{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5675488430095608,
  "eval_steps": 1024,
  "global_step": 12288,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.011823934229365849,
      "grad_norm": 1.1181285381317139,
      "learning_rate": 1.9615384615384617e-05,
      "loss": 10.3794,
      "step": 256
    },
    {
      "epoch": 0.023647868458731697,
      "grad_norm": 1.0416101217269897,
      "learning_rate": 3.930769230769231e-05,
      "loss": 7.9323,
      "step": 512
    },
    {
      "epoch": 0.03547180268809755,
      "grad_norm": 0.8387451171875,
      "learning_rate": 4.999617095521894e-05,
      "loss": 5.6072,
      "step": 768
    },
    {
      "epoch": 0.047295736917463395,
      "grad_norm": 0.559028148651123,
      "learning_rate": 4.9961092368776736e-05,
      "loss": 3.8081,
      "step": 1024
    },
    {
      "epoch": 0.047295736917463395,
      "eval_acr_loss": 0.9904639322191613,
      "eval_across_var": 0.004779557895753107,
      "eval_bleu": 0.5734765992441645,
      "eval_ce_loss": 2.3428709033417374,
      "eval_cos_loss": 0.9217384038450511,
      "eval_cov": 0.07704977793236302,
      "eval_cov_loss": 0.009401565545225797,
      "eval_glb_loss": 0.3516415548106851,
      "eval_global_kurtosis": 3.025142035527861,
      "eval_global_mean": 0.0031828225747635376,
      "eval_global_var": 0.4070052752211758,
      "eval_krt_loss": 0.0007855186364893808,
      "eval_loss": 2.8181393391465486,
      "eval_mse_loss": 1.9018371273937835,
      "eval_per_loss": 0.36342586774259944,
      "eval_per_var": 0.3971644013984018,
      "eval_within_var": 0.4021891167038652,
      "eval_wth_loss": 0.35738333345276035,
      "step": 1024
    },
    {
      "epoch": 0.047295736917463395,
      "eval_acr_loss": 0.9904639322191613,
      "eval_across_var": 0.004779557895753107,
      "eval_bleu": 0.5734765992441645,
      "eval_ce_loss": 2.3428709033417374,
      "eval_cos_loss": 0.9217384038450511,
      "eval_cov": 0.07704977793236302,
      "eval_cov_loss": 0.009401565545225797,
      "eval_glb_loss": 0.3516415548106851,
      "eval_global_kurtosis": 3.025142035527861,
      "eval_global_mean": 0.0031828225747635376,
      "eval_global_var": 0.4070052752211758,
      "eval_krt_loss": 0.0007855186364893808,
      "eval_loss": 2.8181393391465486,
      "eval_mse_loss": 1.9018371273937835,
      "eval_per_loss": 0.36342586774259944,
      "eval_per_var": 0.3971644013984018,
      "eval_runtime": 158.2867,
      "eval_samples_per_second": 176.85,
      "eval_steps_per_second": 2.767,
      "eval_within_var": 0.4021891167038652,
      "eval_wth_loss": 0.35738333345276035,
      "step": 1024
    },
    {
      "epoch": 0.05911967114682925,
      "grad_norm": 0.3861701786518097,
      "learning_rate": 4.988941132556799e-05,
      "loss": 2.7403,
      "step": 1280
    },
    {
      "epoch": 0.0709436053761951,
      "grad_norm": 0.2994460463523865,
      "learning_rate": 4.9781232937269974e-05,
      "loss": 2.1074,
      "step": 1536
    },
    {
      "epoch": 0.08276753960556095,
      "grad_norm": 0.24324144423007965,
      "learning_rate": 4.963671583455164e-05,
      "loss": 1.697,
      "step": 1792
    },
    {
      "epoch": 0.09459147383492679,
      "grad_norm": 0.2096453607082367,
      "learning_rate": 4.945607193446079e-05,
      "loss": 1.4167,
      "step": 2048
    },
    {
      "epoch": 0.09459147383492679,
      "eval_acr_loss": 0.96901579629885,
      "eval_across_var": 0.015615421451509135,
      "eval_bleu": 0.8100885045319792,
      "eval_ce_loss": 0.6945271475674355,
      "eval_cos_loss": 0.7685760948211635,
      "eval_cov": 0.09640809498965468,
      "eval_cov_loss": 0.014826197332392136,
      "eval_glb_loss": 0.00014464876967478017,
      "eval_global_kurtosis": 3.0129096426375925,
      "eval_global_mean": 0.010630951624482734,
      "eval_global_var": 0.9908018068635844,
      "eval_krt_loss": 0.0002954138530850955,
      "eval_loss": 1.056928829243194,
      "eval_mse_loss": 1.6510266498343584,
      "eval_per_loss": 0.0006002873590547744,
      "eval_per_var": 0.9770318296946348,
      "eval_within_var": 0.9752532089954098,
      "eval_wth_loss": 0.0006678303594037685,
      "step": 2048
    },
    {
      "epoch": 0.09459147383492679,
      "eval_acr_loss": 0.96901579629885,
      "eval_across_var": 0.015615421451509135,
      "eval_bleu": 0.8100885045319792,
      "eval_ce_loss": 0.6945271475674355,
      "eval_cos_loss": 0.7685760948211635,
      "eval_cov": 0.09640809498965468,
      "eval_cov_loss": 0.014826197332392136,
      "eval_glb_loss": 0.00014464876967478017,
      "eval_global_kurtosis": 3.0129096426375925,
      "eval_global_mean": 0.010630951624482734,
      "eval_global_var": 0.9908018068635844,
      "eval_krt_loss": 0.0002954138530850955,
      "eval_loss": 1.056928829243194,
      "eval_mse_loss": 1.6510266498343584,
      "eval_per_loss": 0.0006002873590547744,
      "eval_per_var": 0.9770318296946348,
      "eval_runtime": 153.1917,
      "eval_samples_per_second": 182.732,
      "eval_steps_per_second": 2.859,
      "eval_within_var": 0.9752532089954098,
      "eval_wth_loss": 0.0006678303594037685,
      "step": 2048
    },
    {
      "epoch": 0.10641540806429264,
      "grad_norm": 0.17944040894508362,
      "learning_rate": 4.923956612967301e-05,
      "loss": 1.2141,
      "step": 2304
    },
    {
      "epoch": 0.1182393422936585,
      "grad_norm": 0.1666973978281021,
      "learning_rate": 4.898751590005826e-05,
      "loss": 1.0533,
      "step": 2560
    },
    {
      "epoch": 0.13006327652302435,
      "grad_norm": 0.151300847530365,
      "learning_rate": 4.870029084713462e-05,
      "loss": 0.9289,
      "step": 2816
    },
    {
      "epoch": 0.1418872107523902,
      "grad_norm": 0.12945112586021423,
      "learning_rate": 4.837831215209188e-05,
      "loss": 0.8296,
      "step": 3072
    },
    {
      "epoch": 0.1418872107523902,
      "eval_acr_loss": 0.9623512120823882,
      "eval_across_var": 0.019007292450682077,
      "eval_bleu": 0.8976890806937166,
      "eval_ce_loss": 0.323182335727291,
      "eval_cos_loss": 0.6147451090486082,
      "eval_cov": 0.07890243181899258,
      "eval_cov_loss": 0.010057339255515163,
      "eval_glb_loss": 0.0003083459862835331,
      "eval_global_kurtosis": 3.006787543971789,
      "eval_global_mean": 0.013840801639643978,
      "eval_global_var": 1.0159304544805936,
      "eval_krt_loss": 0.0002449526874326567,
      "eval_loss": 0.6247616745021245,
      "eval_mse_loss": 1.3700275383039153,
      "eval_per_loss": 5.431099025081826e-05,
      "eval_per_var": 0.9992698077197488,
      "eval_within_var": 0.997042422425257,
      "eval_wth_loss": 7.23102295562176e-05,
      "step": 3072
    },
    {
      "epoch": 0.1418872107523902,
      "eval_acr_loss": 0.9623512120823882,
      "eval_across_var": 0.019007292450682077,
      "eval_bleu": 0.8976890806937166,
      "eval_ce_loss": 0.323182335727291,
      "eval_cos_loss": 0.6147451090486082,
      "eval_cov": 0.07890243181899258,
      "eval_cov_loss": 0.010057339255515163,
      "eval_glb_loss": 0.0003083459862835331,
      "eval_global_kurtosis": 3.006787543971789,
      "eval_global_mean": 0.013840801639643978,
      "eval_global_var": 1.0159304544805936,
      "eval_krt_loss": 0.0002449526874326567,
      "eval_loss": 0.6247616745021245,
      "eval_mse_loss": 1.3700275383039153,
      "eval_per_loss": 5.431099025081826e-05,
      "eval_per_var": 0.9992698077197488,
      "eval_runtime": 153.8502,
      "eval_samples_per_second": 181.95,
      "eval_steps_per_second": 2.847,
      "eval_within_var": 0.997042422425257,
      "eval_wth_loss": 7.23102295562176e-05,
      "step": 3072
    },
    {
      "epoch": 0.15371114498175603,
      "grad_norm": 0.12589366734027863,
      "learning_rate": 4.802205195817963e-05,
      "loss": 0.7505,
      "step": 3328
    },
    {
      "epoch": 0.1655350792111219,
      "grad_norm": 0.11897846311330795,
      "learning_rate": 4.763203267836576e-05,
      "loss": 0.6825,
      "step": 3584
    },
    {
      "epoch": 0.17735901344048774,
      "grad_norm": 0.10772541910409927,
      "learning_rate": 4.720882622928019e-05,
      "loss": 0.6261,
      "step": 3840
    },
    {
      "epoch": 0.18918294766985358,
      "grad_norm": 0.10752860456705093,
      "learning_rate": 4.675305319256765e-05,
      "loss": 0.5794,
      "step": 4096
    },
    {
      "epoch": 0.18918294766985358,
      "eval_acr_loss": 0.9460189426870651,
      "eval_across_var": 0.027373042131974,
      "eval_bleu": 0.9357963139785648,
      "eval_ce_loss": 0.18706960009929796,
      "eval_cos_loss": 0.49577669084888615,
      "eval_cov": 0.07429023638163527,
      "eval_cov_loss": 0.0088696325406194,
      "eval_glb_loss": 0.0008705457051595052,
      "eval_global_kurtosis": 3.0126239273646105,
      "eval_global_mean": 0.019379994640611624,
      "eval_global_var": 1.0280014804509132,
      "eval_krt_loss": 0.0004827787981839751,
      "eval_loss": 0.44092568850408403,
      "eval_mse_loss": 1.145899186395619,
      "eval_per_loss": 0.000201601960343313,
      "eval_per_var": 1.0113285709189497,
      "eval_within_var": 1.0008215422499669,
      "eval_wth_loss": 0.00011882184727399328,
      "step": 4096
    },
    {
      "epoch": 0.18918294766985358,
      "eval_acr_loss": 0.9460189426870651,
      "eval_across_var": 0.027373042131974,
      "eval_bleu": 0.9357963139785648,
      "eval_ce_loss": 0.18706960009929796,
      "eval_cos_loss": 0.49577669084888615,
      "eval_cov": 0.07429023638163527,
      "eval_cov_loss": 0.0088696325406194,
      "eval_glb_loss": 0.0008705457051595052,
      "eval_global_kurtosis": 3.0126239273646105,
      "eval_global_mean": 0.019379994640611624,
      "eval_global_var": 1.0280014804509132,
      "eval_krt_loss": 0.0004827787981839751,
      "eval_loss": 0.44092568850408403,
      "eval_mse_loss": 1.145899186395619,
      "eval_per_loss": 0.000201601960343313,
      "eval_per_var": 1.0113285709189497,
      "eval_runtime": 154.1862,
      "eval_samples_per_second": 181.553,
      "eval_steps_per_second": 2.841,
      "eval_within_var": 1.0008215422499669,
      "eval_wth_loss": 0.00011882184727399328,
      "step": 4096
    },
    {
      "epoch": 0.20100688189921945,
      "grad_norm": 0.1053977757692337,
      "learning_rate": 4.6265381904878854e-05,
      "loss": 0.5371,
      "step": 4352
    },
    {
      "epoch": 0.2128308161285853,
      "grad_norm": 0.09793444722890854,
      "learning_rate": 4.57465274778347e-05,
      "loss": 0.5051,
      "step": 4608
    },
    {
      "epoch": 0.22465475035795113,
      "grad_norm": 0.09642181545495987,
      "learning_rate": 4.519725074940068e-05,
      "loss": 0.4712,
      "step": 4864
    },
    {
      "epoch": 0.236478684587317,
      "grad_norm": 0.09021387249231339,
      "learning_rate": 4.461835716820895e-05,
      "loss": 0.4441,
      "step": 5120
    },
    {
      "epoch": 0.236478684587317,
      "eval_acr_loss": 0.8976050177121271,
      "eval_across_var": 0.05263728992460661,
      "eval_bleu": 0.9557004800288741,
      "eval_ce_loss": 0.12269850065038629,
      "eval_cos_loss": 0.4099505911815112,
      "eval_cov": 0.0774190545626427,
      "eval_cov_loss": 0.00986302298595746,
      "eval_glb_loss": 0.002168635799460215,
      "eval_global_kurtosis": 3.0220293960614835,
      "eval_global_mean": 0.026287226916448168,
      "eval_global_var": 1.0453655643550228,
      "eval_krt_loss": 0.0009509958784072548,
      "eval_loss": 0.3404404183500978,
      "eval_mse_loss": 0.9835099409160004,
      "eval_per_loss": 0.0008347246744861342,
      "eval_per_var": 1.0266046393407535,
      "eval_within_var": 0.993057064541943,
      "eval_wth_loss": 0.00020235125328104065,
      "step": 5120
    },
    {
      "epoch": 0.236478684587317,
      "eval_acr_loss": 0.8976050177121271,
      "eval_across_var": 0.05263728992460661,
      "eval_bleu": 0.9557004800288741,
      "eval_ce_loss": 0.12269850065038629,
      "eval_cos_loss": 0.4099505911815112,
      "eval_cov": 0.0774190545626427,
      "eval_cov_loss": 0.00986302298595746,
      "eval_glb_loss": 0.002168635799460215,
      "eval_global_kurtosis": 3.0220293960614835,
      "eval_global_mean": 0.026287226916448168,
      "eval_global_var": 1.0453655643550228,
      "eval_krt_loss": 0.0009509958784072548,
      "eval_loss": 0.3404404183500978,
      "eval_mse_loss": 0.9835099409160004,
      "eval_per_loss": 0.0008347246744861342,
      "eval_per_var": 1.0266046393407535,
      "eval_runtime": 152.7613,
      "eval_samples_per_second": 183.247,
      "eval_steps_per_second": 2.867,
      "eval_within_var": 0.993057064541943,
      "eval_wth_loss": 0.00020235125328104065,
      "step": 5120
    },
    {
      "epoch": 0.24830261881668284,
      "grad_norm": 0.08437229692935944,
      "learning_rate": 4.401069561246422e-05,
      "loss": 0.4198,
      "step": 5376
    },
    {
      "epoch": 0.2601265530460487,
      "grad_norm": 0.09224370121955872,
      "learning_rate": 4.337515714516545e-05,
      "loss": 0.398,
      "step": 5632
    },
    {
      "epoch": 0.27195048727541454,
      "grad_norm": 0.09789060801267624,
      "learning_rate": 4.2712673707468434e-05,
      "loss": 0.3822,
      "step": 5888
    },
    {
      "epoch": 0.2837744215047804,
      "grad_norm": 0.08814027905464172,
      "learning_rate": 4.202421675210565e-05,
      "loss": 0.3652,
      "step": 6144
    },
    {
      "epoch": 0.2837744215047804,
      "eval_acr_loss": 0.8202883884242681,
      "eval_across_var": 0.09450281369733755,
      "eval_bleu": 0.9674697145362666,
      "eval_ce_loss": 0.0872740181817856,
      "eval_cos_loss": 0.35029126064145943,
      "eval_cov": 0.08723331259810217,
      "eval_cov_loss": 0.012916838626917367,
      "eval_glb_loss": 0.00597613696093973,
      "eval_global_kurtosis": 3.024092547969731,
      "eval_global_mean": 0.04041210157141838,
      "eval_global_var": 1.0756590682077625,
      "eval_krt_loss": 0.0019540588812798874,
      "eval_loss": 0.27823208194209015,
      "eval_mse_loss": 0.8730145360781177,
      "eval_per_loss": 0.0027831836378193336,
      "eval_per_var": 1.0494290007848173,
      "eval_within_var": 0.9819663109844678,
      "eval_wth_loss": 0.00048046983986418896,
      "step": 6144
    },
    {
      "epoch": 0.2837744215047804,
      "eval_acr_loss": 0.8202883884242681,
      "eval_across_var": 0.09450281369733755,
      "eval_bleu": 0.9674697145362666,
      "eval_ce_loss": 0.0872740181817856,
      "eval_cos_loss": 0.35029126064145943,
      "eval_cov": 0.08723331259810217,
      "eval_cov_loss": 0.012916838626917367,
      "eval_glb_loss": 0.00597613696093973,
      "eval_global_kurtosis": 3.024092547969731,
      "eval_global_mean": 0.04041210157141838,
      "eval_global_var": 1.0756590682077625,
      "eval_krt_loss": 0.0019540588812798874,
      "eval_loss": 0.27823208194209015,
      "eval_mse_loss": 0.8730145360781177,
      "eval_per_loss": 0.0027831836378193336,
      "eval_per_var": 1.0494290007848173,
      "eval_runtime": 152.787,
      "eval_samples_per_second": 183.216,
      "eval_steps_per_second": 2.867,
      "eval_within_var": 0.9819663109844678,
      "eval_wth_loss": 0.00048046983986418896,
      "step": 6144
    },
    {
      "epoch": 0.2955983557341462,
      "grad_norm": 0.0873865932226181,
      "learning_rate": 4.131362984409538e-05,
      "loss": 0.3471,
      "step": 6400
    },
    {
      "epoch": 0.30742228996351206,
      "grad_norm": 0.07706636935472488,
      "learning_rate": 4.0576382427305324e-05,
      "loss": 0.3341,
      "step": 6656
    },
    {
      "epoch": 0.3192462241928779,
      "grad_norm": 0.09675078094005585,
      "learning_rate": 3.981629410852463e-05,
      "loss": 0.323,
      "step": 6912
    },
    {
      "epoch": 0.3310701584222438,
      "grad_norm": 0.09393124282360077,
      "learning_rate": 3.903447946637093e-05,
      "loss": 0.3086,
      "step": 7168
    },
    {
      "epoch": 0.3310701584222438,
      "eval_acr_loss": 0.726046212868059,
      "eval_across_var": 0.14836800615537113,
      "eval_bleu": 0.9750192561822552,
      "eval_ce_loss": 0.06533130286389018,
      "eval_cos_loss": 0.3090691367785136,
      "eval_cov": 0.0970152415097032,
      "eval_cov_loss": 0.01665312849988812,
      "eval_glb_loss": 0.012594995977671723,
      "eval_global_kurtosis": 3.052053846180711,
      "eval_global_mean": 0.055559001556814536,
      "eval_global_var": 1.1099212507134704,
      "eval_krt_loss": 0.005212522339852683,
      "eval_loss": 0.23639628832062629,
      "eval_mse_loss": 0.800027005204327,
      "eval_per_loss": 0.004818411722575149,
      "eval_per_var": 1.0654642462186072,
      "eval_within_var": 0.9631439711405262,
      "eval_wth_loss": 0.0015324829790224052,
      "step": 7168
    },
    {
      "epoch": 0.3310701584222438,
      "eval_acr_loss": 0.726046212868059,
      "eval_across_var": 0.14836800615537113,
      "eval_bleu": 0.9750192561822552,
      "eval_ce_loss": 0.06533130286389018,
      "eval_cos_loss": 0.3090691367785136,
      "eval_cov": 0.0970152415097032,
      "eval_cov_loss": 0.01665312849988812,
      "eval_glb_loss": 0.012594995977671723,
      "eval_global_kurtosis": 3.052053846180711,
      "eval_global_mean": 0.055559001556814536,
      "eval_global_var": 1.1099212507134704,
      "eval_krt_loss": 0.005212522339852683,
      "eval_loss": 0.23639628832062629,
      "eval_mse_loss": 0.800027005204327,
      "eval_per_loss": 0.004818411722575149,
      "eval_per_var": 1.0654642462186072,
      "eval_runtime": 149.7025,
      "eval_samples_per_second": 186.991,
      "eval_steps_per_second": 2.926,
      "eval_within_var": 0.9631439711405262,
      "eval_wth_loss": 0.0015324829790224052,
      "step": 7168
    },
    {
      "epoch": 0.34289409265160964,
      "grad_norm": 0.08861193060874939,
      "learning_rate": 3.823208493851674e-05,
      "loss": 0.3004,
      "step": 7424
    },
    {
      "epoch": 0.3547180268809755,
      "grad_norm": 0.10392362624406815,
      "learning_rate": 3.741028714057574e-05,
      "loss": 0.2906,
      "step": 7680
    },
    {
      "epoch": 0.3665419611103413,
      "grad_norm": 0.09276238083839417,
      "learning_rate": 3.657029114073668e-05,
      "loss": 0.2807,
      "step": 7936
    },
    {
      "epoch": 0.37836589533970716,
      "grad_norm": 0.08639795333147049,
      "learning_rate": 3.571332869267499e-05,
      "loss": 0.2715,
      "step": 8192
    },
    {
      "epoch": 0.37836589533970716,
      "eval_acr_loss": 0.6639264498127105,
      "eval_across_var": 0.18578778622357267,
      "eval_bleu": 0.9802777908340867,
      "eval_ce_loss": 0.05099814674778752,
      "eval_cos_loss": 0.2809805735333325,
      "eval_cov": 0.10055263292843893,
      "eval_cov_loss": 0.018623638194326532,
      "eval_glb_loss": 0.017183700108637005,
      "eval_global_kurtosis": 3.0215004377713486,
      "eval_global_mean": 0.06458907802355345,
      "eval_global_var": 1.1285004637557077,
      "eval_krt_loss": 0.0026812796775613885,
      "eval_loss": 0.20853681022037654,
      "eval_mse_loss": 0.7543587909169394,
      "eval_per_loss": 0.004951255506576468,
      "eval_per_var": 1.0664095943921232,
      "eval_within_var": 0.9453123360191851,
      "eval_wth_loss": 0.0031645966432090277,
      "step": 8192
    },
    {
      "epoch": 0.37836589533970716,
      "eval_acr_loss": 0.6639264498127105,
      "eval_across_var": 0.18578778622357267,
      "eval_bleu": 0.9802777908340867,
      "eval_ce_loss": 0.05099814674778752,
      "eval_cos_loss": 0.2809805735333325,
      "eval_cov": 0.10055263292843893,
      "eval_cov_loss": 0.018623638194326532,
      "eval_glb_loss": 0.017183700108637005,
      "eval_global_kurtosis": 3.0215004377713486,
      "eval_global_mean": 0.06458907802355345,
      "eval_global_var": 1.1285004637557077,
      "eval_krt_loss": 0.0026812796775613885,
      "eval_loss": 0.20853681022037654,
      "eval_mse_loss": 0.7543587909169394,
      "eval_per_loss": 0.004951255506576468,
      "eval_per_var": 1.0664095943921232,
      "eval_runtime": 151.9683,
      "eval_samples_per_second": 184.203,
      "eval_steps_per_second": 2.882,
      "eval_within_var": 0.9453123360191851,
      "eval_wth_loss": 0.0031645966432090277,
      "step": 8192
    },
    {
      "epoch": 0.390189829569073,
      "grad_norm": 0.09597329795360565,
      "learning_rate": 3.484065642933335e-05,
      "loss": 0.2653,
      "step": 8448
    },
    {
      "epoch": 0.4020137637984389,
      "grad_norm": 0.09884477406740189,
      "learning_rate": 3.395704566041868e-05,
      "loss": 0.2588,
      "step": 8704
    },
    {
      "epoch": 0.41383769802780473,
      "grad_norm": 0.12420962005853653,
      "learning_rate": 3.305686266521002e-05,
      "loss": 0.251,
      "step": 8960
    },
    {
      "epoch": 0.4256616322571706,
      "grad_norm": 0.09434681385755539,
      "learning_rate": 3.214486524445379e-05,
      "loss": 0.2459,
      "step": 9216
    },
    {
      "epoch": 0.4256616322571706,
      "eval_acr_loss": 0.5954188222074073,
      "eval_across_var": 0.22920796821174555,
      "eval_bleu": 0.9836602165246535,
      "eval_ce_loss": 0.041640645169470136,
      "eval_cos_loss": 0.2611509262835054,
      "eval_cov": 0.1050533486283533,
      "eval_cov_loss": 0.02121386145875198,
      "eval_glb_loss": 0.02622999339343206,
      "eval_global_kurtosis": 3.059722278216114,
      "eval_global_mean": 0.07356201894751423,
      "eval_global_var": 1.1591975242579908,
      "eval_krt_loss": 0.0066943209811956075,
      "eval_loss": 0.18931484470764795,
      "eval_mse_loss": 0.7253989612675149,
      "eval_per_loss": 0.007019040791411378,
      "eval_per_var": 1.0800580586472603,
      "eval_within_var": 0.9334883986542758,
      "eval_wth_loss": 0.004603328326420399,
      "step": 9216
    },
    {
      "epoch": 0.4256616322571706,
      "eval_acr_loss": 0.5954188222074073,
      "eval_across_var": 0.22920796821174555,
      "eval_bleu": 0.9836602165246535,
      "eval_ce_loss": 0.041640645169470136,
      "eval_cos_loss": 0.2611509262835054,
      "eval_cov": 0.1050533486283533,
      "eval_cov_loss": 0.02121386145875198,
      "eval_glb_loss": 0.02622999339343206,
      "eval_global_kurtosis": 3.059722278216114,
      "eval_global_mean": 0.07356201894751423,
      "eval_global_var": 1.1591975242579908,
      "eval_krt_loss": 0.0066943209811956075,
      "eval_loss": 0.18931484470764795,
      "eval_mse_loss": 0.7253989612675149,
      "eval_per_loss": 0.007019040791411378,
      "eval_per_var": 1.0800580586472603,
      "eval_runtime": 152.3156,
      "eval_samples_per_second": 183.783,
      "eval_steps_per_second": 2.876,
      "eval_within_var": 0.9334883986542758,
      "eval_wth_loss": 0.004603328326420399,
      "step": 9216
    },
    {
      "epoch": 0.4374855664865364,
      "grad_norm": 0.09029538929462433,
      "learning_rate": 3.122239073329827e-05,
      "loss": 0.241,
      "step": 9472
    },
    {
      "epoch": 0.44930950071590225,
      "grad_norm": 0.0906376913189888,
      "learning_rate": 3.029079183029181e-05,
      "loss": 0.2373,
      "step": 9728
    },
    {
      "epoch": 0.4611334349452681,
      "grad_norm": 0.08700253814458847,
      "learning_rate": 2.935143461381221e-05,
      "loss": 0.2319,
      "step": 9984
    },
    {
      "epoch": 0.472957369174634,
      "grad_norm": 0.12261617183685303,
      "learning_rate": 2.8405696538876124e-05,
      "loss": 0.2291,
      "step": 10240
    },
    {
      "epoch": 0.472957369174634,
      "eval_acr_loss": 0.5550906222024464,
      "eval_across_var": 0.2559552679867505,
      "eval_bleu": 0.9864228314384785,
      "eval_ce_loss": 0.03463799520246855,
      "eval_cos_loss": 0.2463943555815035,
      "eval_cov": 0.10881105187821062,
      "eval_cov_loss": 0.023118084696329894,
      "eval_glb_loss": 0.031056445483203347,
      "eval_global_kurtosis": 3.0601126701320145,
      "eval_global_mean": 0.08328581727258691,
      "eval_global_var": 1.173308629423516,
      "eval_krt_loss": 0.0071202065942054926,
      "eval_loss": 0.17534998668247162,
      "eval_mse_loss": 0.7060943077144013,
      "eval_per_loss": 0.007063817760171411,
      "eval_per_var": 1.0801762271689497,
      "eval_within_var": 0.9215792903344925,
      "eval_wth_loss": 0.006330131607865872,
      "step": 10240
    },
    {
      "epoch": 0.472957369174634,
      "eval_acr_loss": 0.5550906222024464,
      "eval_across_var": 0.2559552679867505,
      "eval_bleu": 0.9864228314384785,
      "eval_ce_loss": 0.03463799520246855,
      "eval_cos_loss": 0.2463943555815035,
      "eval_cov": 0.10881105187821062,
      "eval_cov_loss": 0.023118084696329894,
      "eval_glb_loss": 0.031056445483203347,
      "eval_global_kurtosis": 3.0601126701320145,
      "eval_global_mean": 0.08328581727258691,
      "eval_global_var": 1.173308629423516,
      "eval_krt_loss": 0.0071202065942054926,
      "eval_loss": 0.17534998668247162,
      "eval_mse_loss": 0.7060943077144013,
      "eval_per_loss": 0.007063817760171411,
      "eval_per_var": 1.0801762271689497,
      "eval_runtime": 149.1447,
      "eval_samples_per_second": 187.69,
      "eval_steps_per_second": 2.937,
      "eval_within_var": 0.9215792903344925,
      "eval_wth_loss": 0.006330131607865872,
      "step": 10240
    },
    {
      "epoch": 0.48478130340399983,
      "grad_norm": 0.0960695669054985,
      "learning_rate": 2.7454964417266016e-05,
      "loss": 0.2243,
      "step": 10496
    },
    {
      "epoch": 0.49660523763336567,
      "grad_norm": 0.09179438650608063,
      "learning_rate": 2.6504365436879463e-05,
      "loss": 0.2201,
      "step": 10752
    },
    {
      "epoch": 0.5084291718627315,
      "grad_norm": 0.12633387744426727,
      "learning_rate": 2.5547838773922217e-05,
      "loss": 0.2189,
      "step": 11008
    },
    {
      "epoch": 0.5202531060920974,
      "grad_norm": 0.10988616943359375,
      "learning_rate": 2.4590508770921138e-05,
      "loss": 0.2148,
      "step": 11264
    },
    {
      "epoch": 0.5202531060920974,
      "eval_acr_loss": 0.5188274796558842,
      "eval_across_var": 0.2808882309353515,
      "eval_bleu": 0.9881634510768484,
      "eval_ce_loss": 0.029888733919522806,
      "eval_cos_loss": 0.2354265242202641,
      "eval_cov": 0.1133959347798944,
      "eval_cov_loss": 0.025275619128574383,
      "eval_glb_loss": 0.03796970898702264,
      "eval_global_kurtosis": 3.0807342012178953,
      "eval_global_mean": 0.09449117042158292,
      "eval_global_var": 1.1918365617865296,
      "eval_krt_loss": 0.010437782962688104,
      "eval_loss": 0.16559295880195757,
      "eval_mse_loss": 0.6934955275222047,
      "eval_per_loss": 0.008229013991682497,
      "eval_per_var": 1.0867022509988584,
      "eval_within_var": 0.9157081242293528,
      "eval_wth_loss": 0.007290151238135279,
      "step": 11264
    },
    {
      "epoch": 0.5202531060920974,
      "eval_acr_loss": 0.5188274796558842,
      "eval_across_var": 0.2808882309353515,
      "eval_bleu": 0.9881634510768484,
      "eval_ce_loss": 0.029888733919522806,
      "eval_cos_loss": 0.2354265242202641,
      "eval_cov": 0.1133959347798944,
      "eval_cov_loss": 0.025275619128574383,
      "eval_glb_loss": 0.03796970898702264,
      "eval_global_kurtosis": 3.0807342012178953,
      "eval_global_mean": 0.09449117042158292,
      "eval_global_var": 1.1918365617865296,
      "eval_krt_loss": 0.010437782962688104,
      "eval_loss": 0.16559295880195757,
      "eval_mse_loss": 0.6934955275222047,
      "eval_per_loss": 0.008229013991682497,
      "eval_per_var": 1.0867022509988584,
      "eval_runtime": 150.6787,
      "eval_samples_per_second": 185.779,
      "eval_steps_per_second": 2.907,
      "eval_within_var": 0.9157081242293528,
      "eval_wth_loss": 0.007290151238135279,
      "step": 11264
    },
    {
      "epoch": 0.5320770403214632,
      "grad_norm": 0.10263644903898239,
      "learning_rate": 2.3633779237834874e-05,
      "loss": 0.2118,
      "step": 11520
    },
    {
      "epoch": 0.5439009745508291,
      "grad_norm": 0.0880010798573494,
      "learning_rate": 2.267905310410478e-05,
      "loss": 0.2097,
      "step": 11776
    },
    {
      "epoch": 0.5557249087801949,
      "grad_norm": 0.08736666291952133,
      "learning_rate": 2.172773036142663e-05,
      "loss": 0.2077,
      "step": 12032
    },
    {
      "epoch": 0.5675488430095608,
      "grad_norm": 0.08412963151931763,
      "learning_rate": 2.0781206010830228e-05,
      "loss": 0.2063,
      "step": 12288
    },
    {
      "epoch": 0.5675488430095608,
      "eval_acr_loss": 0.5011714086548923,
      "eval_across_var": 0.2933282603479956,
      "eval_bleu": 0.9892267845036895,
      "eval_ce_loss": 0.026746679018740647,
      "eval_cos_loss": 0.22736865680103432,
      "eval_cov": 0.11331831701269977,
      "eval_cov_loss": 0.02555328495502812,
      "eval_glb_loss": 0.03884035702709738,
      "eval_global_kurtosis": 3.0901281877195452,
      "eval_global_mean": 0.09504469671205842,
      "eval_global_var": 1.193941299229452,
      "eval_krt_loss": 0.012473824904427331,
      "eval_loss": 0.15896797551003766,
      "eval_mse_loss": 0.6858495211220224,
      "eval_per_loss": 0.007296311800882697,
      "eval_per_var": 1.081324468464612,
      "eval_within_var": 0.9059402656881776,
      "eval_wth_loss": 0.009029195172984374,
      "step": 12288
    },
    {
      "epoch": 0.5675488430095608,
      "eval_acr_loss": 0.5011714086548923,
      "eval_across_var": 0.2933282603479956,
      "eval_bleu": 0.9892267845036895,
      "eval_ce_loss": 0.026746679018740647,
      "eval_cos_loss": 0.22736865680103432,
      "eval_cov": 0.11331831701269977,
      "eval_cov_loss": 0.02555328495502812,
      "eval_glb_loss": 0.03884035702709738,
      "eval_global_kurtosis": 3.0901281877195452,
      "eval_global_mean": 0.09504469671205842,
      "eval_global_var": 1.193941299229452,
      "eval_krt_loss": 0.012473824904427331,
      "eval_loss": 0.15896797551003766,
      "eval_mse_loss": 0.6858495211220224,
      "eval_per_loss": 0.007296311800882697,
      "eval_per_var": 1.081324468464612,
      "eval_runtime": 150.6158,
      "eval_samples_per_second": 185.857,
      "eval_steps_per_second": 2.908,
      "eval_within_var": 0.9059402656881776,
      "eval_wth_loss": 0.009029195172984374,
      "step": 12288
    }
  ],
  "logging_steps": 256,
  "max_steps": 21651,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1024,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}