.7728,
      "theoretical_loss": 3.75801151144256,
      "tokens_seen": 740556800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005192736901834689,
      "loss": 2.7565,
      "theoretical_loss": 3.756344761389295,
      "tokens_seen": 743833600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000518446188764895,
      "loss": 2.7275,
      "theoretical_loss": 3.754687383352003,
      "tokens_seen": 747110400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176186873463211,
      "loss": 2.7306,
      "theoretical_loss": 3.7530392838725097,
      "tokens_seen": 750387200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005167911859277473,
      "loss": 2.7465,
      "theoretical_loss": 3.751400370827529,
      "tokens_seen": 753664000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005159636845091734,
      "loss": 2.7684,
      "theoretical_loss": 3.749770553403895,
      "tokens_seen": 756940800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005151361830905996,
      "loss": 2.7327,
      "theoretical_loss": 3.748149742074355,
      "tokens_seen": 760217600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005143086816720257,
      "loss": 2.7618,
      "theoretical_loss": 3.746537848573908,
      "tokens_seen": 763494400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005134811802534518,
      "loss": 2.7765,
      "theoretical_loss": 3.744934785876686,
      "tokens_seen": 766771200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005126536788348779,
      "loss": 2.7845,
      "theoretical_loss": 3.7433404681733475,
      "tokens_seen": 770048000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118261774163041,
      "loss": 2.7901,
      "theoretical_loss": 3.7417548108489846,
      "tokens_seen": 773324800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109986759977303,
      "loss": 2.7636,
      "theoretical_loss": 3.740177730461517,
      "tokens_seen": 776601600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005101711745791564,
      "loss": 2.8129,
      "theoretical_loss": 3.73860914472057,
      "tokens_seen": 779878400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005093436731605825,
      "loss": 2.7786,
      "theoretical_loss": 3.7370489724668197,
      "tokens_seen": 783155200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085161717420087,
      "loss": 2.7891,
      "theoretical_loss": 3.735497133651788,
      "tokens_seen": 786432000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076886703234348,
      "loss": 2.7787,
      "theoretical_loss": 3.733953549318091,
      "tokens_seen": 789708800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506861168904861,
      "loss": 2.7923,
      "theoretical_loss": 3.7324181415801094,
      "tokens_seen": 792985600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060336674862872,
      "loss": 2.7873,
      "theoretical_loss": 3.7308908336050814,
      "tokens_seen": 796262400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052061660677133,
      "loss": 2.772,
      "theoretical_loss": 3.729371549594614,
      "tokens_seen": 799539200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005043786646491394,
      "loss": 2.7862,
      "theoretical_loss": 3.7278602147665776,
      "tokens_seen": 802816000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005035511632305655,
      "loss": 2.7841,
      "theoretical_loss": 3.726356755337407,
      "tokens_seen": 806092800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027236618119917,
      "loss": 2.7735,
      "theoretical_loss": 3.724861098504767,
      "tokens_seen": 809369600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018961603934179,
      "loss": 2.7389,
      "theoretical_loss": 3.7233731724305974,
      "tokens_seen": 812646400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501068658974844,
      "loss": 2.7629,
      "theoretical_loss": 3.7218929062245105,
      "tokens_seen": 815923200
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.009068925864994526,
      "objective/train/avg_weight": 0.9902427196502686,
      "objective/train/docs_used": 471128,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6760623455047607,
      "objective/train/std_weight": 0.06257461756467819,
      "objective/train/theoretical_loss": 3.7204202299275475,
      "objective/train/tokens_used": 839660000,
      "theoretical_loss": 3.7204202299275475,
      "tokens_seen": 819200000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005002411575562701,
      "loss": 2.7057,
      "theoretical_loss": 3.7204202299275475,
      "tokens_seen": 819200000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994136561376962,
      "loss": 2.7306,
      "theoretical_loss": 3.7189550744962707,
      "tokens_seen": 822476800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004985861547191224,
      "loss": 2.7091,
      "theoretical_loss": 3.717497371787192,
      "tokens_seen": 825753600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004977586533005485,
      "loss": 2.7075,
      "theoretical_loss": 3.7160470545415274,
      "tokens_seen": 829030400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969311518819746,
      "loss": 2.7036,
      "theoretical_loss": 3.714604056370267,
      "tokens_seen": 832307200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961036504634008,
      "loss": 2.716,
      "theoretical_loss": 3.713168311739558,
      "tokens_seen": 835584000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004952761490448269,
      "loss": 2.7011,
      "theoretical_loss": 3.7117397559563843,
      "tokens_seen": 838860800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004944486476262531,
      "loss": 2.6838,
      "theoretical_loss": 3.710318325154545,
      "tokens_seen": 842137600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936211462076792,
      "loss": 2.7304,
      "theoretical_loss": 3.7089039562809223,
      "tokens_seen": 845414400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927936447891053,
      "loss": 2.7027,
      "theoretical_loss": 3.7074965870820193,
      "tokens_seen": 848691200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004919661433705314,
      "loss": 2.7073,
      "theoretical_loss": 3.7060961560907857,
      "tokens_seen": 851968000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911386419519576,
      "loss": 2.7664,
      "theoretical_loss": 3.7047026026137,
      "tokens_seen": 855244800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903111405333838,
      "loss": 2.7282,
      "theoretical_loss": 3.7033158667181154,
      "tokens_seen": 858521600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004894836391148099,
      "loss": 2.7546,
      "theoretical_loss": 3.701935889219863,
      "tokens_seen": 861798400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488656137696236,
      "loss": 2.7233,
      "theoretical_loss": 3.7005626116710966,
      "tokens_seen": 865075200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048782863627766217,
      "loss": 2.7292,
      "theoretical_loss": 3.69919597634839,
      "tokens_seen": 868352000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004870011348590883,
      "loss": 2.7075,
      "theoretical_loss": 3.6978359262410603,
      "tokens_seen": 871628800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004861736334405145,
      "loss": 2.7358,
      "theoretical_loss": 3.6964824050397276,
      "tokens_seen": 874905600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048534613202194063,
      "loss": 2.676,
      "theoretical_loss": 3.6951353571251015,
      "tokens_seen": 878182400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048451863060336675,
      "loss": 2.6779,
      "theoretical_loss": 3.693794727556988,
      "tokens_seen": 881459200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048369112918479287,
      "loss": 2.6825,
      "theoretical_loss": 3.692460462063506,
      "tokens_seen": 884736000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000482863627766219,
      "loss": 2.7369,
      "theoretical_loss": 3.691132507030521,
      "tokens_seen": 888012800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048203612634764516,
      "loss": 2.7514,
      "theoretical_loss": 3.6898108094912816,
      "tokens_seen": 891289600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004812086249290713,
      "loss": 2.7082,
      "theoretical_loss": 3.6884953171162556,
      "tokens_seen": 894566400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048038112351049745,
      "loss": 2.7144,
      "theoretical_loss": 3.6871859782031624,
      "tokens_seen": 897843200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004795536220919236,
      "loss": 2.7252,
      "theoretical_loss": 3.685882741667202,
      "tokens_seen": 901120000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047872612067334974,
      "loss": 2.7262,
      "theoretical_loss": 3.684585557031461,
      "tokens_seen": 904396800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047789861925477586,
      "loss": 2.7381,
      "theoretical_loss": 3.6832943744175126,
      "tokens_seen": 907673600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000477071117836202,
      "loss": 2.7141,
      "theoretical_loss": 3.682009144536188,
      "tokens_seen": 910950400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004762436164176281,
      "loss": 2.7418,
      "theoretical_loss": 3.680729818678526,
      "tokens_seen": 914227200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004754161149990543,
      "loss": 2.7204,
      "theoretical_loss": 3.6794563487068936,
      "tokens_seen": 917504000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047458861358048044,
      "loss": 2.7297,
      "theoretical_loss": 3.6781886870462692,
      "tokens_seen": 920780800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047376111216190656,
      "loss": 2.7465,
      "theoretical_loss": 3.676926786675698,
      "tokens_seen": 924057600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004729336107433327,
      "loss": 2.6863,
      "theoretical_loss": 3.6756706011198963,
      "tokens_seen": 927334400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004721061093247588,
      "loss": 2.6538,
      "theoretical_loss": 3.6744200844410217,
      "tokens_seen": 930611200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047127860790618497,
      "loss": 2.6555,
      "theoretical_loss": 3.6731751912305914,
      "tokens_seen": 933888000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047045110648761114,
      "loss": 2.688,
      "theoretical_loss": 3.671935876601547,
      "tokens_seen": 937164800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046962360506903726,
      "loss": 2.6286,
      "theoretical_loss": 3.6707020961804715,
      "tokens_seen": 940441600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046879610365046343,
      "loss": 2.6828,
      "theoretical_loss": 3.6694738060999468,
      "tokens_seen": 943718400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046796860223188955,
      "loss": 2.7023,
      "theoretical_loss": 3.668250962991049,
      "tokens_seen": 946995200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046714110081331567,
      "loss": 2.7018,
      "theoretical_loss": 3.667033523975983,
      "tokens_seen": 950272000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004663135993947418,
      "loss": 2.6876,
      "theoretical_loss": 3.66582144666085,
      "tokens_seen": 953548800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004654860979761679,
      "loss": 2.687,
      "theoretical_loss": 3.664614689128546,
      "tokens_seen": 956825600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046465859655759413,
      "loss": 2.6865,
      "theoretical_loss": 3.6634132099317886,
      "tokens_seen": 960102400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046383109513902025,
      "loss": 2.6943,
      "theoretical_loss": 3.662216968086267,
      "tokens_seen": 963379200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046300359372044637,
      "loss": 2.7232,
      "theoretical_loss": 3.6610259230639217,
      "tokens_seen": 966656000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004621760923018725,
      "loss": 2.7559,
      "theoretical_loss": 3.659840034786333,
      "tokens_seen": 969932800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046134859088329866,
      "loss": 2.8156,
      "theoretical_loss": 3.6586592636182376,
      "tokens_seen": 973209600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004605210894647248,
      "loss": 2.8084,
      "theoretical_loss": 3.6574835703611566,
      "tokens_seen": 976486400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045969358804615095,
      "loss": 2.793,
      "theoretical_loss": 3.6563129162471313,
      "tokens_seen": 979763200
    },
    {
      "debugging/Self-BLEU-5": 0.4286046663919377,
      "debugging/distinct-1-grams": 0.8147567798871364,
      "debugging/distinct-2-grams": 0.9823269374342457,
      "debugging/entropy-1-grams": 6.1671920556004824,
      "debugging/entropy-2-grams": 6.947028138756313,
      "debugging/length": 477.53333333333336,
      "debugging/num_segments": 15,
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.020611366257071495,
      "objective/train/avg_weight": 0.9778223633766174,
      "objective/train/docs_used": 560408,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.926975965499878,
      "objective/train/std_weight": 0.09140268713235855,
      "objective/train/theoretical_loss": 3.6551472629325787,
      "objective/train/tokens_used": 1003500000,
      "theoretical_loss": 3.6551472629325787,
      "tokens_seen": 983040000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045886608662757707,
      "loss": 2.7908,
      "theoretical_loss": 3.6551472629325787,
      "tokens_seen": 983040000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045803858520900324,
      "loss": 2.7934,
      "theoretical_loss": 3.653986572492247,
      "tokens_seen": 986316800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045721108379042936,
      "loss": 2.7583,
      "theoretical_loss": 3.65283080741328,
      "tokens_seen": 989593600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045640013240022694,
      "loss": 2.7744,
      "theoretical_loss": 3.6516799305893866,
      "tokens_seen": 992870400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045557263098165306,
      "loss": 2.7683,
      "theoretical_loss": 3.6505339053151076,
      "tokens_seen": 996147200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004547451295630793,
      "loss": 2.7755,
      "theoretical_loss": 3.649392695280186,
      "tokens_seen": 999424000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004539176281445054,
      "loss": 2.7407,
      "theoretical_loss": 3.6482562645640337,
      "tokens_seen": 1002700800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004530901267259315,
      "loss": 2.7456,
      "theoretical_loss": 3.6471245776302883,
      "tokens_seen": 1005977600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045226262530735764,
      "loss": 2.7827,
      "theoretical_loss": 3.6459975993214724,
      "tokens_seen": 1009254400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004514351238887838,
      "loss": 2.7414,
      "theoretical_loss": 3.6448752948537377,
      "tokens_seen": 1012531200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045060762247020993,
      "loss": 2.7774,
      "theoretical_loss": 3.6437576298116996,
      "tokens_seen": 1015808000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044978012105163605,
      "loss": 2.7705,
      "theoretical_loss": 3.6426445701433607,
      "tokens_seen": 1019084800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004489526196330623,
      "loss": 2.79,
      "theoretical_loss": 3.6415360821551226,
      "tokens_seen": 1022361600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004481251182144884,
      "loss": 2.7458,
      "theoretical_loss": 3.6404321325068754,
      "tokens_seen": 1025638400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004472976167959145,
      "loss": 2.7668,
      "theoretical_loss": 3.639332688207178,
      "tokens_seen": 1028915200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044647011537734063,
      "loss": 2.7474,
      "theoretical_loss": 3.6382377166085096,
      "tokens_seen": 1032192000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044564261395876675,
      "loss": 2.7098,
      "theoretical_loss": 3.6371471854026147,
      "tokens_seen": 1035468800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044481511254019287,
      "loss": 2.7161,
      "theoretical_loss": 3.6360610626159087,
      "tokens_seen": 1038745600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004439876111216191,
      "loss": 2.6831,
      "theoretical_loss": 3.634979316604973,
      "tokens_seen": 1042022400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004431601097030452,
      "loss": 2.6983,
      "theoretical_loss": 3.6339019160521198,
      "tokens_seen": 1045299200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044233260828447133,
      "loss": 2.6901,
      "theoretical_loss": 3.632828829961029,
      "tokens_seen": 1048576000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044150510686589745,
      "loss": 2.7065,
      "theoretical_loss": 3.631760027652461,
      "tokens_seen": 1051852800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004406776054473236,
      "loss": 2.7373,
      "theoretical_loss": 3.630695478760034,
      "tokens_seen": 1055129600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043985010402874974,
      "loss": 2.709,
      "theoretical_loss": 3.6296351532260767,
      "tokens_seen": 1058406400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004390226026101759,
      "loss": 2.7033,
      "theoretical_loss": 3.6285790212975435,
      "tokens_seen": 1061683200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004381951011916021,
      "loss": 2.6803,
      "theoretical_loss": 3.6275270535220008,
      "tokens_seen": 1064960000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004373675997730282,
      "loss": 2.7044,
      "theoretical_loss": 3.626479220743673,
      "tokens_seen": 1068236800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004365400983544543,
      "loss": 2.7078,
      "theoretical_loss": 3.6254354940995586,
      "tokens_seen": 1071513600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043571259693588044,
      "loss": 2.7014,
      "theoretical_loss": 3.624395845015602,
      "tokens_seen": 1074790400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043488509551730656,
      "loss": 2.6599,
      "theoretical_loss": 3.6233602452029348,
      "tokens_seen": 1078067200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004340575940987327,
      "loss": 2.6797,
      "theoretical_loss": 3.6223286666541683,
      "tokens_seen": 1081344000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004332300926801589,
      "loss": 2.7048,
      "theoretical_loss": 3.621301081639753,
      "tokens_seen": 1084620800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432402591261585,
      "loss": 2.7439,
      "theoretical_loss": 3.6202774627043923,
      "tokens_seen": 1087897600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043157508984301114,
      "loss": 2.6985,
      "theoretical_loss": 3.619257782663513,
      "tokens_seen": 1091174400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004307475884244373,
      "loss": 2.7095,
      "theoretical_loss": 3.618242014599793,
      "tokens_seen": 1094451200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042992008700586343,
      "loss": 2.6983,
      "theoretical_loss": 3.617230131859743,
      "tokens_seen": 1097728000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042909258558728955,
      "loss": 2.6823,
      "theoretical_loss": 3.6162221080503416,
      "tokens_seen": 1101004800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004282650841687157,
      "loss": 2.6665,
      "theoretical_loss": 3.615217917035726,
      "tokens_seen": 1104281600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004274375827501419,
      "loss": 2.6981,
      "theoretical_loss": 3.614217532933929,
      "tokens_seen": 1107558400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000426610081331568,
      "loss": 2.6861,
      "theoretical_loss": 3.6132209301136715,
      "tokens_seen": 1110835200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042578257991299413,
      "loss": 2.6717,
      "theoretical_loss": 3.612228083191205,
      "tokens_seen": 1114112000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042495507849442025,
      "loss": 2.7291,
      "theoretical_loss": 3.611238967027199,
      "tokens_seen": 1117388800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042412757707584636,
      "loss": 2.7692,
      "theoretical_loss": 3.610253556723679,
      "tokens_seen": 1120665600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004233000756572726,
      "loss": 2.7318,
      "theoretical_loss": 3.609271827621014,
      "tokens_seen": 1123942400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004224725742386987,
      "loss": 2.7372,
      "theoretical_loss": 3.6082937552949463,
      "tokens_seen": 1127219200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042164507282012483,
      "loss": 2.7052,
      "theoretical_loss": 3.607319315553669,
      "tokens_seen": 1130496000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042081757140155095,
      "loss": 2.668,
      "theoretical_loss": 3.6063484844349456,
      "tokens_seen": 1133772800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004199900699829771,
      "loss": 2.6911,
      "theoretical_loss": 3.605381238203279,
      "tokens_seen": 1137049600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041916256856440324,
      "loss": 2.7014,
      "theoretical_loss": 3.604417553347117,
      "tokens_seen": 1140326400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004183350671458294,
      "loss": 2.6926,
      "theoretical_loss": 3.603457406576106,
      "tokens_seen": 1143603200
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.022877871990203857,
      "objective/train/avg_weight": 0.97539883852005,
      "objective/train/docs_used": 649861,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6508138179779053,
      "objective/train/std_weight": 0.10265526175498962,
      "objective/train/theoretical_loss": 3.602500774818379,
      "objective/train/tokens_used": 1167340000,
      "theoretical_loss": 3.602500774818379,
      "tokens_seen": 1146880000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041750756572725553,
      "loss": 2.6792,
      "theoretical_loss": 3.602500774818379,
      "tokens_seen": 1146880000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004166800643086817,
      "loss": 2.7197,
      "theoretical_loss": 3.601547635217892,
      "tokens_seen": 1150156800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004158525628901078,
      "loss": 2.7201,
      "theoretical_loss": 3.6005979651317976,
      "tokens_seen": 1153433600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041502506147153394,
      "loss": 2.7286,
      "theoretical_loss": 3.599651742127855,
      "tokens_seen": 1156710400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041419756005296005,
      "loss": 2.681,
      "theoretical_loss": 3.5987089439818805,
      "tokens_seen": 1159987200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041337005863438617,
      "loss": 2.6621,
      "theoretical_loss": 3.5977695486752426,
      "tokens_seen": 1163264000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004125425572158124,
      "loss": 2.6926,
      "theoretical_loss": 3.596833534392379,
      "tokens_seen": 1166540800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004117150557972385,
      "loss": 2.7096,
      "theoretical_loss": 3.595900879518368,
      "tokens_seen": 1169817600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041088755437866464,
      "loss": 2.7241,
      "theoretical_loss": 3.594971562636521,
      "tokens_seen": 1173094400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041006005296009075,
      "loss": 2.6816,
      "theoretical_loss": 3.5940455625260226,
      "tokens_seen": 1176371200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004092325515415169,
      "loss": 2.6928,
      "theoretical_loss": 3.5931228581595938,
      "tokens_seen": 1179648000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040840505012294305,
      "loss": 2.6899,
      "theoretical_loss": 3.5922034287011995,
      "tokens_seen": 1182924800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004075775487043692,
      "loss": 2.6567,
      "theoretical_loss": 3.5912872535037828,
      "tokens_seen": 1186201600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004067500472857954,
      "loss": 2.739,
      "theoretical_loss": 3.590374312107035,
      "tokens_seen": 1189478400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004059225458672215,
      "loss": 2.724,
      "theoretical_loss": 3.5894645842351993,
      "tokens_seen": 1192755200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004050950444486476,
      "loss": 2.7097,
      "theoretical_loss": 3.588558049794902,
      "tokens_seen": 1196032000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040426754303007374,
      "loss": 2.689,
      "theoretical_loss": 3.5876546888730187,
      "tokens_seen": 1199308800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040344004161149986,
      "loss": 2.7083,
      "theoretical_loss": 3.5867544817345713,
      "tokens_seen": 1202585600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004026125401929261,
      "loss": 2.7336,
      "theoretical_loss": 3.585857408820652,
      "tokens_seen": 1205862400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040180158880272367,
      "loss": 2.7425,
      "theoretical_loss": 3.58496345074638,
      "tokens_seen": 1209139200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004009740873841498,
      "loss": 2.7536,
      "theoretical_loss": 3.5840725882988873,
      "tokens_seen": 1212416000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040014658596557596,
      "loss": 2.7189,
      "theoretical_loss": 3.5831848024353317,
      "tokens_seen": 1215692800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003993190845470021,
      "loss": 2.7442,
      "theoretical_loss": 3.5823000742809374,
      "tokens_seen": 1218969600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003984915831284282,
      "loss": 2.7158,
      "theoretical_loss": 3.5814183851270673,
      "tokens_seen": 1222246400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003976640817098543,
      "loss": 2.7064,
      "theoretical_loss": 3.5805397164293167,
      "tokens_seen": 1225523200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039683658029128054,
      "loss": 2.6876,
      "theoretical_loss": 3.5796640498056407,
      "tokens_seen": 1228800000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039600907887270666,
      "loss": 2.696,
      "theoretical_loss": 3.5787913670345013,
      "tokens_seen": 1232076800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039519812748250424,
      "loss": 2.6741,
      "theoretical_loss": 3.577921650053045,
      "tokens_seen": 1235353600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039437062606393036,
      "loss": 2.6706,
      "theoretical_loss": 3.577054880955303,
      "tokens_seen": 1238630400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039354312464535653,
      "loss": 2.657,
      "theoretical_loss": 3.5761910419904193,
      "tokens_seen": 1241907200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039271562322678265,
      "loss": 2.7193,
      "theoretical_loss": 3.5753301155609014,
      "tokens_seen": 1245184000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003918881218082088,
      "loss": 2.6606,
      "theoretical_loss": 3.574472084220896,
      "tokens_seen": 1248460800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039106062038963494,
      "loss": 2.6497,
      "theoretical_loss": 3.5736169306744885,
      "tokens_seen": 1251737600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003902331189710611,
      "loss": 2.6164,
      "theoretical_loss": 3.572764637774024,
      "tokens_seen": 1255014400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038940561755248723,
      "loss": 2.6288,
      "theoretical_loss": 3.571915188518457,
      "tokens_seen": 1258291200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038857811613391335,
      "loss": 2.6475,
      "theoretical_loss": 3.571068566051716,
      "tokens_seen": 1261568000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038775061471533947,
      "loss": 2.6866,
      "theoretical_loss": 3.5702247536610976,
      "tokens_seen": 1264844800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003869231132967657,
      "loss": 2.667,
      "theoretical_loss": 3.5693837347756783,
      "tokens_seen": 1268121600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003860956118781918,
      "loss": 2.6401,
      "theoretical_loss": 3.5685454929647475,
      "tokens_seen": 1271398400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038526811045961793,
      "loss": 2.6509,
      "theoretical_loss": 3.5677100119362675,
      "tokens_seen": 1274675200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038444060904104405,
      "loss": 2.6335,
      "theoretical_loss": 3.566877275535345,
      "tokens_seen": 1277952000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038361310762247017,
      "loss": 2.6285,
      "theoretical_loss": 3.566047267742733,
      "tokens_seen": 1281228800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038278560620389634,
      "loss": 2.6087,
      "theoretical_loss": 3.5652199726733453,
      "tokens_seen": 1284505600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003819581047853225,
      "loss": 2.6032,
      "theoretical_loss": 3.564395374574796,
      "tokens_seen": 1287782400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038113060336674863,
      "loss": 2.6624,
      "theoretical_loss": 3.5635734578259557,
      "tokens_seen": 1291059200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003803031019481748,
      "loss": 2.6763,
      "theoretical_loss": 3.5627542069355282,
      "tokens_seen": 1294336000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003794756005296009,
      "loss": 2.6398,
      "theoretical_loss": 3.5619376065406474,
      "tokens_seen": 1297612800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037864809911102704,
      "loss": 2.6559,
      "theoretical_loss": 3.5611236414054868,
      "tokens_seen": 1300889600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037782059769245316,
      "loss": 2.6598,
      "theoretical_loss": 3.560312296419899,
      "tokens_seen": 1304166400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003769930962738793,
      "loss": 2.6584,
      "theoretical_loss": 3.55950355659806,
      "tokens_seen": 1307443200
    },
    {
      "debugging/Self-BLEU-5": 0.49020908264157476,
      "debugging/distinct-1-grams": 0.768901113497886,
      "debugging/distinct-2-grams": 0.9428782333551957,
      "debugging/entropy-1-grams": 6.085999550681761,
      "debugging/entropy-2-grams": 7.0033060167714964,
      "debugging/length": 490.2352941176471,
      "debugging/num_segments": 17,
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.02056093141436577,
      "objective/train/avg_weight": 0.977873682975769,
      "objective/train/docs_used": 741674,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.746752977371216,
      "objective/train/std_weight": 0.11813095211982727,
      "objective/train/theoretical_loss": 3.558697407077142,
      "objective/train/tokens_used": 1331180000,
      "theoretical_loss": 3.558697407077142,
      "tokens_seen": 1310720000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003761655948553055,
      "loss": 2.6483,
      "theoretical_loss": 3.558697407077142,
      "tokens_seen": 1310720000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003753380934367316,
      "loss": 2.6646,
      "theoretical_loss": 3.5578938331159975,
      "tokens_seen": 1313996800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037451059201815774,
      "loss": 2.6382,
      "theoretical_loss": 3.557092820093863,
      "tokens_seen": 1317273600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037368309059958386,
      "loss": 2.6577,
      "theoretical_loss": 3.556294353509079,
      "tokens_seen": 1320550400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037285558918101,
      "loss": 2.6493,
      "theoretical_loss": 3.555498418977828,
      "tokens_seen": 1323827200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037202808776243615,
      "loss": 2.6434,
      "theoretical_loss": 3.5547050022328874,
      "tokens_seen": 1327104000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712005863438623,
      "loss": 2.639,
      "theoretical_loss": 3.553914089122399,
      "tokens_seen": 1330380800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037037308492528844,
      "loss": 2.6797,
      "theoretical_loss": 3.553125665608655,
      "tokens_seen": 1333657600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003695455835067146,
      "loss": 2.6617,
      "theoretical_loss": 3.5523397177669005,
      "tokens_seen": 1336934400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036871808208814073,
      "loss": 2.6747,
      "theoretical_loss": 3.551556231784149,
      "tokens_seen": 1340211200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036789058066956685,
      "loss": 2.6276,
      "theoretical_loss": 3.5507751939580148,
      "tokens_seen": 1343488000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036706307925099297,
      "loss": 2.6325,
      "theoretical_loss": 3.5499965906955606,
      "tokens_seen": 1346764800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003662355778324192,
      "loss": 2.6789,
      "theoretical_loss": 3.549220408512161,
      "tokens_seen": 1350041600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003654080764138453,
      "loss": 2.6544,
      "theoretical_loss": 3.5484466340303755,
      "tokens_seen": 1353318400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036458057499527143,
      "loss": 2.6411,
      "theoretical_loss": 3.547675253978843,
      "tokens_seen": 1356595200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036375307357669755,
      "loss": 2.6751,
      "theoretical_loss": 3.5469062551911854,
      "tokens_seen": 1359872000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036292557215812367,
      "loss": 2.6842,
      "theoretical_loss": 3.5461396246049244,
      "tokens_seen": 1363148800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003620980707395498,
      "loss": 2.6504,
      "theoretical_loss": 3.545375349260419,
      "tokens_seen": 1366425600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036127056932097596,
      "loss": 2.6894,
      "theoretical_loss": 3.544613416299808,
      "tokens_seen": 1369702400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036044306790240213,
      "loss": 2.6629,
      "theoretical_loss": 3.5438538129659687,
      "tokens_seen": 1372979200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035961556648382825,
      "loss": 2.6823,
      "theoretical_loss": 3.5430965266014933,
      "tokens_seen": 1376256000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003587880650652544,
      "loss": 2.6879,
      "theoretical_loss": 3.5423415446476705,
      "tokens_seen": 1379532800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035796056364668054,
      "loss": 2.6256,
      "theoretical_loss": 3.541588854643487,
      "tokens_seen": 1382809600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035713306222810666,
      "loss": 2.6521,
      "theoretical_loss": 3.5408384442246343,
      "tokens_seen": 1386086400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035632211083790424,
      "loss": 2.6065,
      "theoretical_loss": 3.540090301122535,
      "tokens_seen": 1389363200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035549460941933046,
      "loss": 2.6299,
      "theoretical_loss": 3.5393444131633762,
      "tokens_seen": 1392640000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003546671080007566,
      "loss": 2.676,
      "theoretical_loss": 3.5386007682671576,
      "tokens_seen": 1395916800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003538396065821827,
      "loss": 2.6589,
      "theoretical_loss": 3.5378593544467494,
      "tokens_seen": 1399193600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003530121051636088,
      "loss": 2.641,
      "theoretical_loss": 3.537120159806962,
      "tokens_seen": 1402470400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000352184603745035,
      "loss": 2.6321,
      "theoretical_loss": 3.5363831725436317,
      "tokens_seen": 1405747200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003513571023264611,
      "loss": 2.6062,
      "theoretical_loss": 3.5356483809427104,
      "tokens_seen": 1409024000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003505296009078873,
      "loss": 2.603,
      "theoretical_loss": 3.53491577337937,
      "tokens_seen": 1412300800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003497020994893134,
      "loss": 2.6142,
      "theoretical_loss": 3.5341853383171205,
      "tokens_seen": 1415577600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003488745980707396,
      "loss": 2.6449,
      "theoretical_loss": 3.5334570643069325,
      "tokens_seen": 1418854400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003480470966521657,
      "loss": 2.628,
      "theoretical_loss": 3.5327309399863767,
      "tokens_seen": 1422131200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003472195952335918,
      "loss": 2.6463,
      "theoretical_loss": 3.5320069540787706,
      "tokens_seen": 1425408000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000346392093815018,
      "loss": 2.6097,
      "theoretical_loss": 3.531285095392335,
      "tokens_seen": 1428684800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003455645923964441,
      "loss": 2.6261,
      "theoretical_loss": 3.5305653528193623,
      "tokens_seen": 1431961600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003447370909778702,
      "loss": 2.6001,
      "theoretical_loss": 3.5298477153353955,
      "tokens_seen": 1435238400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003439095895592964,
      "loss": 2.5685,
      "theoretical_loss": 3.5291321719984134,
      "tokens_seen": 1438515200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003430820881407225,
      "loss": 2.5861,
      "theoretical_loss": 3.5284187119480315,
      "tokens_seen": 1441792000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034225458672214863,
      "loss": 2.5687,
      "theoretical_loss": 3.527707324404704,
      "tokens_seen": 1445068800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003414270853035748,
      "loss": 2.5948,
      "theoretical_loss": 3.5269979986689464,
      "tokens_seen": 1448345600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000340599583885001,
      "loss": 2.605,
      "theoretical_loss": 3.5262907241205568,
      "tokens_seen": 1451622400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003397720824664271,
      "loss": 2.5807,
      "theoretical_loss": 3.5255854902178534,
      "tokens_seen": 1454899200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033894458104785326,
      "loss": 2.5755,
      "theoretical_loss": 3.524882286496917,
      "tokens_seen": 1458176000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003381170796292794,
      "loss": 2.5977,
      "theoretical_loss": 3.524181102570846,
      "tokens_seen": 1461452800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003372895782107055,
      "loss": 2.596,
      "theoretical_loss": 3.5234819281290166,
      "tokens_seen": 1464729600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003364620767921316,
      "loss": 2.6195,
      "theoretical_loss": 3.5227847529363547,
      "tokens_seen": 1468006400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003356345753735578,
      "loss": 2.6149,
      "theoretical_loss": 3.522089566832615,
      "tokens_seen": 1471283200
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.03047107346355915,
      "objective/train/avg_weight": 0.9672204852104187,
      "objective/train/docs_used": 835915,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.794424533843994,
      "objective/train/std_weight": 0.12225707620382309,
      "objective/train/theoretical_loss": 3.521396359731668,
      "objective/train/tokens_used": 1495020000,
      "theoretical_loss": 3.521396359731668,
      "tokens_seen": 1474560000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003348070739549839,
      "loss": 2.6119,
      "theoretical_loss": 3.521396359731668,
      "tokens_seen": 1474560000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033397957253641003,
      "loss": 2.6421,
      "theoretical_loss": 3.520705121620797,
      "tokens_seen": 1477836800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003331520711178362,
      "loss": 2.611,
      "theoretical_loss": 3.520015842560001,
      "tokens_seen": 1481113600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003323245696992623,
      "loss": 2.6349,
      "theoretical_loss": 3.5193285126813105,
      "tokens_seen": 1484390400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003314970682806885,
      "loss": 2.6584,
      "theoretical_loss": 3.5186431221881023,
      "tokens_seen": 1487667200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003306695668621146,
      "loss": 2.6635,
      "theoretical_loss": 3.517959661354433,
      "tokens_seen": 1490944000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003298420654435408,
      "loss": 2.6585,
      "theoretical_loss": 3.517278120524372,
      "tokens_seen": 1494220800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003290145640249669,
      "loss": 2.6678,
      "theoretical_loss": 3.516598490111348,
      "tokens_seen": 1497497600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032818706260639307,
      "loss": 2.6823,
      "theoretical_loss": 3.515920760597498,
      "tokens_seen": 1500774400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003273595611878192,
      "loss": 2.6309,
      "theoretical_loss": 3.515244922533025,
      "tokens_seen": 1504051200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003265320597692453,
      "loss": 2.6216,
      "theoretical_loss": 3.514570966535569,
      "tokens_seen": 1507328000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003257045583506715,
      "loss": 2.6695,
      "theoretical_loss": 3.513898883289575,
      "tokens_seen": 1510604800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003248770569320976,
      "loss": 2.6468,
      "theoretical_loss": 3.513228663545678,
      "tokens_seen": 1513881600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003240495555135237,
      "loss": 2.6609,
      "theoretical_loss": 3.5125602981200874,
      "tokens_seen": 1517158400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003232220540949499,
      "loss": 2.6232,
      "theoretical_loss": 3.5118937778939836,
      "tokens_seen": 1520435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000322394552676376,
      "loss": 2.6252,
      "theoretical_loss": 3.5112290938129176,
      "tokens_seen": 1523712000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003215670512578021,
      "loss": 2.6457,
      "theoretical_loss": 3.5105662368862234,
      "tokens_seen": 1526988800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003207395498392283,
      "loss": 2.6211,
      "theoretical_loss": 3.5099051981864275,
      "tokens_seen": 1530265600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003199120484206544,
      "loss": 2.6021,
      "theoretical_loss": 3.509245968848674,
      "tokens_seen": 1533542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003190845470020806,
      "loss": 2.601,
      "theoretical_loss": 3.5085885400701535,
      "tokens_seen": 1536819200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003182570455835067,
      "loss": 2.5826,
      "theoretical_loss": 3.507932903109534,
      "tokens_seen": 1540096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003174295441649329,
      "loss": 2.5916,
      "theoretical_loss": 3.5072790492864057,
      "tokens_seen": 1543372800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000316602042746359,
      "loss": 2.6108,
      "theoretical_loss": 3.506626969980724,
      "tokens_seen": 1546649600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003157745413277851,
      "loss": 2.6397,
      "theoretical_loss": 3.505976656632269,
      "tokens_seen": 1549926400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003149470399092113,
      "loss": 2.6174,
      "theoretical_loss": 3.505328100740097,
      "tokens_seen": 1553203200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003141195384906374,
      "loss": 2.6173,
      "theoretical_loss": 3.5046812938620135,
      "tokens_seen": 1556480000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003132920370720635,
      "loss": 2.6152,
      "theoretical_loss": 3.5040362276140393,
      "tokens_seen": 1559756800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003124645356534897,
      "loss": 2.5919,
      "theoretical_loss": 3.5033928936698904,
      "tokens_seen": 1563033600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003116370342349158,
      "loss": 2.6195,
      "theoretical_loss": 3.50275128376046,
      "tokens_seen": 1566310400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031080953281634194,
      "loss": 2.5736,
      "theoretical_loss": 3.5021113896733076,
      "tokens_seen": 1569587200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003099820313977681,
      "loss": 2.5816,
      "theoretical_loss": 3.501473203252153,
      "tokens_seen": 1572864000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003091545299791942,
      "loss": 2.639,
      "theoretical_loss": 3.500836716396377,
      "tokens_seen": 1576140800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003083270285606204,
      "loss": 2.6309,
      "theoretical_loss": 3.5002019210605253,
      "tokens_seen": 1579417600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003074995271420465,
      "loss": 2.6344,
      "theoretical_loss": 3.4995688092538213,
      "tokens_seen": 1582694400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003066720257234727,
      "loss": 2.5819,
      "theoretical_loss": 3.49893737303968,
      "tokens_seen": 1585971200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003058445243048988,
      "loss": 2.5641,
      "theoretical_loss": 3.4983076045352317,
      "tokens_seen": 1589248000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003050170228863249,
      "loss": 2.5544,
      "theoretical_loss": 3.4976794959108464,
      "tokens_seen": 1592524800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003041895214677511,
      "loss": 2.6117,
      "theoretical_loss": 3.4970530393896673,
      "tokens_seen": 1595801600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003033620200491772,
      "loss": 2.5828,
      "theoretical_loss": 3.4964282272471467,
      "tokens_seen": 1599078400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030253451863060333,
      "loss": 2.5688,
      "theoretical_loss": 3.495805051810588,
      "tokens_seen": 1602355200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003017070172120295,
      "loss": 2.5939,
      "theoretical_loss": 3.4951835054586926,
      "tokens_seen": 1605632000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003008795157934556,
      "loss": 2.5762,
      "theoretical_loss": 3.4945635806211115,
      "tokens_seen": 1608908800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030005201437488174,
      "loss": 2.6168,
      "theoretical_loss": 3.493945269778002,
      "tokens_seen": 1612185600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002992245129563079,
      "loss": 2.6195,
      "theoretical_loss": 3.4933285654595867,
      "tokens_seen": 1615462400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029839701153773403,
      "loss": 2.6088,
      "theoretical_loss": 3.4927134602457244,
      "tokens_seen": 1618739200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002975695101191602,
      "loss": 2.6144,
      "theoretical_loss": 3.4920999467654754,
      "tokens_seen": 1622016000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002967420087005864,
      "loss": 2.6003,
      "theoretical_loss": 3.49148801769668,
      "tokens_seen": 1625292800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002959145072820125,
      "loss": 2.63,
      "theoretical_loss": 3.4908776657655354,
      "tokens_seen": 1628569600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002950870058634386,
      "loss": 2.5355,
      "theoretical_loss": 3.4902688837461824,
      "tokens_seen": 1631846400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002942595044448648,
      "loss": 2.556,
      "theoretical_loss": 3.489661664460292,
      "tokens_seen": 1635123200
    },
    {
      "debugging/Self-BLEU-5": 0.4862537472725578,
      "debugging/distinct-1-grams": 0.7771784690107355,
      "debugging/distinct-2-grams": 0.9668669456652452,
      "debugging/entropy-1-grams": 6.202714981817353,
      "debugging/entropy-2-grams": 7.211883548596637,
      "debugging/length": 471.0,
      "debugging/num_segments": 21,
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.013972360640764236,
      "objective/train/avg_weight": 0.9849681258201599,
      "objective/train/docs_used": 927402,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8893728256225586,
      "objective/train/std_weight": 0.06725912541151047,
      "objective/train/theoretical_loss": 3.4890560007766607,
      "objective/train/tokens_used": 1658860000,
      "theoretical_loss": 3.4890560007766607,
      "tokens_seen": 1638400000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002934320030262909,
      "loss": 2.5529,
      "theoretical_loss": 3.4890560007766607,
      "tokens_seen": 1638400000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000292604501607717,
      "loss": 2.5676,
      "theoretical_loss": 3.4884518856108038,
      "tokens_seen": 1641676800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002917770001891432,
      "loss": 2.5749,
      "theoretical_loss": 3.4878493119245606,
      "tokens_seen": 1644953600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002909494987705693,
      "loss": 2.5696,
      "theoretical_loss": 3.4872482727256964,
      "tokens_seen": 1648230400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029012199735199543,
      "loss": 2.5879,
      "theoretical_loss": 3.4866487610675163,
      "tokens_seen": 1651507200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002892944959334216,
      "loss": 2.5732,
      "theoretical_loss": 3.4860507700484757,
      "tokens_seen": 1654784000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002884669945148477,
      "loss": 2.5287,
      "theoretical_loss": 3.4854542928117986,
      "tokens_seen": 1658060800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002876394930962739,
      "loss": 2.5417,
      "theoretical_loss": 3.484859322545101,
      "tokens_seen": 1661337600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002868119916777,
      "loss": 2.5627,
      "theoretical_loss": 3.484265852480015,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002859844902591262,
      "loss": 2.5806,
      "theoretical_loss": 3.48367387589182,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002851569888405523,
      "loss": 2.5934,
      "theoretical_loss": 3.4830833860990733,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002843294874219784,
      "loss": 2.595,
      "theoretical_loss": 3.4824943764632503,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002835019860034046,
      "loss": 2.6038,
      "theoretical_loss": 3.4819068403883855,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002826744845848307,
      "loss": 2.5881,
      "theoretical_loss": 3.481320771320714,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028184698316625683,
      "loss": 2.5792,
      "theoretical_loss": 3.4807361627483235,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000281019481747683,
      "loss": 2.6324,
      "theoretical_loss": 3.480153008200804,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002801919803291091,
      "loss": 2.5892,
      "theoretical_loss": 3.4795713012489045,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027936447891053524,
      "loss": 2.5599,
      "theoretical_loss": 3.478991035504191,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002785369774919614,
      "loss": 2.6314,
      "theoretical_loss": 3.478412204618709,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027770947607338753,
      "loss": 2.5939,
      "theoretical_loss": 3.4778348022846517,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002768819746548137,
      "loss": 2.6271,
      "theoretical_loss": 3.477258822234026,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002760544732362398,
      "loss": 2.5764,
      "theoretical_loss": 3.4766842582383277,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000275226971817666,
      "loss": 2.5701,
      "theoretical_loss": 3.476111104108216,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002743994703990921,
      "loss": 2.5806,
      "theoretical_loss": 3.4755393536931942,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002735719689805183,
      "loss": 2.5904,
      "theoretical_loss": 3.4749690008812903,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002727444675619444,
      "loss": 2.5876,
      "theoretical_loss": 3.474400039598745,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002719169661433705,
      "loss": 2.6179,
      "theoretical_loss": 3.473832463809698,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027108946472479664,
      "loss": 2.6163,
      "theoretical_loss": 3.473266267515883,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002702619633062228,
      "loss": 2.6371,
      "theoretical_loss": 3.4727014447563196,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026943446188764893,
      "loss": 2.6279,
      "theoretical_loss": 3.4721379896070133,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026860696046907505,
      "loss": 2.6422,
      "theoretical_loss": 3.4715758961806573,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002677794590505012,
      "loss": 2.6189,
      "theoretical_loss": 3.471015158626335,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026695195763192734,
      "loss": 2.6253,
      "theoretical_loss": 3.4704557711292265,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002661244562133535,
      "loss": 2.6349,
      "theoretical_loss": 3.469897727910321,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026529695479477963,
      "loss": 2.6548,
      "theoretical_loss": 3.4693410232261277,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002644694533762058,
      "loss": 2.6662,
      "theoretical_loss": 3.4687856513683917,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002636419519576319,
      "loss": 2.6461,
      "theoretical_loss": 3.468231606663813,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002628144505390581,
      "loss": 2.6367,
      "theoretical_loss": 3.4676788834737677,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002619869491204842,
      "loss": 2.6133,
      "theoretical_loss": 3.4671274761940314,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026115944770191033,
      "loss": 2.5947,
      "theoretical_loss": 3.466577379254506,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002603319462833365,
      "loss": 2.5994,
      "theoretical_loss": 3.4660285871189496,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002595044448647626,
      "loss": 2.6165,
      "theoretical_loss": 3.465481094284707,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025867694344618874,
      "loss": 2.6054,
      "theoretical_loss": 3.4649348952824477,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002578494420276149,
      "loss": 2.587,
      "theoretical_loss": 3.4643899846758983,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025702194060904103,
      "loss": 2.5389,
      "theoretical_loss": 3.4638463570615854,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025621098921883867,
      "loss": 2.5921,
      "theoretical_loss": 3.4633040070685777,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025538348780026484,
      "loss": 2.5911,
      "theoretical_loss": 3.46276292935823,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025455598638169096,
      "loss": 2.5718,
      "theoretical_loss": 3.4622231186239296,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025374503499148854,
      "loss": 2.5735,
      "theoretical_loss": 3.461684569590848,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002529175335729147,
      "loss": 2.5525,
      "theoretical_loss": 3.4611472770156917,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.013695425353944302,
      "objective/train/avg_weight": 0.9852670431137085,
      "objective/train/docs_used": 1018281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.786952018737793,
      "objective/train/std_weight": 0.06961064785718918,
      "objective/train/theoretical_loss": 3.460611235686456,
      "objective/train/tokens_used": 1822700000,
      "theoretical_loss": 3.460611235686456,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025209003215434083,
      "loss": 2.5599,
      "theoretical_loss": 3.460611235686456,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025126253073576695,
      "loss": 2.5413,
      "theoretical_loss": 3.4600764404221835,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002504350293171931,
      "loss": 2.5337,
      "theoretical_loss": 3.4595428860727218,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024960752789861924,
      "loss": 2.5177,
      "theoretical_loss": 3.4590105675184866,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002487965765084169,
      "loss": 2.5605,
      "theoretical_loss": 3.4584794796702223,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000247969075089843,
      "loss": 2.5608,
      "theoretical_loss": 3.457949617468771,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024714157367126916,
      "loss": 2.6132,
      "theoretical_loss": 3.4574209758848395,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024633062228106674,
      "loss": 2.5523,
      "theoretical_loss": 3.4568935499187696,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024550312086249286,
      "loss": 2.564,
      "theoretical_loss": 3.456367334600311,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024467561944391904,
      "loss": 2.5575,
      "theoretical_loss": 3.455842324988396,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024384811802534518,
      "loss": 2.566,
      "theoretical_loss": 3.455318516170914,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002430206166067713,
      "loss": 2.5524,
      "theoretical_loss": 3.454795903264496,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024219311518819747,
      "loss": 2.572,
      "theoretical_loss": 3.4542744814142896,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002413656137696236,
      "loss": 2.593,
      "theoretical_loss": 3.4537542457937453,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024053811235104973,
      "loss": 2.6142,
      "theoretical_loss": 3.4532351916044,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002397106109324759,
      "loss": 2.6223,
      "theoretical_loss": 3.4527173140756666,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023888310951390203,
      "loss": 2.6359,
      "theoretical_loss": 3.4522006084646204,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023805560809532814,
      "loss": 2.6417,
      "theoretical_loss": 3.4516850700557917,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023722810667675432,
      "loss": 2.6799,
      "theoretical_loss": 3.4511706941609566,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023640060525818043,
      "loss": 2.6135,
      "theoretical_loss": 3.450657476118936,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023557310383960658,
      "loss": 2.5851,
      "theoretical_loss": 3.450145411295389,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023474560242103273,
      "loss": 2.5753,
      "theoretical_loss": 3.4496344950826114,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023391810100245887,
      "loss": 2.5719,
      "theoretical_loss": 3.449124722899339,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000233090599583885,
      "loss": 2.5541,
      "theoretical_loss": 3.4486160901905487,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002322630981653111,
      "loss": 2.5474,
      "theoretical_loss": 3.448108592427261,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023143559674673728,
      "loss": 2.5327,
      "theoretical_loss": 3.4476022251063485,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023060809532816343,
      "loss": 2.5256,
      "theoretical_loss": 3.447096983750342,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022978059390958954,
      "loss": 2.5316,
      "theoretical_loss": 3.4465928639072416,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022895309249101572,
      "loss": 2.5481,
      "theoretical_loss": 3.446089861150326,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022812559107244183,
      "loss": 2.578,
      "theoretical_loss": 3.445587971077968,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022729808965386795,
      "loss": 2.5376,
      "theoretical_loss": 3.4450871893134467,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022647058823529412,
      "loss": 2.5419,
      "theoretical_loss": 3.4445875115047655,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022564308681672024,
      "loss": 2.5076,
      "theoretical_loss": 3.4440889333244695,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002248155853981464,
      "loss": 2.4827,
      "theoretical_loss": 3.443591450469466,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022398808397957256,
      "loss": 2.5315,
      "theoretical_loss": 3.4430950586608438,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022316058256099868,
      "loss": 2.5303,
      "theoretical_loss": 3.4425997536437,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002223330811424248,
      "loss": 2.5554,
      "theoretical_loss": 3.4421055311869604,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022150557972385097,
      "loss": 2.5813,
      "theoretical_loss": 3.441612387083208,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002206780783052771,
      "loss": 2.6066,
      "theoretical_loss": 3.44112031714851,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021985057688670323,
      "loss": 2.6411,
      "theoretical_loss": 3.4406293172222466,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021902307546812938,
      "loss": 2.6013,
      "theoretical_loss": 3.4401393831669447,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021819557404955552,
      "loss": 2.5705,
      "theoretical_loss": 3.4396505108681046,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021736807263098164,
      "loss": 2.5822,
      "theoretical_loss": 3.4391626962340394,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021654057121240782,
      "loss": 2.6193,
      "theoretical_loss": 3.4386759351957066,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021571306979383393,
      "loss": 2.5569,
      "theoretical_loss": 3.438190223706548,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021488556837526005,
      "loss": 2.5863,
      "theoretical_loss": 3.437705557742323,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002140580669566862,
      "loss": 2.5405,
      "theoretical_loss": 3.4372219333009535,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021323056553811237,
      "loss": 2.5992,
      "theoretical_loss": 3.436739346402362,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002124030641195385,
      "loss": 2.6134,
      "theoretical_loss": 3.4362577930883154,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002115755627009646,
      "loss": 2.5962,
      "theoretical_loss": 3.435777269422266,
      "tokens_seen": 1962803200
    },
    {
      "debugging/Self-BLEU-5": 0.4796244782337672,
      "debugging/distinct-1-grams": 0.7874466763773558,
      "debugging/distinct-2-grams": 0.9537592626603371,
      "debugging/entropy-1-grams": 6.225892551400493,
      "debugging/entropy-2-grams": 7.143796200176747,
      "debugging/length": 450.05,
      "debugging/num_segments": 20,
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.028914527967572212,
      "objective/train/avg_weight": 0.9689064621925354,
      "objective/train/docs_used": 1112341,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6380088329315186,
      "objective/train/std_weight": 0.10137132555246353,
      "objective/train/theoretical_loss": 3.4352977714892017,
      "objective/train/tokens_used": 1986540000,
      "theoretical_loss": 3.4352977714892017,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021074806128239078,
      "loss": 2.5954,
      "theoretical_loss": 3.4352977714892017,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002099205598638169,
      "loss": 2.5442,
      "theoretical_loss": 3.434819295395487,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020909305844524304,
      "loss": 2.5713,
      "theoretical_loss": 3.4343418372687156,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002082655570266692,
      "loss": 2.5697,
      "theoretical_loss": 3.433865393257557,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020743805560809533,
      "loss": 2.5261,
      "theoretical_loss": 3.433389959531607,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020661055418952145,
      "loss": 2.5397,
      "theoretical_loss": 3.4329155322812417,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020578305277094762,
      "loss": 2.5351,
      "theoretical_loss": 3.432442107717467,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020495555135237374,
      "loss": 2.5727,
      "theoretical_loss": 3.4319696820717778,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002041280499337999,
      "loss": 2.6081,
      "theoretical_loss": 3.4314982515960093,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020330054851522603,
      "loss": 2.62,
      "theoretical_loss": 3.4310278125621965,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020247304709665218,
      "loss": 2.6214,
      "theoretical_loss": 3.430558361262433,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002016455456780783,
      "loss": 2.5939,
      "theoretical_loss": 3.4300898940087254,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020081804425950447,
      "loss": 2.6077,
      "theoretical_loss": 3.4296224071328623,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001999905428409306,
      "loss": 2.5915,
      "theoretical_loss": 3.429155896986268,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001991630414223567,
      "loss": 2.5622,
      "theoretical_loss": 3.4286903599398704,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019833554000378285,
      "loss": 2.5999,
      "theoretical_loss": 3.428225792383964,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000197508038585209,
      "loss": 2.5942,
      "theoretical_loss": 3.4277621907280746,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019668053716663514,
      "loss": 2.5485,
      "theoretical_loss": 3.427299551400826,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019585303574806126,
      "loss": 2.5684,
      "theoretical_loss": 3.4268378708498086,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019502553432948743,
      "loss": 2.5498,
      "theoretical_loss": 3.426377145541448,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019419803291091355,
      "loss": 2.5284,
      "theoretical_loss": 3.425917371960872,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001933705314923397,
      "loss": 2.4971,
      "theoretical_loss": 3.425458546611787,
      "tokens_seen": 2034892800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019254303007376584,
      "loss": 2.4992,
      "theoretical_loss": 3.425000666016346,
      "tokens_seen": 2038169600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000191715528655192,
      "loss": 2.5058,
      "theoretical_loss": 3.4245437267150214,
      "tokens_seen": 2041446400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001908880272366181,
      "loss": 2.5501,
      "theoretical_loss": 3.4240877252664834,
      "tokens_seen": 2044723200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019006052581804428,
      "loss": 2.5238,
      "theoretical_loss": 3.4236326582474716,
      "tokens_seen": 2048000000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001892330243994704,
      "loss": 2.5157,
      "theoretical_loss": 3.423178522252672,
      "tokens_seen": 2051276800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018840552298089651,
      "loss": 2.5179,
      "theoretical_loss": 3.4227253138945946,
      "tokens_seen": 2054553600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018757802156232269,
      "loss": 2.5231,
      "theoretical_loss": 3.4222730298034545,
      "tokens_seen": 2057830400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018675052014374883,
      "loss": 2.5179,
      "theoretical_loss": 3.421821666627046,
      "tokens_seen": 2061107200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018592301872517495,
      "loss": 2.5355,
      "theoretical_loss": 3.4213712210306286,
      "tokens_seen": 2064384000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018509551730660112,
      "loss": 2.545,
      "theoretical_loss": 3.4209216896968044,
      "tokens_seen": 2067660800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018426801588802724,
      "loss": 2.4954,
      "theoretical_loss": 3.4204730693254035,
      "tokens_seen": 2070937600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018344051446945336,
      "loss": 2.5687,
      "theoretical_loss": 3.4200253566333645,
      "tokens_seen": 2074214400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001826130130508795,
      "loss": 2.5969,
      "theoretical_loss": 3.4195785483546226,
      "tokens_seen": 2077491200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018178551163230565,
      "loss": 2.5938,
      "theoretical_loss": 3.419132641239991,
      "tokens_seen": 2080768000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001809580102137318,
      "loss": 2.5466,
      "theoretical_loss": 3.4186876320570505,
      "tokens_seen": 2084044800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001801305087951579,
      "loss": 2.5695,
      "theoretical_loss": 3.418243517590036,
      "tokens_seen": 2087321600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017930300737658409,
      "loss": 2.5454,
      "theoretical_loss": 3.4178002946397217,
      "tokens_seen": 2090598400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001784755059580102,
      "loss": 2.5807,
      "theoretical_loss": 3.4173579600233155,
      "tokens_seen": 2093875200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017764800453943635,
      "loss": 2.5565,
      "theoretical_loss": 3.416916510574345,
      "tokens_seen": 2097152000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001768205031208625,
      "loss": 2.5393,
      "theoretical_loss": 3.4164759431425495,
      "tokens_seen": 2100428800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017599300170228864,
      "loss": 2.5671,
      "theoretical_loss": 3.4160362545937732,
      "tokens_seen": 2103705600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017516550028371476,
      "loss": 2.5602,
      "theoretical_loss": 3.4155974418098562,
      "tokens_seen": 2106982400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001743379988651409,
      "loss": 2.5624,
      "theoretical_loss": 3.415159501688528,
      "tokens_seen": 2110259200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017351049744656705,
      "loss": 2.5924,
      "theoretical_loss": 3.4147224311433053,
      "tokens_seen": 2113536000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001726829960279932,
      "loss": 2.6017,
      "theoretical_loss": 3.414286227103381,
      "tokens_seen": 2116812800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001718554946094193,
      "loss": 2.5792,
      "theoretical_loss": 3.413850886513529,
      "tokens_seen": 2120089600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017102799319084546,
      "loss": 2.5506,
      "theoretical_loss": 3.413416406333992,
      "tokens_seen": 2123366400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001702004917722716,
      "loss": 2.5629,
      "theoretical_loss": 3.412982783540387,
      "tokens_seen": 2126643200
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.011589955538511276,
      "objective/train/avg_weight": 0.9875407814979553,
      "objective/train/docs_used": 1206282,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.480736255645752,
      "objective/train/std_weight": 0.053698983043432236,
      "objective/train/theoretical_loss": 3.4125500151236,
      "objective/train/tokens_used": 2150380000,
      "theoretical_loss": 3.4125500151236,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016937299035369775,
      "loss": 2.5177,
      "theoretical_loss": 3.4125500151236,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001685454889351239,
      "loss": 2.542,
      "theoretical_loss": 3.412118098089686,
      "tokens_seen": 2133196800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016771798751655,
      "loss": 2.5607,
      "theoretical_loss": 3.411687029459771,
      "tokens_seen": 2136473600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016689048609797616,
      "loss": 2.5259,
      "theoretical_loss": 3.411256806269951,
      "tokens_seen": 2139750400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001660629846794023,
      "loss": 2.5255,
      "theoretical_loss": 3.4108274255711954,
      "tokens_seen": 2143027200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016523548326082845,
      "loss": 2.5475,
      "theoretical_loss": 3.4103988844292497,
      "tokens_seen": 2146304000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001644079818422546,
      "loss": 2.5731,
      "theoretical_loss": 3.4099711799245385,
      "tokens_seen": 2149580800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016358048042368074,
      "loss": 2.5864,
      "theoretical_loss": 3.4095443091520705,
      "tokens_seen": 2152857600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016275297900510686,
      "loss": 2.5488,
      "theoretical_loss": 3.4091182692213415,
      "tokens_seen": 2156134400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000161925477586533,
      "loss": 2.5084,
      "theoretical_loss": 3.4086930572562446,
      "tokens_seen": 2159411200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016109797616795915,
      "loss": 2.5382,
      "theoretical_loss": 3.4082686703949716,
      "tokens_seen": 2162688000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001602704747493853,
      "loss": 2.5694,
      "theoretical_loss": 3.4078451057899244,
      "tokens_seen": 2165964800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015944297333081144,
      "loss": 2.5917,
      "theoretical_loss": 3.4074223606076206,
      "tokens_seen": 2169241600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015861547191223756,
      "loss": 2.5855,
      "theoretical_loss": 3.4070004320286036,
      "tokens_seen": 2172518400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001577879704936637,
      "loss": 2.5467,
      "theoretical_loss": 3.4065793172473526,
      "tokens_seen": 2175795200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015696046907508985,
      "loss": 2.528,
      "theoretical_loss": 3.40615901347219,
      "tokens_seen": 2179072000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015613296765651597,
      "loss": 2.5646,
      "theoretical_loss": 3.405739517925197,
      "tokens_seen": 2182348800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001553054662379421,
      "loss": 2.5627,
      "theoretical_loss": 3.4053208278421208,
      "tokens_seen": 2185625600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015447796481936826,
      "loss": 2.5287,
      "theoretical_loss": 3.4049029404722897,
      "tokens_seen": 2188902400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001536504634007944,
      "loss": 2.5082,
      "theoretical_loss": 3.4044858530785254,
      "tokens_seen": 2192179200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015282296198222055,
      "loss": 2.5695,
      "theoretical_loss": 3.4040695629370563,
      "tokens_seen": 2195456000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015199546056364667,
      "loss": 2.5637,
      "theoretical_loss": 3.403654067337433,
      "tokens_seen": 2198732800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001511679591450728,
      "loss": 2.5283,
      "theoretical_loss": 3.4032393635824416,
      "tokens_seen": 2202009600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015034045772649896,
      "loss": 2.5736,
      "theoretical_loss": 3.4028254489880227,
      "tokens_seen": 2205286400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001495129563079251,
      "loss": 2.7156,
      "theoretical_loss": 3.4024123208831836,
      "tokens_seen": 2208563200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014868545488935125,
      "loss": 2.7728,
      "theoretical_loss": 3.4019999766099187,
      "tokens_seen": 2211840000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001478579534707774,
      "loss": 2.8195,
      "theoretical_loss": 3.4015884135231262,
      "tokens_seen": 2215116800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001470304520522035,
      "loss": 2.7383,
      "theoretical_loss": 3.401177628990526,
      "tokens_seen": 2218393600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014620295063362966,
      "loss": 2.7418,
      "theoretical_loss": 3.400767620392579,
      "tokens_seen": 2221670400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014537544921505577,
      "loss": 2.706,
      "theoretical_loss": 3.4003583851224075,
      "tokens_seen": 2224947200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014454794779648192,
      "loss": 2.6929,
      "theoretical_loss": 3.3999499205857138,
      "tokens_seen": 2228224000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001437204463779081,
      "loss": 2.6266,
      "theoretical_loss": 3.3995422242007036,
      "tokens_seen": 2231500800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001428929449593342,
      "loss": 2.6501,
      "theoretical_loss": 3.399135293398004,
      "tokens_seen": 2234777600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014206544354076036,
      "loss": 2.6064,
      "theoretical_loss": 3.39872912562059,
      "tokens_seen": 2238054400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001412379421221865,
      "loss": 2.5993,
      "theoretical_loss": 3.3983237183237023,
      "tokens_seen": 2241331200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014041044070361262,
      "loss": 2.5668,
      "theoretical_loss": 3.397919068974775,
      "tokens_seen": 2244608000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013958293928503877,
      "loss": 2.5433,
      "theoretical_loss": 3.3975151750533574,
      "tokens_seen": 2247884800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001387554378664649,
      "loss": 2.516,
      "theoretical_loss": 3.3971120340510375,
      "tokens_seen": 2251161600
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013794448647626252,
      "loss": 2.5333,
      "theoretical_loss": 3.39670964347137,
      "tokens_seen": 2254438400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013711698505768866,
      "loss": 2.5386,
      "theoretical_loss": 3.396308000829799,
      "tokens_seen": 2257715200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001362894836391148,
      "loss": 2.5126,
      "theoretical_loss": 3.395907103653585,
      "tokens_seen": 2260992000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013546198222054095,
      "loss": 2.5085,
      "theoretical_loss": 3.395506949481734,
      "tokens_seen": 2264268800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013465103083033856,
      "loss": 2.5255,
      "theoretical_loss": 3.3951075358649208,
      "tokens_seen": 2267545600
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001338235294117647,
      "loss": 2.4971,
      "theoretical_loss": 3.3947088603654203,
      "tokens_seen": 2270822400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013299602799319085,
      "loss": 2.5504,
      "theoretical_loss": 3.394310920557034,
      "tokens_seen": 2274099200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000132168526574617,
      "loss": 2.5251,
      "theoretical_loss": 3.393913714025021,
      "tokens_seen": 2277376000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013134102515604312,
      "loss": 2.54,
      "theoretical_loss": 3.393517238366023,
      "tokens_seen": 2280652800
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013051352373746926,
      "loss": 2.5642,
      "theoretical_loss": 3.3931214911880008,
      "tokens_seen": 2283929600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001296860223188954,
      "loss": 2.5427,
      "theoretical_loss": 3.3927264701101594,
      "tokens_seen": 2287206400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012885852090032153,
      "loss": 2.5809,
      "theoretical_loss": 3.392332172762883,
      "tokens_seen": 2290483200
    },
    {
      "debugging/Self-BLEU-5": 0.515096219338814,
      "debugging/distinct-1-grams": 0.7923118197529954,
      "debugging/distinct-2-grams": 0.9542743672238616,
      "debugging/entropy-1-grams": 6.147062465574793,
      "debugging/entropy-2-grams": 7.107820582543805,
      "debugging/length": 504.1666666666667,
      "debugging/num_segments": 18,
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.012544393539428711,
      "objective/train/avg_weight": 0.9865128397941589,
      "objective/train/docs_used": 1292636,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.50348162651062,
      "objective/train/std_weight": 0.048928581178188324,
      "objective/train/theoretical_loss": 3.3919385967876625,
      "objective/train/tokens_used": 2314220000,
      "theoretical_loss": 3.3919385967876625,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012803101948174767,
      "loss": 2.5491,
      "theoretical_loss": 3.3919385967876625,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012720351806317384,
      "loss": 2.5855,
      "theoretical_loss": 3.3915457398370323,
      "tokens_seen": 2297036800
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012639256667297142,
      "loss": 2.6482,
      "theoretical_loss": 3.3911535995744995,
      "tokens_seen": 2300313600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012556506525439757,
      "loss": 2.6972,
      "theoretical_loss": 3.390762173674479,
      "tokens_seen": 2303590400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012473756383582372,
      "loss": 2.6858,
      "theoretical_loss": 3.3903714598222243,
      "tokens_seen": 2306867200
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012391006241724986,
      "loss": 2.632,
      "theoretical_loss": 3.389981455713767,
      "tokens_seen": 2310144000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.000123082560998676,
      "loss": 2.6549,
      "theoretical_loss": 3.389592159055845,
      "tokens_seen": 2313420800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012225505958010212,
      "loss": 2.6161,
      "theoretical_loss": 3.389203567565843,
      "tokens_seen": 2316697600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012142755816152827,
      "loss": 2.5388,
      "theoretical_loss": 3.3888156789717243,
      "tokens_seen": 2319974400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012060005674295443,
      "loss": 2.6067,
      "theoretical_loss": 3.3884284910119704,
      "tokens_seen": 2323251200
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011977255532438055,
      "loss": 2.5811,
      "theoretical_loss": 3.3880420014355135,
      "tokens_seen": 2326528000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011894505390580669,
      "loss": 2.5567,
      "theoretical_loss": 3.387656208001678,
      "tokens_seen": 2329804800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011811755248723284,
      "loss": 2.5653,
      "theoretical_loss": 3.387271108480113,
      "tokens_seen": 2333081600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011729005106865897,
      "loss": 2.5541,
      "theoretical_loss": 3.386886700650736,
      "tokens_seen": 2336358400
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011646254965008511,
      "loss": 2.5613,
      "theoretical_loss": 3.386502982303667,
      "tokens_seen": 2339635200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011563504823151125,
      "loss": 2.5225,
      "theoretical_loss": 3.3861199512391673,
      "tokens_seen": 2342912000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011480754681293739,
      "loss": 2.5891,
      "theoretical_loss": 3.385737605267583,
      "tokens_seen": 2346188800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011398004539436354,
      "loss": 2.5543,
      "theoretical_loss": 3.38535594220928,
      "tokens_seen": 2349465600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011315254397578967,
      "loss": 2.5464,
      "theoretical_loss": 3.3849749598945875,
      "tokens_seen": 2352742400
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011232504255721581,
      "loss": 2.5367,
      "theoretical_loss": 3.3845946561637366,
      "tokens_seen": 2356019200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011149754113864196,
      "loss": 2.6255,
      "theoretical_loss": 3.384215028866803,
      "tokens_seen": 2359296000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011067003972006809,
      "loss": 2.6472,
      "theoretical_loss": 3.3838360758636474,
      "tokens_seen": 2362572800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010984253830149424,
      "loss": 2.5815,
      "theoretical_loss": 3.3834577950238582,
      "tokens_seen": 2365849600
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010901503688292038,
      "loss": 2.5966,
      "theoretical_loss": 3.383080184226693,
      "tokens_seen": 2369126400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001081875354643465,
      "loss": 2.6123,
      "theoretical_loss": 3.3827032413610225,
      "tokens_seen": 2372403200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010736003404577266,
      "loss": 2.5812,
      "theoretical_loss": 3.382326964325273,
      "tokens_seen": 2375680000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010653253262719878,
      "loss": 2.5612,
      "theoretical_loss": 3.3819513510273698,
      "tokens_seen": 2378956800
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010570503120862492,
      "loss": 2.5643,
      "theoretical_loss": 3.381576399384682,
      "tokens_seen": 2382233600
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010487752979005107,
      "loss": 2.6045,
      "theoretical_loss": 3.3812021073239675,
      "tokens_seen": 2385510400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001040500283714772,
      "loss": 2.6014,
      "theoretical_loss": 3.380828472781316,
      "tokens_seen": 2388787200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010322252695290335,
      "loss": 2.6035,
      "theoretical_loss": 3.380455493702095,
      "tokens_seen": 2392064000
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010239502553432949,
      "loss": 2.5852,
      "theoretical_loss": 3.3800831680408963,
      "tokens_seen": 2395340800
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010156752411575562,
      "loss": 2.5794,
      "theoretical_loss": 3.3797114937614814,
      "tokens_seen": 2398617600
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010074002269718177,
      "loss": 2.584,
      "theoretical_loss": 3.3793404688367277,
      "tokens_seen": 2401894400
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.991252127860791e-05,
      "loss": 2.5674,
      "theoretical_loss": 3.378970091248576,
      "tokens_seen": 2405171200
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.908501986003405e-05,
      "loss": 2.5548,
      "theoretical_loss": 3.3786003589879763,
      "tokens_seen": 2408448000
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.825751844146019e-05,
      "loss": 2.5132,
      "theoretical_loss": 3.3782312700548376,
      "tokens_seen": 2411724800
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.743001702288632e-05,
      "loss": 2.5159,
      "theoretical_loss": 3.377862822457974,
      "tokens_seen": 2415001600
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.660251560431247e-05,
      "loss": 2.5208,
      "theoretical_loss": 3.3774950142150537,
      "tokens_seen": 2418278400
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.577501418573861e-05,
      "loss": 2.5313,
      "theoretical_loss": 3.3771278433525476,
      "tokens_seen": 2421555200
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.494751276716473e-05,
      "loss": 2.5446,
      "theoretical_loss": 3.376761307905678,
      "tokens_seen": 2424832000
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.412001134859089e-05,
      "loss": 2.5497,
      "theoretical_loss": 3.376395405918369,
      "tokens_seen": 2428108800
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.329250993001704e-05,
      "loss": 2.5346,
      "theoretical_loss": 3.3760301354431963,
      "tokens_seen": 2431385600
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.246500851144315e-05,
      "loss": 2.5372,
      "theoretical_loss": 3.375665494541335,
      "tokens_seen": 2434662400
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.16375070928693e-05,
      "loss": 2.488,
      "theoretical_loss": 3.3753014812825133,
      "tokens_seen": 2437939200
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.081000567429543e-05,
      "loss": 2.4755,
      "theoretical_loss": 3.374938093744962,
      "tokens_seen": 2441216000
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.998250425572158e-05,
      "loss": 2.5082,
      "theoretical_loss": 3.3745753300153636,
      "tokens_seen": 2444492800
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.915500283714772e-05,
      "loss": 2.5035,
      "theoretical_loss": 3.374213188188808,
      "tokens_seen": 2447769600
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.834405144694533e-05,
      "loss": 2.507,
      "theoretical_loss": 3.373851666368742,
      "tokens_seen": 2451046400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.751655002837148e-05,
      "loss": 2.5011,
      "theoretical_loss": 3.373490762666921,
      "tokens_seen": 2454323200
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.009017497301101685,
      "objective/train/avg_weight": 0.9902958273887634,
      "objective/train/docs_used": 1380428,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6355104446411133,
      "objective/train/std_weight": 0.07502805441617966,
      "objective/train/theoretical_loss": 3.3731304752033617,
      "objective/train/tokens_used": 2478060000,
      "theoretical_loss": 3.3731304752033617,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.668904860979761e-05,
      "loss": 2.5452,
      "theoretical_loss": 3.3731304752033617,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.586154719122375e-05,
      "loss": 2.5376,
      "theoretical_loss": 3.372770802106297,
      "tokens_seen": 2460876800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.50340457726499e-05,
      "loss": 2.5238,
      "theoretical_loss": 3.3724117415121273,
      "tokens_seen": 2464153600
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.420654435407603e-05,
      "loss": 2.5297,
      "theoretical_loss": 3.372053291565374,
      "tokens_seen": 2467430400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.337904293550217e-05,
      "loss": 2.5479,
      "theoretical_loss": 3.3716954504186347,
      "tokens_seen": 2470707200
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.255154151692832e-05,
      "loss": 2.5133,
      "theoretical_loss": 3.371338216232536,
      "tokens_seen": 2473984000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.172404009835445e-05,
      "loss": 2.5265,
      "theoretical_loss": 3.370981587175689,
      "tokens_seen": 2477260800
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.089653867978058e-05,
      "loss": 2.5089,
      "theoretical_loss": 3.370625561424644,
      "tokens_seen": 2480537600
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.006903726120674e-05,
      "loss": 2.4709,
      "theoretical_loss": 3.3702701371638453,
      "tokens_seen": 2483814400
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.924153584263287e-05,
      "loss": 2.4729,
      "theoretical_loss": 3.369915312585586,
      "tokens_seen": 2487091200
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.8414034424059e-05,
      "loss": 2.4748,
      "theoretical_loss": 3.3695610858899663,
      "tokens_seen": 2490368000
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.758653300548515e-05,
      "loss": 2.447,
      "theoretical_loss": 3.3692074552848457,
      "tokens_seen": 2493644800
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.67590315869113e-05,
      "loss": 2.4936,
      "theoretical_loss": 3.368854418985804,
      "tokens_seen": 2496921600
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.593153016833743e-05,
      "loss": 2.4549,
      "theoretical_loss": 3.368501975216093,
      "tokens_seen": 2500198400
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.510402874976357e-05,
      "loss": 2.4805,
      "theoretical_loss": 3.3681501222065977,
      "tokens_seen": 2503475200
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.42765273311897e-05,
      "loss": 2.4758,
      "theoretical_loss": 3.367798858195792,
      "tokens_seen": 2506752000
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.344902591261585e-05,
      "loss": 2.5068,
      "theoretical_loss": 3.367448181429695,
      "tokens_seen": 2510028800
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.262152449404198e-05,
      "loss": 2.508,
      "theoretical_loss": 3.36709809016183,
      "tokens_seen": 2513305600
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.179402307546813e-05,
      "loss": 2.5267,
      "theoretical_loss": 3.3667485826531847,
      "tokens_seen": 2516582400
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.096652165689426e-05,
      "loss": 2.4982,
      "theoretical_loss": 3.366399657172166,
      "tokens_seen": 2519859200
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.01390202383204e-05,
      "loss": 2.481,
      "theoretical_loss": 3.36605131199456,
      "tokens_seen": 2523136000
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.931151881974655e-05,
      "loss": 2.4286,
      "theoretical_loss": 3.365703545403493,
      "tokens_seen": 2526412800
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.848401740117268e-05,
      "loss": 2.4159,
      "theoretical_loss": 3.3653563556893875,
      "tokens_seen": 2529689600
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.765651598259883e-05,
      "loss": 2.4417,
      "theoretical_loss": 3.365009741149925,
      "tokens_seen": 2532966400
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.682901456402497e-05,
      "loss": 2.4824,
      "theoretical_loss": 3.364663700090004,
      "tokens_seen": 2536243200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.60015131454511e-05,
      "loss": 2.4548,
      "theoretical_loss": 3.3643182308216986,
      "tokens_seen": 2539520000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.517401172687724e-05,
      "loss": 2.4356,
      "theoretical_loss": 3.3639733316642233,
      "tokens_seen": 2542796800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.434651030830338e-05,
      "loss": 2.4523,
      "theoretical_loss": 3.36362900094389,
      "tokens_seen": 2546073600
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.351900888972953e-05,
      "loss": 2.4694,
      "theoretical_loss": 3.3632852369940696,
      "tokens_seen": 2549350400
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.269150747115566e-05,
      "loss": 2.4794,
      "theoretical_loss": 3.362942038155154,
      "tokens_seen": 2552627200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.18640060525818e-05,
      "loss": 2.4109,
      "theoretical_loss": 3.362599402774518,
      "tokens_seen": 2555904000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.103650463400795e-05,
      "loss": 2.4748,
      "theoretical_loss": 3.362257329206479,
      "tokens_seen": 2559180800
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.020900321543408e-05,
      "loss": 2.4892,
      "theoretical_loss": 3.36191581581226,
      "tokens_seen": 2562457600
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.938150179686022e-05,
      "loss": 2.4737,
      "theoretical_loss": 3.3615748609599545,
      "tokens_seen": 2565734400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.857055040665784e-05,
      "loss": 2.4504,
      "theoretical_loss": 3.361234463024483,
      "tokens_seen": 2569011200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.774304898808398e-05,
      "loss": 2.4363,
      "theoretical_loss": 3.3608946203875614,
      "tokens_seen": 2572288000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.691554756951012e-05,
      "loss": 2.4679,
      "theoretical_loss": 3.360555331437662,
      "tokens_seen": 2575564800
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.608804615093625e-05,
      "loss": 2.5041,
      "theoretical_loss": 3.360216594569975,
      "tokens_seen": 2578841600
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5260544732362404e-05,
      "loss": 2.4293,
      "theoretical_loss": 3.3598784081863755,
      "tokens_seen": 2582118400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4433043313788535e-05,
      "loss": 2.4697,
      "theoretical_loss": 3.3595407706953835,
      "tokens_seen": 2585395200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3638641951957626e-05,
      "loss": 2.4687,
      "theoretical_loss": 3.359203680512131,
      "tokens_seen": 2588672000
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.281114053338377e-05,
      "loss": 2.4762,
      "theoretical_loss": 3.3588671360583247,
      "tokens_seen": 2591948800
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.198363911480991e-05,
      "loss": 2.4965,
      "theoretical_loss": 3.3585311357622096,
      "tokens_seen": 2595225600
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.115613769623605e-05,
      "loss": 2.4964,
      "theoretical_loss": 3.358195678058536,
      "tokens_seen": 2598502400
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0328636277662194e-05,
      "loss": 2.5347,
      "theoretical_loss": 3.3578607613885225,
      "tokens_seen": 2601779200
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.950113485908833e-05,
      "loss": 2.5302,
      "theoretical_loss": 3.357526384199822,
      "tokens_seen": 2605056000
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.867363344051447e-05,
      "loss": 2.5242,
      "theoretical_loss": 3.3571925449464874,
      "tokens_seen": 2608332800
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.78461320219406e-05,
      "loss": 2.5059,
      "theoretical_loss": 3.356859242088935,
      "tokens_seen": 2611609600
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7018630603366755e-05,
      "loss": 2.4959,
      "theoretical_loss": 3.356526474093914,
      "tokens_seen": 2614886400
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.619112918479289e-05,
      "loss": 2.5325,
      "theoretical_loss": 3.35619423943447,
      "tokens_seen": 2618163200
    },
    {
      "debugging/Self-BLEU-5": 0.449855913696806,
      "debugging/distinct-1-grams": 0.769778305351557,
      "debugging/distinct-2-grams": 0.9467712904639874,
      "debugging/entropy-1-grams": 5.998633443012235,
      "debugging/entropy-2-grams": 6.951908950814323,
      "debugging/length": 469.11764705882354,
      "debugging/num_segments": 17,
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.039095163345336914,
      "objective/train/avg_weight": 0.9579753279685974,
      "objective/train/docs_used": 1474904,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.611696243286133,
      "objective/train/std_weight": 0.11919951438903809,
      "objective/train/theoretical_loss": 3.3558625365899117,
      "objective/train/tokens_used": 2641900000,
      "theoretical_loss": 3.3558625365899117,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5363627766219026e-05,
      "loss": 2.5499,
      "theoretical_loss": 3.3558625365899117,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4536126347645164e-05,
      "loss": 2.5376,
      "theoretical_loss": 3.3555313640457776,
      "tokens_seen": 2624716800
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.37086249290713e-05,
      "loss": 2.4876,
      "theoretical_loss": 3.355200720293803,
      "tokens_seen": 2627993600
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.288112351049745e-05,
      "loss": 2.5103,
      "theoretical_loss": 3.3548706038318867,
      "tokens_seen": 2631270400
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.205362209192359e-05,
      "loss": 2.4624,
      "theoretical_loss": 3.3545410131640585,
      "tokens_seen": 2634547200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1226120673349725e-05,
      "loss": 2.489,
      "theoretical_loss": 3.354211946800446,
      "tokens_seen": 2637824000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0398619254775864e-05,
      "loss": 2.4885,
      "theoretical_loss": 3.3538834032572424,
      "tokens_seen": 2641100800
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9571117836202e-05,
      "loss": 2.5156,
      "theoretical_loss": 3.3535553810566743,
      "tokens_seen": 2644377600
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.874361641762815e-05,
      "loss": 2.5058,
      "theoretical_loss": 3.3532278787269707,
      "tokens_seen": 2647654400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7916114999054286e-05,
      "loss": 2.484,
      "theoretical_loss": 3.35290089480233,
      "tokens_seen": 2650931200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7088613580480425e-05,
      "loss": 2.4951,
      "theoretical_loss": 3.352574427822888,
      "tokens_seen": 2654208000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6261112161906564e-05,
      "loss": 2.4939,
      "theoretical_loss": 3.3522484763346885,
      "tokens_seen": 2657484800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.54336107433327e-05,
      "loss": 2.4904,
      "theoretical_loss": 3.3519230388896504,
      "tokens_seen": 2660761600
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.460610932475884e-05,
      "loss": 2.4957,
      "theoretical_loss": 3.3515981140455375,
      "tokens_seen": 2664038400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3778607906184986e-05,
      "loss": 2.5041,
      "theoretical_loss": 3.3512737003659288,
      "tokens_seen": 2667315200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.295110648761112e-05,
      "loss": 2.504,
      "theoretical_loss": 3.350949796420184,
      "tokens_seen": 2670592000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.212360506903726e-05,
      "loss": 2.5119,
      "theoretical_loss": 3.3506264007834186,
      "tokens_seen": 2673868800
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.12961036504634e-05,
      "loss": 2.5272,
      "theoretical_loss": 3.3503035120364704,
      "tokens_seen": 2677145600
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.046860223188954e-05,
      "loss": 2.5123,
      "theoretical_loss": 3.349981128765869,
      "tokens_seen": 2680422400
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.964110081331568e-05,
      "loss": 2.5086,
      "theoretical_loss": 3.3496592495638096,
      "tokens_seen": 2683699200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.881359939474182e-05,
      "loss": 2.4753,
      "theoretical_loss": 3.349337873028119,
      "tokens_seen": 2686976000
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7986097976167956e-05,
      "loss": 2.5089,
      "theoretical_loss": 3.349016997762229,
      "tokens_seen": 2690252800
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.71585965575941e-05,
      "loss": 2.4705,
      "theoretical_loss": 3.348696622375147,
      "tokens_seen": 2693529600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.633109513902024e-05,
      "loss": 2.484,
      "theoretical_loss": 3.3483767454814255,
      "tokens_seen": 2696806400
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.550359372044638e-05,
      "loss": 2.4954,
      "theoretical_loss": 3.3480573657011354,
      "tokens_seen": 2700083200
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.467609230187252e-05,
      "loss": 2.4715,
      "theoretical_loss": 3.347738481659836,
      "tokens_seen": 2703360000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3848590883298656e-05,
      "loss": 2.4621,
      "theoretical_loss": 3.347420091988546,
      "tokens_seen": 2706636800
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3021089464724798e-05,
      "loss": 2.4846,
      "theoretical_loss": 3.3471021953237177,
      "tokens_seen": 2709913600
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2193588046150937e-05,
      "loss": 2.5031,
      "theoretical_loss": 3.346784790307205,
      "tokens_seen": 2713190400
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1366086627577075e-05,
      "loss": 2.4827,
      "theoretical_loss": 3.346467875586239,
      "tokens_seen": 2716467200
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0538585209003217e-05,
      "loss": 2.562,
      "theoretical_loss": 3.3461514498133997,
      "tokens_seen": 2719744000
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9711083790429356e-05,
      "loss": 2.5616,
      "theoretical_loss": 3.345835511646587,
      "tokens_seen": 2723020800
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8883582371855494e-05,
      "loss": 2.5486,
      "theoretical_loss": 3.345520059748993,
      "tokens_seen": 2726297600
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8056080953281636e-05,
      "loss": 2.5346,
      "theoretical_loss": 3.3452050927890777,
      "tokens_seen": 2729574400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7228579534707775e-05,
      "loss": 2.5467,
      "theoretical_loss": 3.3448906094405393,
      "tokens_seen": 2732851200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6401078116133914e-05,
      "loss": 2.5381,
      "theoretical_loss": 3.344576608382288,
      "tokens_seen": 2736128000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5573576697560052e-05,
      "loss": 2.5619,
      "theoretical_loss": 3.34426308829842,
      "tokens_seen": 2739404800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4746075278986192e-05,
      "loss": 2.5578,
      "theoretical_loss": 3.343950047878189,
      "tokens_seen": 2742681600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3918573860412333e-05,
      "loss": 2.5669,
      "theoretical_loss": 3.343637485815983,
      "tokens_seen": 2745958400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3091072441838471e-05,
      "loss": 2.5565,
      "theoretical_loss": 3.3433254008112963,
      "tokens_seen": 2749235200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2263571023264612e-05,
      "loss": 2.5601,
      "theoretical_loss": 3.343013791568702,
      "tokens_seen": 2752512000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.143606960469075e-05,
      "loss": 2.5311,
      "theoretical_loss": 3.342702656797829,
      "tokens_seen": 2755788800
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.060856818611689e-05,
      "loss": 2.5079,
      "theoretical_loss": 3.342391995213335,
      "tokens_seen": 2759065600
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.78106676754303e-06,
      "loss": 2.5169,
      "theoretical_loss": 3.34208180553488,
      "tokens_seen": 2762342400
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.95356534896917e-06,
      "loss": 2.5037,
      "theoretical_loss": 3.341772086487102,
      "tokens_seen": 2765619200
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.12606393039531e-06,
      "loss": 2.5614,
      "theoretical_loss": 3.341462836799593,
      "tokens_seen": 2768896000
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.298562511821448e-06,
      "loss": 2.5543,
      "theoretical_loss": 3.341154055206871,
      "tokens_seen": 2772172800
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.471061093247588e-06,
      "loss": 2.6017,
      "theoretical_loss": 3.3408457404483576,
      "tokens_seen": 2775449600
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.643559674673727e-06,
      "loss": 2.5745,
      "theoretical_loss": 3.3405378912683514,
      "tokens_seen": 2778726400
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.8160582560998674e-06,
      "loss": 2.5935,
      "theoretical_loss": 3.340230506416005,
      "tokens_seen": 2782003200
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.025402138009667397,
      "objective/train/avg_weight": 0.9726998209953308,
      "objective/train/docs_used": 1566749,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6003384590148926,
      "objective/train/std_weight": 0.12505510449409485,
      "objective/train/theoretical_loss": 3.3399235846453004,
      "objective/train/tokens_used": 2805740000,
      "theoretical_loss": 3.3399235846453004,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.988556837526007e-06,
      "loss": 2.6075,
      "theoretical_loss": 3.3399235846453004,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.161055418952147e-06,
      "loss": 2.5666,
      "theoretical_loss": 3.3396171247150224,
      "tokens_seen": 2788556800
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3335540003782866e-06,
      "loss": 2.5112,
      "theoretical_loss": 3.3393111253887375,
      "tokens_seen": 2791833600
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5060525818044258e-06,
      "loss": 2.5268,
      "theoretical_loss": 3.339005585434769,
      "tokens_seen": 2795110400
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.785511632305656e-07,
      "loss": 2.4972,
      "theoretical_loss": 3.338700503626171,
      "tokens_seen": 2798387200
    }
  ],
  "max_steps": 42724,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 1.428920419221504e+18,
  "trial_name": null,
  "trial_params": null
}