{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.3310701584222438,
  "eval_steps": 1024,
  "global_step": 7168,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.011823934229365849,
      "grad_norm": 1.1487267017364502,
      "learning_rate": 0.000498046875,
      "loss": 11.798027992248535,
      "step": 256
    },
    {
      "epoch": 0.023647868458731697,
      "grad_norm": 0.8386015295982361,
      "learning_rate": 0.000998046875,
      "loss": 1.7853779792785645,
      "step": 512
    },
    {
      "epoch": 0.03547180268809755,
      "grad_norm": 0.7344363331794739,
      "learning_rate": 0.000999640996023194,
      "loss": 1.103014588356018,
      "step": 768
    },
    {
      "epoch": 0.047295736917463395,
      "grad_norm": 1.1188315153121948,
      "learning_rate": 0.0009985588674043958,
      "loss": 0.9580796360969543,
      "step": 1024
    },
    {
      "epoch": 0.047295736917463395,
      "eval_bleu": 0.9192375404443994,
      "eval_ce_loss": 0.26167990953648745,
      "eval_cos_loss": 0.26406568816127296,
      "eval_loss": 0.9037111119864738,
      "eval_mse_loss": 0.6016385473617135,
      "eval_rec_loss": 0.013986090569199833,
      "flow/cos_sim": 0.7359343250048215,
      "flow/improvement_ratio": 0.9760946458605326,
      "flow/mag_ratio_mean": 0.7269674200717717,
      "flow/mag_ratio_std": 0.1390784539316343,
      "step": 1024
    },
    {
      "epoch": 0.047295736917463395,
      "eval_bleu": 0.9192375404443994,
      "eval_ce_loss": 0.26167990953648745,
      "eval_cos_loss": 0.26406568816127296,
      "eval_loss": 0.9037111119864738,
      "eval_mse_loss": 0.6016385473617135,
      "eval_rec_loss": 0.013986090569199833,
      "eval_runtime": 144.0156,
      "eval_samples_per_second": 194.375,
      "eval_steps_per_second": 3.041,
      "flow/cos_sim": 0.7359343250048215,
      "flow/improvement_ratio": 0.9760946458605326,
      "flow/mag_ratio_mean": 0.7269674200717717,
      "flow/mag_ratio_std": 0.1390784539316343,
      "step": 1024
    },
    {
      "epoch": 0.05911967114682925,
      "grad_norm": 1.0113043785095215,
      "learning_rate": 0.0009967551747861387,
      "loss": 0.8836896419525146,
      "step": 1280
    },
    {
      "epoch": 0.0709436053761951,
      "grad_norm": 0.9680636525154114,
      "learning_rate": 0.000994232528651847,
      "loss": 0.8432819247245789,
      "step": 1536
    },
    {
      "epoch": 0.08276753960556095,
      "grad_norm": 1.166627049446106,
      "learning_rate": 0.0009909945800260092,
      "loss": 0.7870283126831055,
      "step": 1792
    },
    {
      "epoch": 0.09459147383492679,
      "grad_norm": 0.7747617363929749,
      "learning_rate": 0.0009870460151900522,
      "loss": 0.7735522389411926,
      "step": 2048
    },
    {
      "epoch": 0.09459147383492679,
      "eval_bleu": 0.9264483676285591,
      "eval_ce_loss": 0.22021640589690372,
      "eval_cos_loss": 0.15934634184864557,
      "eval_loss": 0.751849727815689,
      "eval_mse_loss": 0.51190055103879,
      "eval_rec_loss": 0.0037981332031229603,
      "flow/cos_sim": 0.8406536742432477,
      "flow/improvement_ratio": 0.9754726998337871,
      "flow/mag_ratio_mean": 0.8395581131112085,
      "flow/mag_ratio_std": 0.09706963221096013,
      "step": 2048
    },
    {
      "epoch": 0.09459147383492679,
      "eval_bleu": 0.9264483676285591,
      "eval_ce_loss": 0.22021640589690372,
      "eval_cos_loss": 0.15934634184864557,
      "eval_loss": 0.751849727815689,
      "eval_mse_loss": 0.51190055103879,
      "eval_rec_loss": 0.0037981332031229603,
      "eval_runtime": 139.3758,
      "eval_samples_per_second": 200.845,
      "eval_steps_per_second": 3.143,
      "flow/cos_sim": 0.8406536742432477,
      "flow/improvement_ratio": 0.9754726998337871,
      "flow/mag_ratio_mean": 0.8395581131112085,
      "flow/mag_ratio_std": 0.09706963221096013,
      "step": 2048
    },
    {
      "epoch": 0.10641540806429264,
      "grad_norm": 1.3507742881774902,
      "learning_rate": 0.0009823925488998885,
      "loss": 0.7531520128250122,
      "step": 2304
    },
    {
      "epoch": 0.1182393422936585,
      "grad_norm": 1.090326189994812,
      "learning_rate": 0.0009770409161149525,
      "loss": 0.7384664416313171,
      "step": 2560
    },
    {
      "epoch": 0.13006327652302435,
      "grad_norm": 1.6648627519607544,
      "learning_rate": 0.0009709988622506973,
      "loss": 0.7159472107887268,
      "step": 2816
    },
    {
      "epoch": 0.1418872107523902,
      "grad_norm": 0.720405638217926,
      "learning_rate": 0.000964275131968659,
      "loss": 0.7134207487106323,
      "step": 3072
    },
    {
      "epoch": 0.1418872107523902,
      "eval_bleu": 0.9362046037171927,
      "eval_ce_loss": 0.18669388993094638,
      "eval_cos_loss": 0.09599785676829892,
      "eval_loss": 0.7009708360177741,
      "eval_mse_loss": 0.5027340843797274,
      "eval_rec_loss": 0.0019430727923829022,
      "flow/cos_sim": 0.9040021625555814,
      "flow/improvement_ratio": 0.9753212502830104,
      "flow/mag_ratio_mean": 0.906250716208323,
      "flow/mag_ratio_std": 0.07307879087519428,
      "step": 3072
    },
    {
      "epoch": 0.1418872107523902,
      "eval_bleu": 0.9362046037171927,
      "eval_ce_loss": 0.18669388993094638,
      "eval_cos_loss": 0.09599785676829892,
      "eval_loss": 0.7009708360177741,
      "eval_mse_loss": 0.5027340843797274,
      "eval_rec_loss": 0.0019430727923829022,
      "eval_runtime": 139.2617,
      "eval_samples_per_second": 201.01,
      "eval_steps_per_second": 3.145,
      "flow/cos_sim": 0.9040021625555814,
      "flow/improvement_ratio": 0.9753212502830104,
      "flow/mag_ratio_mean": 0.906250716208323,
      "flow/mag_ratio_std": 0.07307879087519428,
      "step": 3072
    },
    {
      "epoch": 0.15371114498175603,
      "grad_norm": 1.171225905418396,
      "learning_rate": 0.0009568794565203123,
      "loss": 0.6967981457710266,
      "step": 3328
    },
    {
      "epoch": 0.1655350792111219,
      "grad_norm": 0.9184499979019165,
      "learning_rate": 0.0009488225396630347,
      "loss": 0.6859588623046875,
      "step": 3584
    },
    {
      "epoch": 0.17735901344048774,
      "grad_norm": 1.0972322225570679,
      "learning_rate": 0.0009401160421685646,
      "loss": 0.68483966588974,
      "step": 3840
    },
    {
      "epoch": 0.18918294766985358,
      "grad_norm": 1.2944236993789673,
      "learning_rate": 0.0009307725649463714,
      "loss": 0.6722217202186584,
      "step": 4096
    },
    {
      "epoch": 0.18918294766985358,
      "eval_bleu": 0.9367529454304925,
      "eval_ce_loss": 0.17851220300531687,
      "eval_cos_loss": 0.0685597131088308,
      "eval_loss": 0.6638682314522191,
      "eval_mse_loss": 0.4771829223660029,
      "eval_rec_loss": 0.0013171346048419428,
      "flow/cos_sim": 0.9314403127045392,
      "flow/improvement_ratio": 0.9747336862021929,
      "flow/mag_ratio_mean": 0.9314602082722807,
      "flow/mag_ratio_std": 0.05983651272068013,
      "step": 4096
    },
    {
      "epoch": 0.18918294766985358,
      "eval_bleu": 0.9367529454304925,
      "eval_ce_loss": 0.17851220300531687,
      "eval_cos_loss": 0.0685597131088308,
      "eval_loss": 0.6638682314522191,
      "eval_mse_loss": 0.4771829223660029,
      "eval_rec_loss": 0.0013171346048419428,
      "eval_runtime": 143.247,
      "eval_samples_per_second": 195.418,
      "eval_steps_per_second": 3.058,
      "flow/cos_sim": 0.9314403127045392,
      "flow/improvement_ratio": 0.9747336862021929,
      "flow/mag_ratio_mean": 0.9314602082722807,
      "flow/mag_ratio_std": 0.05983651272068013,
      "step": 4096
    },
    {
      "epoch": 0.20100688189921945,
      "grad_norm": 1.0950664281845093,
      "learning_rate": 0.0009208056308063659,
      "loss": 0.6635431051254272,
      "step": 4352
    },
    {
      "epoch": 0.2128308161285853,
      "grad_norm": 1.0510311126708984,
      "learning_rate": 0.0009102296648873445,
      "loss": 0.652130126953125,
      "step": 4608
    },
    {
      "epoch": 0.22465475035795113,
      "grad_norm": 0.7107524275779724,
      "learning_rate": 0.0008990599737794927,
      "loss": 0.6548014283180237,
      "step": 4864
    },
    {
      "epoch": 0.236478684587317,
      "grad_norm": 1.119279146194458,
      "learning_rate": 0.0008873127233711644,
      "loss": 0.644295871257782,
      "step": 5120
    },
    {
      "epoch": 0.236478684587317,
      "eval_bleu": 0.9375087809294717,
      "eval_ce_loss": 0.17866700632629498,
      "eval_cos_loss": 0.05568918508379699,
      "eval_loss": 0.6384022356304404,
      "eval_mse_loss": 0.4531491862856634,
      "eval_rec_loss": 0.0010171248973892112,
      "flow/cos_sim": 0.9443108406785417,
      "flow/improvement_ratio": 0.9754358607612245,
      "flow/mag_ratio_mean": 0.9432625220790846,
      "flow/mag_ratio_std": 0.0532344374550532,
      "step": 5120
    },
    {
      "epoch": 0.236478684587317,
      "eval_bleu": 0.9375087809294717,
      "eval_ce_loss": 0.17866700632629498,
      "eval_cos_loss": 0.05568918508379699,
      "eval_loss": 0.6384022356304404,
      "eval_mse_loss": 0.4531491862856634,
      "eval_rec_loss": 0.0010171248973892112,
      "eval_runtime": 140.7376,
      "eval_samples_per_second": 198.902,
      "eval_steps_per_second": 3.112,
      "flow/cos_sim": 0.9443108406785417,
      "flow/improvement_ratio": 0.9754358607612245,
      "flow/mag_ratio_mean": 0.9432625220790846,
      "flow/mag_ratio_std": 0.0532344374550532,
      "step": 5120
    },
    {
      "epoch": 0.24830261881668284,
      "grad_norm": 1.223329782485962,
      "learning_rate": 0.0008750049154520011,
      "loss": 0.6385497450828552,
      "step": 5376
    },
    {
      "epoch": 0.2601265530460487,
      "grad_norm": 0.7951129078865051,
      "learning_rate": 0.0008621543631062487,
      "loss": 0.6350575089454651,
      "step": 5632
    },
    {
      "epoch": 0.27195048727541454,
      "grad_norm": 0.8830247521400452,
      "learning_rate": 0.0008487796649318904,
      "loss": 0.6269800066947937,
      "step": 5888
    },
    {
      "epoch": 0.2837744215047804,
      "grad_norm": 1.0399079322814941,
      "learning_rate": 0.0008349001781229053,
      "loss": 0.6236906051635742,
      "step": 6144
    },
    {
      "epoch": 0.2837744215047804,
      "eval_bleu": 0.9406288888554537,
      "eval_ce_loss": 0.16271350685439018,
      "eval_cos_loss": 0.04946032597696128,
      "eval_loss": 0.6158577807962078,
      "eval_mse_loss": 0.4472781540868489,
      "eval_rec_loss": 0.0009200886164281037,
      "flow/cos_sim": 0.9505396935765602,
      "flow/improvement_ratio": 0.9752184091365501,
      "flow/mag_ratio_mean": 0.9574754819205907,
      "flow/mag_ratio_std": 0.04891788842131014,
      "step": 6144
    },
    {
      "epoch": 0.2837744215047804,
      "eval_bleu": 0.9406288888554537,
      "eval_ce_loss": 0.16271350685439018,
      "eval_cos_loss": 0.04946032597696128,
      "eval_loss": 0.6158577807962078,
      "eval_mse_loss": 0.4472781540868489,
      "eval_rec_loss": 0.0009200886164281037,
      "eval_runtime": 138.7881,
      "eval_samples_per_second": 201.696,
      "eval_steps_per_second": 3.156,
      "flow/cos_sim": 0.9505396935765602,
      "flow/improvement_ratio": 0.9752184091365501,
      "flow/mag_ratio_mean": 0.9574754819205907,
      "flow/mag_ratio_std": 0.04891788842131014,
      "step": 6144
    },
    {
      "epoch": 0.2955983557341462,
      "grad_norm": 1.2804330587387085,
      "learning_rate": 0.0008205359904536107,
      "loss": 0.6104704737663269,
      "step": 6400
    },
    {
      "epoch": 0.30742228996351206,
      "grad_norm": 1.038807988166809,
      "learning_rate": 0.0008057078912056363,
      "loss": 0.6035579442977905,
      "step": 6656
    },
    {
      "epoch": 0.3192462241928779,
      "grad_norm": 1.1162539720535278,
      "learning_rate": 0.0007904373410796086,
      "loss": 0.6099694967269897,
      "step": 6912
    },
    {
      "epoch": 0.3310701584222438,
      "grad_norm": 0.8053554892539978,
      "learning_rate": 0.0007747464411350876,
      "loss": 0.6010444760322571,
      "step": 7168
    },
    {
      "epoch": 0.3310701584222438,
      "eval_bleu": 0.9403609007223014,
      "eval_ce_loss": 0.16489822024130793,
      "eval_cos_loss": 0.042615741474307293,
      "eval_loss": 0.5965314044799979,
      "eval_mse_loss": 0.4266453656839998,
      "eval_rec_loss": 0.0007262465627901213,
      "flow/cos_sim": 0.9573842790573155,
      "flow/improvement_ratio": 0.9754182146564466,
      "flow/mag_ratio_mean": 0.9583017167435389,
      "flow/mag_ratio_std": 0.044361623012584096,
      "step": 7168
    },
    {
      "epoch": 0.3310701584222438,
      "eval_bleu": 0.9403609007223014,
      "eval_ce_loss": 0.16489822024130793,
      "eval_cos_loss": 0.042615741474307293,
      "eval_loss": 0.5965314044799979,
      "eval_mse_loss": 0.4266453656839998,
      "eval_rec_loss": 0.0007262465627901213,
      "eval_runtime": 141.0373,
      "eval_samples_per_second": 198.479,
      "eval_steps_per_second": 3.106,
      "flow/cos_sim": 0.9573842790573155,
      "flow/improvement_ratio": 0.9754182146564466,
      "flow/mag_ratio_mean": 0.9583017167435389,
      "flow/mag_ratio_std": 0.044361623012584096,
      "step": 7168
    }
  ],
  "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
}