{
  "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": 1.1330132484436035,
      "learning_rate": 1.9615384615384617e-05,
      "loss": 10.3987,
      "step": 256
    },
    {
      "epoch": 0.023647868458731697,
      "grad_norm": 1.1089212894439697,
      "learning_rate": 3.930769230769231e-05,
      "loss": 7.8889,
      "step": 512
    },
    {
      "epoch": 0.03547180268809755,
      "grad_norm": 0.8628233075141907,
      "learning_rate": 4.999617095521894e-05,
      "loss": 5.5834,
      "step": 768
    },
    {
      "epoch": 0.047295736917463395,
      "grad_norm": 0.5602379441261292,
      "learning_rate": 4.9961092368776736e-05,
      "loss": 3.7587,
      "step": 1024
    },
    {
      "epoch": 0.047295736917463395,
      "eval_acr_loss": 0.472869769419165,
      "eval_across_var": 0.004283991528007299,
      "eval_bleu": 0.5621407857406262,
      "eval_ce_loss": 2.3296360490528962,
      "eval_cos_loss": 0.9237595144744333,
      "eval_cov": 0.07785117789490582,
      "eval_cov_loss": 0.009674696304907674,
      "eval_glb_loss": 0.3056357570979149,
      "eval_global_kurtosis": 3.016345693640513,
      "eval_global_mean": -0.0019390732730360336,
      "eval_global_var": 0.3470509150256849,
      "eval_krt_loss": 0.00041134808841301906,
      "eval_loss": 2.7735937942652944,
      "eval_mean_loss": 5.7365890913968915e-06,
      "eval_mse_loss": 1.926733713443965,
      "eval_per_loss": 0.3164680907715401,
      "eval_per_var": 0.337350505672089,
      "eval_within_var": 0.342682637784579,
      "eval_wth_loss": 0.310607642547725,
      "step": 1024
    },
    {
      "epoch": 0.047295736917463395,
      "eval_acr_loss": 0.472869769419165,
      "eval_across_var": 0.004283991528007299,
      "eval_bleu": 0.5621407857406262,
      "eval_ce_loss": 2.3296360490528962,
      "eval_cos_loss": 0.9237595144744333,
      "eval_cov": 0.07785117789490582,
      "eval_cov_loss": 0.009674696304907674,
      "eval_glb_loss": 0.3056357570979149,
      "eval_global_kurtosis": 3.016345693640513,
      "eval_global_mean": -0.0019390732730360336,
      "eval_global_var": 0.3470509150256849,
      "eval_krt_loss": 0.00041134808841301906,
      "eval_loss": 2.7735937942652944,
      "eval_mean_loss": 5.7365890913968915e-06,
      "eval_mse_loss": 1.926733713443965,
      "eval_per_loss": 0.3164680907715401,
      "eval_per_var": 0.337350505672089,
      "eval_runtime": 158.0332,
      "eval_samples_per_second": 177.134,
      "eval_steps_per_second": 2.772,
      "eval_within_var": 0.342682637784579,
      "eval_wth_loss": 0.310607642547725,
      "step": 1024
    },
    {
      "epoch": 0.05911967114682925,
      "grad_norm": 0.3792508840560913,
      "learning_rate": 4.988941132556799e-05,
      "loss": 2.6954,
      "step": 1280
    },
    {
      "epoch": 0.0709436053761951,
      "grad_norm": 0.30784618854522705,
      "learning_rate": 4.9781232937269974e-05,
      "loss": 2.0719,
      "step": 1536
    },
    {
      "epoch": 0.08276753960556095,
      "grad_norm": 0.24569711089134216,
      "learning_rate": 4.963671583455164e-05,
      "loss": 1.664,
      "step": 1792
    },
    {
      "epoch": 0.09459147383492679,
      "grad_norm": 0.20954662561416626,
      "learning_rate": 4.945607193446079e-05,
      "loss": 1.385,
      "step": 2048
    },
    {
      "epoch": 0.09459147383492679,
      "eval_acr_loss": 0.467126801201742,
      "eval_across_var": 0.014505732470706582,
      "eval_bleu": 0.8102335908319817,
      "eval_ce_loss": 0.6907969389331939,
      "eval_cos_loss": 0.7647272306217995,
      "eval_cov": 0.10773917420269692,
      "eval_cov_loss": 0.017931701802567805,
      "eval_glb_loss": 0.000625664478229841,
      "eval_global_kurtosis": 3.000208995113634,
      "eval_global_mean": -0.007508264829034675,
      "eval_global_var": 0.8768238085045662,
      "eval_krt_loss": 7.712448770311225e-05,
      "eval_loss": 1.0269616710268743,
      "eval_mean_loss": 6.29432441826079e-05,
      "eval_mse_loss": 1.6563664652985526,
      "eval_per_loss": 0.0019870867933301875,
      "eval_per_var": 0.8565188891267124,
      "eval_within_var": 0.8617800931407981,
      "eval_wth_loss": 0.0015375203581266992,
      "step": 2048
    },
    {
      "epoch": 0.09459147383492679,
      "eval_acr_loss": 0.467126801201742,
      "eval_across_var": 0.014505732470706582,
      "eval_bleu": 0.8102335908319817,
      "eval_ce_loss": 0.6907969389331939,
      "eval_cos_loss": 0.7647272306217995,
      "eval_cov": 0.10773917420269692,
      "eval_cov_loss": 0.017931701802567805,
      "eval_glb_loss": 0.000625664478229841,
      "eval_global_kurtosis": 3.000208995113634,
      "eval_global_mean": -0.007508264829034675,
      "eval_global_var": 0.8768238085045662,
      "eval_krt_loss": 7.712448770311225e-05,
      "eval_loss": 1.0269616710268743,
      "eval_mean_loss": 6.29432441826079e-05,
      "eval_mse_loss": 1.6563664652985526,
      "eval_per_loss": 0.0019870867933301875,
      "eval_per_var": 0.8565188891267124,
      "eval_runtime": 153.9643,
      "eval_samples_per_second": 181.815,
      "eval_steps_per_second": 2.845,
      "eval_within_var": 0.8617800931407981,
      "eval_wth_loss": 0.0015375203581266992,
      "step": 2048
    },
    {
      "epoch": 0.10641540806429264,
      "grad_norm": 0.181256964802742,
      "learning_rate": 4.923956612967301e-05,
      "loss": 1.1836,
      "step": 2304
    },
    {
      "epoch": 0.1182393422936585,
      "grad_norm": 0.1728077083826065,
      "learning_rate": 4.898751590005826e-05,
      "loss": 1.024,
      "step": 2560
    },
    {
      "epoch": 0.13006327652302435,
      "grad_norm": 0.14891040325164795,
      "learning_rate": 4.870029084713462e-05,
      "loss": 0.9004,
      "step": 2816
    },
    {
      "epoch": 0.1418872107523902,
      "grad_norm": 0.12918385863304138,
      "learning_rate": 4.837831215209188e-05,
      "loss": 0.8021,
      "step": 3072
    },
    {
      "epoch": 0.1418872107523902,
      "eval_acr_loss": 0.46483140122400574,
      "eval_across_var": 0.01694616893323385,
      "eval_bleu": 0.8960603905933928,
      "eval_ce_loss": 0.32335185586181403,
      "eval_cos_loss": 0.6106928648197487,
      "eval_cov": 0.09281384890482305,
      "eval_cov_loss": 0.013490723856388841,
      "eval_glb_loss": 0.0,
      "eval_global_kurtosis": 3.000047793127086,
      "eval_global_mean": -0.0062043639622866836,
      "eval_global_var": 0.9301077786101598,
      "eval_krt_loss": 0.0001482390632322396,
      "eval_loss": 0.5985009211655621,
      "eval_mean_loss": 4.6196501266608506e-05,
      "eval_mse_loss": 1.3684135288952692,
      "eval_per_loss": 6.070195510882551e-06,
      "eval_per_var": 0.906668049015411,
      "eval_within_var": 0.9128873076068756,
      "eval_wth_loss": 7.710680918438013e-07,
      "step": 3072
    },
    {
      "epoch": 0.1418872107523902,
      "eval_acr_loss": 0.46483140122400574,
      "eval_across_var": 0.01694616893323385,
      "eval_bleu": 0.8960603905933928,
      "eval_ce_loss": 0.32335185586181403,
      "eval_cos_loss": 0.6106928648197487,
      "eval_cov": 0.09281384890482305,
      "eval_cov_loss": 0.013490723856388841,
      "eval_glb_loss": 0.0,
      "eval_global_kurtosis": 3.000047793127086,
      "eval_global_mean": -0.0062043639622866836,
      "eval_global_var": 0.9301077786101598,
      "eval_krt_loss": 0.0001482390632322396,
      "eval_loss": 0.5985009211655621,
      "eval_mean_loss": 4.6196501266608506e-05,
      "eval_mse_loss": 1.3684135288952692,
      "eval_per_loss": 6.070195510882551e-06,
      "eval_per_var": 0.906668049015411,
      "eval_runtime": 154.7557,
      "eval_samples_per_second": 180.885,
      "eval_steps_per_second": 2.83,
      "eval_within_var": 0.9128873076068756,
      "eval_wth_loss": 7.710680918438013e-07,
      "step": 3072
    },
    {
      "epoch": 0.15371114498175603,
      "grad_norm": 0.12574972212314606,
      "learning_rate": 4.802205195817963e-05,
      "loss": 0.7236,
      "step": 3328
    },
    {
      "epoch": 0.1655350792111219,
      "grad_norm": 0.12267891317605972,
      "learning_rate": 4.763203267836576e-05,
      "loss": 0.656,
      "step": 3584
    },
    {
      "epoch": 0.17735901344048774,
      "grad_norm": 0.10911328345537186,
      "learning_rate": 4.720882622928019e-05,
      "loss": 0.6002,
      "step": 3840
    },
    {
      "epoch": 0.18918294766985358,
      "grad_norm": 0.10669343918561935,
      "learning_rate": 4.675305319256765e-05,
      "loss": 0.5541,
      "step": 4096
    },
    {
      "epoch": 0.18918294766985358,
      "eval_acr_loss": 0.461427516861049,
      "eval_across_var": 0.02085038360216778,
      "eval_bleu": 0.9355169660485144,
      "eval_ce_loss": 0.18811083091720598,
      "eval_cos_loss": 0.4924817247886092,
      "eval_cov": 0.08738018715218322,
      "eval_cov_loss": 0.011900046778647186,
      "eval_glb_loss": 0.0,
      "eval_global_kurtosis": 3.0061902341232996,
      "eval_global_mean": -0.007297369579201964,
      "eval_global_var": 1.0064546767979452,
      "eval_krt_loss": 0.00036125986233864483,
      "eval_loss": 0.41625422633946213,
      "eval_mean_loss": 6.224468990837522e-05,
      "eval_mse_loss": 1.1413913132393196,
      "eval_per_loss": 0.0,
      "eval_per_var": 0.9814564604737442,
      "eval_within_var": 0.9855326701244808,
      "eval_wth_loss": 0.0,
      "step": 4096
    },
    {
      "epoch": 0.18918294766985358,
      "eval_acr_loss": 0.461427516861049,
      "eval_across_var": 0.02085038360216778,
      "eval_bleu": 0.9355169660485144,
      "eval_ce_loss": 0.18811083091720598,
      "eval_cos_loss": 0.4924817247886092,
      "eval_cov": 0.08738018715218322,
      "eval_cov_loss": 0.011900046778647186,
      "eval_glb_loss": 0.0,
      "eval_global_kurtosis": 3.0061902341232996,
      "eval_global_mean": -0.007297369579201964,
      "eval_global_var": 1.0064546767979452,
      "eval_krt_loss": 0.00036125986233864483,
      "eval_loss": 0.41625422633946213,
      "eval_mean_loss": 6.224468990837522e-05,
      "eval_mse_loss": 1.1413913132393196,
      "eval_per_loss": 0.0,
      "eval_per_var": 0.9814564604737442,
      "eval_runtime": 155.5272,
      "eval_samples_per_second": 179.988,
      "eval_steps_per_second": 2.816,
      "eval_within_var": 0.9855326701244808,
      "eval_wth_loss": 0.0,
      "step": 4096
    },
    {
      "epoch": 0.20100688189921945,
      "grad_norm": 0.10487012565135956,
      "learning_rate": 4.6265381904878854e-05,
      "loss": 0.5121,
      "step": 4352
    },
    {
      "epoch": 0.2128308161285853,
      "grad_norm": 0.09705886989831924,
      "learning_rate": 4.57465274778347e-05,
      "loss": 0.4805,
      "step": 4608
    },
    {
      "epoch": 0.22465475035795113,
      "grad_norm": 0.10024245083332062,
      "learning_rate": 4.519725074940068e-05,
      "loss": 0.4469,
      "step": 4864
    },
    {
      "epoch": 0.236478684587317,
      "grad_norm": 0.08834318816661835,
      "learning_rate": 4.461835716820895e-05,
      "loss": 0.42,
      "step": 5120
    },
    {
      "epoch": 0.236478684587317,
      "eval_acr_loss": 0.4548332993978779,
      "eval_across_var": 0.028460351478087304,
      "eval_bleu": 0.9552258623563333,
      "eval_ce_loss": 0.12340653128071463,
      "eval_cos_loss": 0.40689416910143206,
      "eval_cov": 0.08788992825164098,
      "eval_cov_loss": 0.01204854997563852,
      "eval_glb_loss": 0.0002671638515636362,
      "eval_global_kurtosis": 2.9975850026901454,
      "eval_global_mean": -0.006530939716182343,
      "eval_global_var": 1.1122132741152968,
      "eval_krt_loss": 0.0002464600221175453,
      "eval_loss": 0.3173055400451024,
      "eval_mean_loss": 5.356442062072988e-05,
      "eval_mse_loss": 0.9764057996610528,
      "eval_per_loss": 5.57458498690526e-06,
      "eval_per_var": 1.0869274400684932,
      "eval_within_var": 1.0837734279022913,
      "eval_wth_loss": 5.189207881040453e-06,
      "step": 5120
    },
    {
      "epoch": 0.236478684587317,
      "eval_acr_loss": 0.4548332993978779,
      "eval_across_var": 0.028460351478087304,
      "eval_bleu": 0.9552258623563333,
      "eval_ce_loss": 0.12340653128071463,
      "eval_cos_loss": 0.40689416910143206,
      "eval_cov": 0.08788992825164098,
      "eval_cov_loss": 0.01204854997563852,
      "eval_glb_loss": 0.0002671638515636362,
      "eval_global_kurtosis": 2.9975850026901454,
      "eval_global_mean": -0.006530939716182343,
      "eval_global_var": 1.1122132741152968,
      "eval_krt_loss": 0.0002464600221175453,
      "eval_loss": 0.3173055400451024,
      "eval_mean_loss": 5.356442062072988e-05,
      "eval_mse_loss": 0.9764057996610528,
      "eval_per_loss": 5.57458498690526e-06,
      "eval_per_var": 1.0869274400684932,
      "eval_runtime": 154.1747,
      "eval_samples_per_second": 181.567,
      "eval_steps_per_second": 2.841,
      "eval_within_var": 1.0837734279022913,
      "eval_wth_loss": 5.189207881040453e-06,
      "step": 5120
    },
    {
      "epoch": 0.24830261881668284,
      "grad_norm": 0.08900712430477142,
      "learning_rate": 4.401069561246422e-05,
      "loss": 0.3965,
      "step": 5376
    },
    {
      "epoch": 0.2601265530460487,
      "grad_norm": 0.08696399629116058,
      "learning_rate": 4.337515714516545e-05,
      "loss": 0.3747,
      "step": 5632
    },
    {
      "epoch": 0.27195048727541454,
      "grad_norm": 0.09633397310972214,
      "learning_rate": 4.2712673707468434e-05,
      "loss": 0.3594,
      "step": 5888
    },
    {
      "epoch": 0.2837744215047804,
      "grad_norm": 0.08299960196018219,
      "learning_rate": 4.202421675210565e-05,
      "loss": 0.3425,
      "step": 6144
    },
    {
      "epoch": 0.2837744215047804,
      "eval_acr_loss": 0.444739139651599,
      "eval_across_var": 0.037298334392476575,
      "eval_bleu": 0.9670970291714691,
      "eval_ce_loss": 0.08759757838122649,
      "eval_cos_loss": 0.34691714662122947,
      "eval_cov": 0.08670684953802797,
      "eval_cov_loss": 0.011868421341288307,
      "eval_glb_loss": 0.0008387476386425913,
      "eval_global_kurtosis": 2.998767622529644,
      "eval_global_mean": -0.007462886234396669,
      "eval_global_var": 1.1261125677796804,
      "eval_krt_loss": 0.00038161360164841305,
      "eval_loss": 0.2573624641111452,
      "eval_mean_loss": 6.83625881451915e-05,
      "eval_mse_loss": 0.8629201449215684,
      "eval_per_loss": 0.00010140475162446005,
      "eval_per_var": 1.1030652468607305,
      "eval_within_var": 1.0889683330440085,
      "eval_wth_loss": 2.5263947968048936e-05,
      "step": 6144
    },
    {
      "epoch": 0.2837744215047804,
      "eval_acr_loss": 0.444739139651599,
      "eval_across_var": 0.037298334392476575,
      "eval_bleu": 0.9670970291714691,
      "eval_ce_loss": 0.08759757838122649,
      "eval_cos_loss": 0.34691714662122947,
      "eval_cov": 0.08670684953802797,
      "eval_cov_loss": 0.011868421341288307,
      "eval_glb_loss": 0.0008387476386425913,
      "eval_global_kurtosis": 2.998767622529644,
      "eval_global_mean": -0.007462886234396669,
      "eval_global_var": 1.1261125677796804,
      "eval_krt_loss": 0.00038161360164841305,
      "eval_loss": 0.2573624641111452,
      "eval_mean_loss": 6.83625881451915e-05,
      "eval_mse_loss": 0.8629201449215684,
      "eval_per_loss": 0.00010140475162446005,
      "eval_per_var": 1.1030652468607305,
      "eval_runtime": 153.8249,
      "eval_samples_per_second": 181.98,
      "eval_steps_per_second": 2.847,
      "eval_within_var": 1.0889683330440085,
      "eval_wth_loss": 2.5263947968048936e-05,
      "step": 6144
    },
    {
      "epoch": 0.2955983557341462,
      "grad_norm": 0.08428088575601578,
      "learning_rate": 4.131079581886694e-05,
      "loss": 0.325,
      "step": 6400
    },
    {
      "epoch": 0.30742228996351206,
      "grad_norm": 0.08620952814817429,
      "learning_rate": 4.057345705423016e-05,
      "loss": 0.3121,
      "step": 6656
    },
    {
      "epoch": 0.3192462241928779,
      "grad_norm": 0.08939851075410843,
      "learning_rate": 3.981328167731251e-05,
      "loss": 0.3016,
      "step": 6912
    },
    {
      "epoch": 0.3310701584222438,
      "grad_norm": 0.08685296773910522,
      "learning_rate": 3.9031384394391954e-05,
      "loss": 0.2875,
      "step": 7168
    },
    {
      "epoch": 0.3310701584222438,
      "eval_acr_loss": 0.4308900043164214,
      "eval_across_var": 0.04932180811230042,
      "eval_bleu": 0.9757814026870458,
      "eval_ce_loss": 0.0650875303026748,
      "eval_cos_loss": 0.3049790498600703,
      "eval_cov": 0.0888557608269121,
      "eval_cov_loss": 0.012648627446904847,
      "eval_glb_loss": 0.0011189478131933299,
      "eval_global_kurtosis": 3.0107862796957634,
      "eval_global_mean": -0.007322731763804885,
      "eval_global_var": 1.130718910530822,
      "eval_krt_loss": 0.0008007638278780262,
      "eval_loss": 0.217820766683855,
      "eval_mean_loss": 6.851153396431181e-05,
      "eval_mse_loss": 0.7866716142658774,
      "eval_per_loss": 0.0001979704842660462,
      "eval_per_var": 1.1079815032819635,
      "eval_within_var": 1.0816122696279935,
      "eval_wth_loss": 1.0462231934273432e-05,
      "step": 7168
    },
    {
      "epoch": 0.3310701584222438,
      "eval_acr_loss": 0.4308900043164214,
      "eval_across_var": 0.04932180811230042,
      "eval_bleu": 0.9757814026870458,
      "eval_ce_loss": 0.0650875303026748,
      "eval_cos_loss": 0.3049790498600703,
      "eval_cov": 0.0888557608269121,
      "eval_cov_loss": 0.012648627446904847,
      "eval_glb_loss": 0.0011189478131933299,
      "eval_global_kurtosis": 3.0107862796957634,
      "eval_global_mean": -0.007322731763804885,
      "eval_global_var": 1.130718910530822,
      "eval_krt_loss": 0.0008007638278780262,
      "eval_loss": 0.217820766683855,
      "eval_mean_loss": 6.851153396431181e-05,
      "eval_mse_loss": 0.7866716142658774,
      "eval_per_loss": 0.0001979704842660462,
      "eval_per_var": 1.1079815032819635,
      "eval_runtime": 151.7662,
      "eval_samples_per_second": 184.448,
      "eval_steps_per_second": 2.886,
      "eval_within_var": 1.0816122696279935,
      "eval_wth_loss": 1.0462231934273432e-05,
      "step": 7168
    },
    {
      "epoch": 0.34289409265160964,
      "grad_norm": 0.08307478576898575,
      "learning_rate": 3.822891176432382e-05,
      "loss": 0.2798,
      "step": 7424
    },
    {
      "epoch": 0.3547180268809755,
      "grad_norm": 0.07975339144468307,
      "learning_rate": 3.7407040517249335e-05,
      "loss": 0.2702,
      "step": 7680
    },
    {
      "epoch": 0.3665419611103413,
      "grad_norm": 0.08212888985872269,
      "learning_rate": 3.6566975829061614e-05,
      "loss": 0.2607,
      "step": 7936
    },
    {
      "epoch": 0.37836589533970716,
      "grad_norm": 0.07644044607877731,
      "learning_rate": 3.5709949554159355e-05,
      "loss": 0.2519,
      "step": 8192
    },
    {
      "epoch": 0.37836589533970716,
      "eval_acr_loss": 0.41207800308863324,
      "eval_across_var": 0.0661766531167033,
      "eval_bleu": 0.9805728687340939,
      "eval_ce_loss": 0.05073687865411582,
      "eval_cos_loss": 0.2765354519905565,
      "eval_cov": 0.09245990073844178,
      "eval_cov_loss": 0.013950884125310264,
      "eval_glb_loss": 0.0015832305986751307,
      "eval_global_kurtosis": 3.0033102030079113,
      "eval_global_mean": -0.008325645232309489,
      "eval_global_var": 1.1371356842180365,
      "eval_krt_loss": 0.0005490112146340372,
      "eval_loss": 0.19166131559164012,
      "eval_mean_loss": 8.809174024773519e-05,
      "eval_mse_loss": 0.7389907115670644,
      "eval_per_loss": 0.00044034952490991304,
      "eval_per_var": 1.1152143086472603,
      "eval_within_var": 1.0713012561406174,
      "eval_wth_loss": 1.391253488354359e-06,
      "step": 8192
    },
    {
      "epoch": 0.37836589533970716,
      "eval_acr_loss": 0.41207800308863324,
      "eval_across_var": 0.0661766531167033,
      "eval_bleu": 0.9805728687340939,
      "eval_ce_loss": 0.05073687865411582,
      "eval_cos_loss": 0.2765354519905565,
      "eval_cov": 0.09245990073844178,
      "eval_cov_loss": 0.013950884125310264,
      "eval_glb_loss": 0.0015832305986751307,
      "eval_global_kurtosis": 3.0033102030079113,
      "eval_global_mean": -0.008325645232309489,
      "eval_global_var": 1.1371356842180365,
      "eval_krt_loss": 0.0005490112146340372,
      "eval_loss": 0.19166131559164012,
      "eval_mean_loss": 8.809174024773519e-05,
      "eval_mse_loss": 0.7389907115670644,
      "eval_per_loss": 0.00044034952490991304,
      "eval_per_var": 1.1152143086472603,
      "eval_runtime": 151.0158,
      "eval_samples_per_second": 185.365,
      "eval_steps_per_second": 2.9,
      "eval_within_var": 1.0713012561406174,
      "eval_wth_loss": 1.391253488354359e-06,
      "step": 8192
    },
    {
      "epoch": 0.390189829569073,
      "grad_norm": 0.07752202451229095,
      "learning_rate": 3.483721841907964e-05,
      "loss": 0.2462,
      "step": 8448
    },
    {
      "epoch": 0.4020137637984389,
      "grad_norm": 0.08402163535356522,
      "learning_rate": 3.3953554020219734e-05,
      "loss": 0.2397,
      "step": 8704
    },
    {
      "epoch": 0.41383769802780473,
      "grad_norm": 0.08829676359891891,
      "learning_rate": 3.305332229492516e-05,
      "loss": 0.2323,
      "step": 8960
    },
    {
      "epoch": 0.4256616322571706,
      "grad_norm": 0.07489398866891861,
      "learning_rate": 3.214128133561262e-05,
      "loss": 0.2275,
      "step": 9216
    },
    {
      "epoch": 0.4256616322571706,
      "eval_acr_loss": 0.3899558907094067,
      "eval_across_var": 0.08639241476037186,
      "eval_bleu": 0.9839023939205267,
      "eval_ce_loss": 0.04116568632730066,
      "eval_cos_loss": 0.25615145446366916,
      "eval_cov": 0.09534583244149543,
      "eval_cov_loss": 0.015230727917833687,
      "eval_glb_loss": 0.001839591442792664,
      "eval_global_kurtosis": 3.0162572811727655,
      "eval_global_mean": -0.008430217362974333,
      "eval_global_var": 1.1393898045091324,
      "eval_krt_loss": 0.001037847468692951,
      "eval_loss": 0.17338106053196675,
      "eval_mean_loss": 9.661306929678656e-05,
      "eval_mse_loss": 0.7077833269829075,
      "eval_per_loss": 0.0006317324198246723,
      "eval_per_var": 1.1172164847317352,
      "eval_within_var": 1.053544073616533,
      "eval_wth_loss": 0.0,
      "step": 9216
    },
    {
      "epoch": 0.4256616322571706,
      "eval_acr_loss": 0.3899558907094067,
      "eval_across_var": 0.08639241476037186,
      "eval_bleu": 0.9839023939205267,
      "eval_ce_loss": 0.04116568632730066,
      "eval_cos_loss": 0.25615145446366916,
      "eval_cov": 0.09534583244149543,
      "eval_cov_loss": 0.015230727917833687,
      "eval_glb_loss": 0.001839591442792664,
      "eval_global_kurtosis": 3.0162572811727655,
      "eval_global_mean": -0.008430217362974333,
      "eval_global_var": 1.1393898045091324,
      "eval_krt_loss": 0.001037847468692951,
      "eval_loss": 0.17338106053196675,
      "eval_mean_loss": 9.661306929678656e-05,
      "eval_mse_loss": 0.7077833269829075,
      "eval_per_loss": 0.0006317324198246723,
      "eval_per_var": 1.1172164847317352,
      "eval_runtime": 149.7127,
      "eval_samples_per_second": 186.978,
      "eval_steps_per_second": 2.926,
      "eval_within_var": 1.053544073616533,
      "eval_wth_loss": 0.0,
      "step": 9216
    },
    {
      "epoch": 0.4374855664865364,
      "grad_norm": 0.07958604395389557,
      "learning_rate": 3.1218768541274476e-05,
      "loss": 0.2225,
      "step": 9472
    },
    {
      "epoch": 0.44930950071590225,
      "grad_norm": 0.08125226199626923,
      "learning_rate": 3.028713666659683e-05,
      "loss": 0.219,
      "step": 9728
    },
    {
      "epoch": 0.4611334349452681,
      "grad_norm": 0.08044584840536118,
      "learning_rate": 2.9347751838306454e-05,
      "loss": 0.2136,
      "step": 9984
    },
    {
      "epoch": 0.472957369174634,
      "grad_norm": 0.09431233257055283,
      "learning_rate": 2.840199155190943e-05,
      "loss": 0.2109,
      "step": 10240
    },
    {
      "epoch": 0.472957369174634,
      "eval_acr_loss": 0.371521062094327,
      "eval_across_var": 0.10420215363746092,
      "eval_bleu": 0.9864977390539889,
      "eval_ce_loss": 0.03421062290813076,
      "eval_cos_loss": 0.24110814265600622,
      "eval_cov": 0.09938139893692922,
      "eval_cov_loss": 0.016782419545550462,
      "eval_glb_loss": 0.0025077612427728364,
      "eval_global_kurtosis": 3.000300804229632,
      "eval_global_mean": -0.008482733817949686,
      "eval_global_var": 1.1459448130707763,
      "eval_krt_loss": 0.0005616036675252787,
      "eval_loss": 0.16001523779407484,
      "eval_mean_loss": 0.00010505710818644507,
      "eval_mse_loss": 0.6870357463621113,
      "eval_per_loss": 0.0010411720005544605,
      "eval_per_var": 1.1233344927226028,
      "eval_within_var": 1.042447058849683,
      "eval_wth_loss": 0.0,
      "step": 10240
    },
    {
      "epoch": 0.472957369174634,
      "eval_acr_loss": 0.371521062094327,
      "eval_across_var": 0.10420215363746092,
      "eval_bleu": 0.9864977390539889,
      "eval_ce_loss": 0.03421062290813076,
      "eval_cos_loss": 0.24110814265600622,
      "eval_cov": 0.09938139893692922,
      "eval_cov_loss": 0.016782419545550462,
      "eval_glb_loss": 0.0025077612427728364,
      "eval_global_kurtosis": 3.000300804229632,
      "eval_global_mean": -0.008482733817949686,
      "eval_global_var": 1.1459448130707763,
      "eval_krt_loss": 0.0005616036675252787,
      "eval_loss": 0.16001523779407484,
      "eval_mean_loss": 0.00010505710818644507,
      "eval_mse_loss": 0.6870357463621113,
      "eval_per_loss": 0.0010411720005544605,
      "eval_per_var": 1.1233344927226028,
      "eval_runtime": 150.7627,
      "eval_samples_per_second": 185.676,
      "eval_steps_per_second": 2.905,
      "eval_within_var": 1.042447058849683,
      "eval_wth_loss": 0.0,
      "step": 10240
    },
    {
      "epoch": 0.48478130340399983,
      "grad_norm": 0.07165589183568954,
      "learning_rate": 2.745124265175868e-05,
      "loss": 0.2062,
      "step": 10496
    },
    {
      "epoch": 0.49660523763336567,
      "grad_norm": 0.07731898874044418,
      "learning_rate": 2.6496899297412598e-05,
      "loss": 0.2023,
      "step": 10752
    },
    {
      "epoch": 0.5084291718627315,
      "grad_norm": 0.08252639323472977,
      "learning_rate": 2.5544099852682395e-05,
      "loss": 0.2011,
      "step": 11008
    },
    {
      "epoch": 0.5202531060920974,
      "grad_norm": 0.09438033401966095,
      "learning_rate": 2.4586769464065254e-05,
      "loss": 0.1971,
      "step": 11264
    },
    {
      "epoch": 0.5202531060920974,
      "eval_acr_loss": 0.3526782588735563,
      "eval_across_var": 0.1233543821240533,
      "eval_bleu": 0.9885170657049208,
      "eval_ce_loss": 0.02941183759256789,
      "eval_cos_loss": 0.2296683646432341,
      "eval_cov": 0.10448926115689212,
      "eval_cov_loss": 0.018818640007667208,
      "eval_glb_loss": 0.0035778092502643057,
      "eval_global_kurtosis": 3.010114621898355,
      "eval_global_mean": -0.011101974472063318,
      "eval_global_var": 1.1552310751997716,
      "eval_krt_loss": 0.0011141615731220227,
      "eval_loss": 0.15028831549839342,
      "eval_mean_loss": 0.0001692649679361995,
      "eval_mse_loss": 0.6726388316176254,
      "eval_per_loss": 0.0018812230319723378,
      "eval_per_var": 1.1340878281963471,
      "eval_within_var": 1.0327191478041209,
      "eval_wth_loss": 0.0,
      "step": 11264
    },
    {
      "epoch": 0.5202531060920974,
      "eval_acr_loss": 0.3526782588735563,
      "eval_across_var": 0.1233543821240533,
      "eval_bleu": 0.9885170657049208,
      "eval_ce_loss": 0.02941183759256789,
      "eval_cos_loss": 0.2296683646432341,
      "eval_cov": 0.10448926115689212,
      "eval_cov_loss": 0.018818640007667208,
      "eval_glb_loss": 0.0035778092502643057,
      "eval_global_kurtosis": 3.010114621898355,
      "eval_global_mean": -0.011101974472063318,
      "eval_global_var": 1.1552310751997716,
      "eval_krt_loss": 0.0011141615731220227,
      "eval_loss": 0.15028831549839342,
      "eval_mean_loss": 0.0001692649679361995,
      "eval_mse_loss": 0.6726388316176254,
      "eval_per_loss": 0.0018812230319723378,
      "eval_per_var": 1.1340878281963471,
      "eval_runtime": 150.1371,
      "eval_samples_per_second": 186.45,
      "eval_steps_per_second": 2.917,
      "eval_within_var": 1.0327191478041209,
      "eval_wth_loss": 0.0,
      "step": 11264
    },
    {
      "epoch": 0.5320770403214632,
      "grad_norm": 0.08242896944284439,
      "learning_rate": 2.3630045028609248e-05,
      "loss": 0.1941,
      "step": 11520
    },
    {
      "epoch": 0.5439009745508291,
      "grad_norm": 0.12015263736248016,
      "learning_rate": 2.267532946828065e-05,
      "loss": 0.1923,
      "step": 11776
    },
    {
      "epoch": 0.5557249087801949,
      "grad_norm": 0.07551924884319305,
      "learning_rate": 2.1724022759270597e-05,
      "loss": 0.19,
      "step": 12032
    },
    {
      "epoch": 0.5675488430095608,
      "grad_norm": 0.09451667219400406,
      "learning_rate": 2.0777519879097458e-05,
      "loss": 0.1889,
      "step": 12288
    },
    {
      "epoch": 0.5675488430095608,
      "eval_acr_loss": 0.34067982850281614,
      "eval_across_var": 0.1346761514964305,
      "eval_bleu": 0.9895254511324281,
      "eval_ce_loss": 0.0263222230802456,
      "eval_cos_loss": 0.22149913948556604,
      "eval_cov": 0.10440648745184075,
      "eval_cov_loss": 0.019090854430987955,
      "eval_glb_loss": 0.0031285923291081673,
      "eval_global_kurtosis": 3.0161606329216806,
      "eval_global_mean": -0.01319564518318873,
      "eval_global_var": 1.1500695633561644,
      "eval_krt_loss": 0.0014608212088625325,
      "eval_loss": 0.14371125602490825,
      "eval_mean_loss": 0.0002307435466704484,
      "eval_mse_loss": 0.6642545581408287,
      "eval_per_loss": 0.00148715251082838,
      "eval_per_var": 1.1267546910673516,
      "eval_within_var": 1.0164183696655378,
      "eval_wth_loss": 0.0,
      "step": 12288
    },
    {
      "epoch": 0.5675488430095608,
      "eval_acr_loss": 0.34067982850281614,
      "eval_across_var": 0.1346761514964305,
      "eval_bleu": 0.9895254511324281,
      "eval_ce_loss": 0.0263222230802456,
      "eval_cos_loss": 0.22149913948556604,
      "eval_cov": 0.10440648745184075,
      "eval_cov_loss": 0.019090854430987955,
      "eval_glb_loss": 0.0031285923291081673,
      "eval_global_kurtosis": 3.0161606329216806,
      "eval_global_mean": -0.01319564518318873,
      "eval_global_var": 1.1500695633561644,
      "eval_krt_loss": 0.0014608212088625325,
      "eval_loss": 0.14371125602490825,
      "eval_mean_loss": 0.0002307435466704484,
      "eval_mse_loss": 0.6642545581408287,
      "eval_per_loss": 0.00148715251082838,
      "eval_per_var": 1.1267546910673516,
      "eval_runtime": 149.6169,
      "eval_samples_per_second": 187.098,
      "eval_steps_per_second": 2.927,
      "eval_within_var": 1.0164183696655378,
      "eval_wth_loss": 0.0,
      "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
}