{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.44131273299286744,
  "eval_steps": 1024,
  "global_step": 10240,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.011032818324821687,
      "grad_norm": 0.20388168096542358,
      "learning_rate": 0.000498046875,
      "loss": 2.512310028076172,
      "step": 256
    },
    {
      "epoch": 0.022065636649643373,
      "grad_norm": 0.3460715115070343,
      "learning_rate": 0.000998046875,
      "loss": 2.018148422241211,
      "step": 512
    },
    {
      "epoch": 0.03309845497446506,
      "grad_norm": 0.5453425645828247,
      "learning_rate": 0.000999688448778502,
      "loss": 1.8114819526672363,
      "step": 768
    },
    {
      "epoch": 0.04413127329928675,
      "grad_norm": 0.9051710367202759,
      "learning_rate": 0.0009987492950653055,
      "loss": 1.75458824634552,
      "step": 1024
    },
    {
      "epoch": 0.04413127329928675,
      "eval_bleu": 0.9399749239907662,
      "eval_cos_loss": 0.47715059564566054,
      "eval_dec_loss": 0.10647435713885055,
      "eval_loss": 1.73655072051579,
      "eval_mse2_loss": 0.1675516524389863,
      "eval_mse_loss": 1.3505616757407117,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.522849405561683,
      "flow/improvement_ratio": 0.8910323953323527,
      "flow/mag_ratio_mean": 0.5448383128465112,
      "flow/mag_ratio_std": 0.23550461588510826,
      "step": 1024
    },
    {
      "epoch": 0.04413127329928675,
      "eval_bleu": 0.9399749239907662,
      "eval_cos_loss": 0.47715059564566054,
      "eval_dec_loss": 0.10647435713885055,
      "eval_loss": 1.73655072051579,
      "eval_mse2_loss": 0.1675516524389863,
      "eval_mse_loss": 1.3505616757407117,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 152.3582,
      "eval_samples_per_second": 196.904,
      "eval_steps_per_second": 3.078,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.522849405561683,
      "flow/improvement_ratio": 0.8910323953323527,
      "flow/mag_ratio_mean": 0.5448383128465112,
      "flow/mag_ratio_std": 0.23550461588510826,
      "step": 1024
    },
    {
      "epoch": 0.05516409162410843,
      "grad_norm": 0.5948837399482727,
      "learning_rate": 0.0009971837136430763,
      "loss": 1.732498049736023,
      "step": 1280
    },
    {
      "epoch": 0.06619690994893011,
      "grad_norm": 0.6182620525360107,
      "learning_rate": 0.0009949936708776692,
      "loss": 1.7030788660049438,
      "step": 1536
    },
    {
      "epoch": 0.07722972827375181,
      "grad_norm": 1.142866611480713,
      "learning_rate": 0.0009921819174566252,
      "loss": 1.7001720666885376,
      "step": 1792
    },
    {
      "epoch": 0.0882625465985735,
      "grad_norm": 0.8506317734718323,
      "learning_rate": 0.000988751984934317,
      "loss": 1.6855676174163818,
      "step": 2048
    },
    {
      "epoch": 0.0882625465985735,
      "eval_bleu": 0.9434877818736953,
      "eval_cos_loss": 0.46131489716613217,
      "eval_dec_loss": 0.09124524126659388,
      "eval_loss": 1.6677722422553023,
      "eval_mse2_loss": 0.15732173794812992,
      "eval_mse_loss": 1.3088257922800874,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5386851041683002,
      "flow/improvement_ratio": 0.8871048878250855,
      "flow/mag_ratio_mean": 0.5622577369848548,
      "flow/mag_ratio_std": 0.24945266208033573,
      "step": 2048
    },
    {
      "epoch": 0.0882625465985735,
      "eval_bleu": 0.9434877818736953,
      "eval_cos_loss": 0.46131489716613217,
      "eval_dec_loss": 0.09124524126659388,
      "eval_loss": 1.6677722422553023,
      "eval_mse2_loss": 0.15732173794812992,
      "eval_mse_loss": 1.3088257922800874,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 150.2408,
      "eval_samples_per_second": 199.68,
      "eval_steps_per_second": 3.122,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5386851041683002,
      "flow/improvement_ratio": 0.8871048878250855,
      "flow/mag_ratio_mean": 0.5622577369848548,
      "flow/mag_ratio_std": 0.24945266208033573,
      "step": 2048
    },
    {
      "epoch": 0.09929536492339518,
      "grad_norm": 0.7440093159675598,
      "learning_rate": 0.0009847081812963268,
      "loss": 1.6802997589111328,
      "step": 2304
    },
    {
      "epoch": 0.11032818324821686,
      "grad_norm": 0.9319222569465637,
      "learning_rate": 0.0009800555855486275,
      "loss": 1.6744197607040405,
      "step": 2560
    },
    {
      "epoch": 0.12136100157303854,
      "grad_norm": 0.8629500865936279,
      "learning_rate": 0.0009748000413383664,
      "loss": 1.6740639209747314,
      "step": 2816
    },
    {
      "epoch": 0.13239381989786023,
      "grad_norm": 0.9893732666969299,
      "learning_rate": 0.0009689481496142604,
      "loss": 1.664785623550415,
      "step": 3072
    },
    {
      "epoch": 0.13239381989786023,
      "eval_bleu": 0.9404068449586629,
      "eval_cos_loss": 0.4534218231243874,
      "eval_dec_loss": 0.10443712005824614,
      "eval_loss": 1.6572931651621738,
      "eval_mse2_loss": 0.1550013455850229,
      "eval_mse_loss": 1.288264540212749,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5465781726816824,
      "flow/improvement_ratio": 0.8946911343125138,
      "flow/mag_ratio_mean": 0.5628405101517878,
      "flow/mag_ratio_std": 0.24253392000315285,
      "step": 3072
    },
    {
      "epoch": 0.13239381989786023,
      "eval_bleu": 0.9404068449586629,
      "eval_cos_loss": 0.4534218231243874,
      "eval_dec_loss": 0.10443712005824614,
      "eval_loss": 1.6572931651621738,
      "eval_mse2_loss": 0.1550013455850229,
      "eval_mse_loss": 1.288264540212749,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 153.5602,
      "eval_samples_per_second": 195.363,
      "eval_steps_per_second": 3.054,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5465781726816824,
      "flow/improvement_ratio": 0.8946911343125138,
      "flow/mag_ratio_mean": 0.5628405101517878,
      "flow/mag_ratio_std": 0.24253392000315285,
      "step": 3072
    },
    {
      "epoch": 0.14342663822268192,
      "grad_norm": 0.9933224320411682,
      "learning_rate": 0.0009625072603358231,
      "loss": 1.6605451107025146,
      "step": 3328
    },
    {
      "epoch": 0.15445945654750362,
      "grad_norm": 1.221793532371521,
      "learning_rate": 0.0009554854632418371,
      "loss": 1.6490036249160767,
      "step": 3584
    },
    {
      "epoch": 0.1654922748723253,
      "grad_norm": 0.8394345045089722,
      "learning_rate": 0.000947891577689663,
      "loss": 1.649448275566101,
      "step": 3840
    },
    {
      "epoch": 0.176525093197147,
      "grad_norm": 1.245514154434204,
      "learning_rate": 0.0009397351415781539,
      "loss": 1.6489267349243164,
      "step": 4096
    },
    {
      "epoch": 0.176525093197147,
      "eval_bleu": 0.942338221101898,
      "eval_cos_loss": 0.450266804585833,
      "eval_dec_loss": 0.1002394597608048,
      "eval_loss": 1.6427197324187517,
      "eval_mse2_loss": 0.15268841918025697,
      "eval_mse_loss": 1.280517189741643,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5497331933172018,
      "flow/improvement_ratio": 0.8926444684010325,
      "flow/mag_ratio_mean": 0.5659083429175907,
      "flow/mag_ratio_std": 0.24314757854318314,
      "step": 4096
    },
    {
      "epoch": 0.176525093197147,
      "eval_bleu": 0.942338221101898,
      "eval_cos_loss": 0.450266804585833,
      "eval_dec_loss": 0.1002394597608048,
      "eval_loss": 1.6427197324187517,
      "eval_mse2_loss": 0.15268841918025697,
      "eval_mse_loss": 1.280517189741643,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 152.0027,
      "eval_samples_per_second": 197.365,
      "eval_steps_per_second": 3.085,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5497331933172018,
      "flow/improvement_ratio": 0.8926444684010325,
      "flow/mag_ratio_mean": 0.5659083429175907,
      "flow/mag_ratio_std": 0.24314757854318314,
      "step": 4096
    },
    {
      "epoch": 0.18755791152196866,
      "grad_norm": 1.0416312217712402,
      "learning_rate": 0.000931026399368079,
      "loss": 1.6447768211364746,
      "step": 4352
    },
    {
      "epoch": 0.19859072984679035,
      "grad_norm": 1.1173036098480225,
      "learning_rate": 0.0009217762892151117,
      "loss": 1.6489276885986328,
      "step": 4608
    },
    {
      "epoch": 0.20962354817161205,
      "grad_norm": 0.930402934551239,
      "learning_rate": 0.0009119964292315354,
      "loss": 1.6420283317565918,
      "step": 4864
    },
    {
      "epoch": 0.22065636649643372,
      "grad_norm": 0.9209682941436768,
      "learning_rate": 0.0009016991028939279,
      "loss": 1.6357425451278687,
      "step": 5120
    },
    {
      "epoch": 0.22065636649643372,
      "eval_bleu": 0.9431814239032791,
      "eval_cos_loss": 0.4463651056991203,
      "eval_dec_loss": 0.09822412853889755,
      "eval_loss": 1.628653102846288,
      "eval_mse2_loss": 0.14985301353529826,
      "eval_mse_loss": 1.2716914730539708,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5536348958259453,
      "flow/improvement_ratio": 0.8939987117293547,
      "flow/mag_ratio_mean": 0.576050937048662,
      "flow/mag_ratio_std": 0.2491114061397276,
      "step": 5120
    },
    {
      "epoch": 0.22065636649643372,
      "eval_bleu": 0.9431814239032791,
      "eval_cos_loss": 0.4463651056991203,
      "eval_dec_loss": 0.09822412853889755,
      "eval_loss": 1.628653102846288,
      "eval_mse2_loss": 0.14985301353529826,
      "eval_mse_loss": 1.2716914730539708,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 150.4792,
      "eval_samples_per_second": 199.363,
      "eval_steps_per_second": 3.117,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5536348958259453,
      "flow/improvement_ratio": 0.8939987117293547,
      "flow/mag_ratio_mean": 0.576050937048662,
      "flow/mag_ratio_std": 0.2491114061397276,
      "step": 5120
    },
    {
      "epoch": 0.23168918482125542,
      "grad_norm": 0.6372450590133667,
      "learning_rate": 0.0008908972436151494,
      "loss": 1.6375595331192017,
      "step": 5376
    },
    {
      "epoch": 0.2427220031460771,
      "grad_norm": 0.833997368812561,
      "learning_rate": 0.0008796044185000127,
      "loss": 1.6372840404510498,
      "step": 5632
    },
    {
      "epoch": 0.2537548214708988,
      "grad_norm": 0.6318811178207397,
      "learning_rate": 0.0008678348113050368,
      "loss": 1.628332257270813,
      "step": 5888
    },
    {
      "epoch": 0.26478763979572045,
      "grad_norm": 0.7464238405227661,
      "learning_rate": 0.0008556032046236897,
      "loss": 1.6342945098876953,
      "step": 6144
    },
    {
      "epoch": 0.26478763979572045,
      "eval_bleu": 0.941904927802284,
      "eval_cos_loss": 0.44474837520737637,
      "eval_dec_loss": 0.102102437767504,
      "eval_loss": 1.628202569764306,
      "eval_mse2_loss": 0.15077992649411343,
      "eval_mse_loss": 1.266597391954109,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5552516249197124,
      "flow/improvement_ratio": 0.8912770662988935,
      "flow/mag_ratio_mean": 0.5855364076364269,
      "flow/mag_ratio_std": 0.25087345015011364,
      "step": 6144
    },
    {
      "epoch": 0.26478763979572045,
      "eval_bleu": 0.941904927802284,
      "eval_cos_loss": 0.44474837520737637,
      "eval_dec_loss": 0.102102437767504,
      "eval_loss": 1.628202569764306,
      "eval_mse2_loss": 0.15077992649411343,
      "eval_mse_loss": 1.266597391954109,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 151.0104,
      "eval_samples_per_second": 198.662,
      "eval_steps_per_second": 3.106,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5552516249197124,
      "flow/improvement_ratio": 0.8912770662988935,
      "flow/mag_ratio_mean": 0.5855364076364269,
      "flow/mag_ratio_std": 0.25087345015011364,
      "step": 6144
    },
    {
      "epoch": 0.2758204581205422,
      "grad_norm": 1.1199694871902466,
      "learning_rate": 0.000842924961319492,
      "loss": 1.6266489028930664,
      "step": 6400
    },
    {
      "epoch": 0.28685327644536385,
      "grad_norm": 0.8668828010559082,
      "learning_rate": 0.0008298160052303045,
      "loss": 1.62454092502594,
      "step": 6656
    },
    {
      "epoch": 0.2978860947701855,
      "grad_norm": 0.8108460307121277,
      "learning_rate": 0.0008162928011680314,
      "loss": 1.624453067779541,
      "step": 6912
    },
    {
      "epoch": 0.30891891309500724,
      "grad_norm": 0.8465085625648499,
      "learning_rate": 0.000802372334238864,
      "loss": 1.6209194660186768,
      "step": 7168
    },
    {
      "epoch": 0.30891891309500724,
      "eval_bleu": 0.942268661485615,
      "eval_cos_loss": 0.4415781778186115,
      "eval_dec_loss": 0.09969639404813872,
      "eval_loss": 1.61412800298825,
      "eval_mse2_loss": 0.14795255090699774,
      "eval_mse_loss": 1.2580732640935415,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5584218213553114,
      "flow/improvement_ratio": 0.8916484548338949,
      "flow/mag_ratio_mean": 0.5767976182864419,
      "flow/mag_ratio_std": 0.2534670445011623,
      "step": 7168
    },
    {
      "epoch": 0.30891891309500724,
      "eval_bleu": 0.942268661485615,
      "eval_cos_loss": 0.4415781778186115,
      "eval_dec_loss": 0.09969639404813872,
      "eval_loss": 1.61412800298825,
      "eval_mse2_loss": 0.14795255090699774,
      "eval_mse_loss": 1.2580732640935415,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 150.313,
      "eval_samples_per_second": 199.584,
      "eval_steps_per_second": 3.12,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5584218213553114,
      "flow/improvement_ratio": 0.8916484548338949,
      "flow/mag_ratio_mean": 0.5767976182864419,
      "flow/mag_ratio_std": 0.2534670445011623,
      "step": 7168
    },
    {
      "epoch": 0.3199517314198289,
      "grad_norm": 1.8415089845657349,
      "learning_rate": 0.0007880720885100349,
      "loss": 1.6192532777786255,
      "step": 7424
    },
    {
      "epoch": 0.3309845497446506,
      "grad_norm": 0.7575666904449463,
      "learning_rate": 0.0007734100250498788,
      "loss": 1.6192028522491455,
      "step": 7680
    },
    {
      "epoch": 0.3420173680694723,
      "grad_norm": 1.108810544013977,
      "learning_rate": 0.000758404559368781,
      "loss": 1.614426851272583,
      "step": 7936
    },
    {
      "epoch": 0.353050186394294,
      "grad_norm": 1.224976897239685,
      "learning_rate": 0.0007430745382893488,
      "loss": 1.612691879272461,
      "step": 8192
    },
    {
      "epoch": 0.353050186394294,
      "eval_bleu": 0.9417854724917855,
      "eval_cos_loss": 0.4400326657905253,
      "eval_dec_loss": 0.1005224303852743,
      "eval_loss": 1.6120708153954446,
      "eval_mse2_loss": 0.14864133708258429,
      "eval_mse_loss": 1.254655804715431,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5599673331927643,
      "flow/improvement_ratio": 0.8927549306771903,
      "flow/mag_ratio_mean": 0.5825493686488951,
      "flow/mag_ratio_std": 0.25123427366651196,
      "step": 8192
    },
    {
      "epoch": 0.353050186394294,
      "eval_bleu": 0.9417854724917855,
      "eval_cos_loss": 0.4400326657905253,
      "eval_dec_loss": 0.1005224303852743,
      "eval_loss": 1.6120708153954446,
      "eval_mse2_loss": 0.14864133708258429,
      "eval_mse_loss": 1.254655804715431,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 153.0586,
      "eval_samples_per_second": 196.003,
      "eval_steps_per_second": 3.064,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5599673331927643,
      "flow/improvement_ratio": 0.8927549306771903,
      "flow/mag_ratio_mean": 0.5825493686488951,
      "flow/mag_ratio_std": 0.25123427366651196,
      "step": 8192
    },
    {
      "epoch": 0.36408300471911564,
      "grad_norm": 1.2766138315200806,
      "learning_rate": 0.0007274392162748551,
      "loss": 1.6162679195404053,
      "step": 8448
    },
    {
      "epoch": 0.3751158230439373,
      "grad_norm": 0.862872302532196,
      "learning_rate": 0.000711518231245687,
      "loss": 1.6088062524795532,
      "step": 8704
    },
    {
      "epoch": 0.38614864136875904,
      "grad_norm": 0.7975575923919678,
      "learning_rate": 0.0006953315799141723,
      "loss": 1.6033779382705688,
      "step": 8960
    },
    {
      "epoch": 0.3971814596935807,
      "grad_norm": 1.822509765625,
      "learning_rate": 0.0006788995926687669,
      "loss": 1.6062895059585571,
      "step": 9216
    },
    {
      "epoch": 0.3971814596935807,
      "eval_bleu": 0.9410773493346376,
      "eval_cos_loss": 0.4370425479498499,
      "eval_dec_loss": 0.10427019556861188,
      "eval_loss": 1.6075624590997757,
      "eval_mse2_loss": 0.1481303899272927,
      "eval_mse_loss": 1.247209641470838,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5629574498896406,
      "flow/improvement_ratio": 0.8962287469459241,
      "flow/mag_ratio_mean": 0.5783110863364327,
      "flow/mag_ratio_std": 0.2480927841432059,
      "step": 9216
    },
    {
      "epoch": 0.3971814596935807,
      "eval_bleu": 0.9410773493346376,
      "eval_cos_loss": 0.4370425479498499,
      "eval_dec_loss": 0.10427019556861188,
      "eval_loss": 1.6075624590997757,
      "eval_mse2_loss": 0.1481303899272927,
      "eval_mse_loss": 1.247209641470838,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 159.7751,
      "eval_samples_per_second": 187.764,
      "eval_steps_per_second": 2.935,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5629574498896406,
      "flow/improvement_ratio": 0.8962287469459241,
      "flow/mag_ratio_mean": 0.5783110863364327,
      "flow/mag_ratio_std": 0.2480927841432059,
      "step": 9216
    },
    {
      "epoch": 0.4082142780184024,
      "grad_norm": 1.1795552968978882,
      "learning_rate": 0.0006622429080391422,
      "loss": 1.6098705530166626,
      "step": 9472
    },
    {
      "epoch": 0.4192470963432241,
      "grad_norm": 0.8205899000167847,
      "learning_rate": 0.0006453824467742515,
      "loss": 1.6050623655319214,
      "step": 9728
    },
    {
      "epoch": 0.43027991466804577,
      "grad_norm": 0.6470943093299866,
      "learning_rate": 0.0006283393855659275,
      "loss": 1.61065673828125,
      "step": 9984
    },
    {
      "epoch": 0.44131273299286744,
      "grad_norm": 0.9093553423881531,
      "learning_rate": 0.0006111351304510173,
      "loss": 1.6007680892944336,
      "step": 10240
    },
    {
      "epoch": 0.44131273299286744,
      "eval_bleu": 0.9417572640186486,
      "eval_cos_loss": 0.4365569194242644,
      "eval_dec_loss": 0.10090916226905927,
      "eval_loss": 1.6032250524838088,
      "eval_mse2_loss": 0.14803322787478027,
      "eval_mse_loss": 1.2463789934288465,
      "eval_rec_loss": 0.047009017791098624,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5634430805757356,
      "flow/improvement_ratio": 0.8962225704304954,
      "flow/mag_ratio_mean": 0.5846807780042131,
      "flow/mag_ratio_std": 0.25302369241267125,
      "step": 10240
    },
    {
      "epoch": 0.44131273299286744,
      "eval_bleu": 0.9417572640186486,
      "eval_cos_loss": 0.4365569194242644,
      "eval_dec_loss": 0.10090916226905927,
      "eval_loss": 1.6032250524838088,
      "eval_mse2_loss": 0.14803322787478027,
      "eval_mse_loss": 1.2463789934288465,
      "eval_rec_loss": 0.047009017791098624,
      "eval_runtime": 152.6772,
      "eval_samples_per_second": 196.493,
      "eval_steps_per_second": 3.072,
      "eval_var_loss": 0.01723895594080501,
      "flow/cos_sim": 0.5634430805757356,
      "flow/improvement_ratio": 0.8962225704304954,
      "flow/mag_ratio_mean": 0.5846807780042131,
      "flow/mag_ratio_std": 0.25302369241267125,
      "step": 10240
    }
  ],
  "logging_steps": 256,
  "max_steps": 23204,
  "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
}