{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.7592870294793189,
  "eval_steps": 2500,
  "global_step": 20000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01898217573698297,
      "grad_norm": 2.8858866691589355,
      "learning_rate": 4.735843087630496e-05,
      "loss": 6.00573583984375,
      "step": 500
    },
    {
      "epoch": 0.03796435147396594,
      "grad_norm": 0.22312745451927185,
      "learning_rate": 9.481176842771275e-05,
      "loss": 1.7886568603515625,
      "step": 1000
    },
    {
      "epoch": 0.05694652721094892,
      "grad_norm": 0.3883834183216095,
      "learning_rate": 0.0001422651059791205,
      "loss": 1.4103663330078124,
      "step": 1500
    },
    {
      "epoch": 0.07592870294793189,
      "grad_norm": 0.28461548686027527,
      "learning_rate": 0.0001897184435305283,
      "loss": 1.2999722900390625,
      "step": 2000
    },
    {
      "epoch": 0.09491087868491487,
      "grad_norm": 0.3026827871799469,
      "learning_rate": 0.00023717178108193608,
      "loss": 1.27368310546875,
      "step": 2500
    },
    {
      "epoch": 0.09491087868491487,
      "eval_correction_accuracy": 0.9691768119898284,
      "eval_detection_f1": 0.5029481803464545,
      "eval_loss": 1.2979440689086914,
      "eval_runtime": 60.3464,
      "eval_samples_per_second": 371.886,
      "eval_steps_per_second": 2.916,
      "step": 2500
    },
    {
      "epoch": 0.09491087868491487,
      "eval_edit_f0_5": 0.8420576085111199,
      "eval_edit_f0_5_m0": 0.8057950832842375,
      "eval_edit_f0_5_m0.1": 0.8175929454030809,
      "eval_edit_f0_5_m0.2": 0.8316255884796455,
      "eval_edit_f0_5_m0.3": 0.8420576085111199,
      "step": 2500
    },
    {
      "epoch": 0.09491087868491487,
      "eval_real_f0_5": 0.42910447761194037,
      "eval_real_f0_5_m0": 0.3888400702987698,
      "eval_real_f0_5_m0.1": 0.4028041825095057,
      "eval_real_f0_5_m0.2": 0.4175050301810866,
      "eval_real_f0_5_m0.3": 0.42910447761194037,
      "step": 2500
    },
    {
      "epoch": 0.11389305442189784,
      "grad_norm": 0.27056726813316345,
      "learning_rate": 0.00028462511863334386,
      "loss": 1.2630726318359375,
      "step": 3000
    },
    {
      "epoch": 0.1328752301588808,
      "grad_norm": 0.2008272260427475,
      "learning_rate": 0.00029996551748314967,
      "loss": 1.2586767578125,
      "step": 3500
    },
    {
      "epoch": 0.15185740589586377,
      "grad_norm": 0.23608435690402985,
      "learning_rate": 0.00029978808195120705,
      "loss": 1.2485474853515626,
      "step": 4000
    },
    {
      "epoch": 0.17083958163284677,
      "grad_norm": 0.17198926210403442,
      "learning_rate": 0.00029945995039741313,
      "loss": 1.2438245849609375,
      "step": 4500
    },
    {
      "epoch": 0.18982175736982973,
      "grad_norm": 0.20551930367946625,
      "learning_rate": 0.00029898145294228977,
      "loss": 1.2398516845703125,
      "step": 5000
    },
    {
      "epoch": 0.18982175736982973,
      "eval_correction_accuracy": 0.9716923312269569,
      "eval_detection_f1": 0.5269904148108618,
      "eval_loss": 1.2654461860656738,
      "eval_runtime": 59.8792,
      "eval_samples_per_second": 374.788,
      "eval_steps_per_second": 2.939,
      "step": 5000
    },
    {
      "epoch": 0.18982175736982973,
      "eval_edit_f0_5": 0.872664843158936,
      "eval_edit_f0_5_m0": 0.8547840106743758,
      "eval_edit_f0_5_m0.1": 0.8604271144418775,
      "eval_edit_f0_5_m0.2": 0.8685822219678324,
      "eval_edit_f0_5_m0.3": 0.872664843158936,
      "step": 5000
    },
    {
      "epoch": 0.18982175736982973,
      "eval_real_f0_5": 0.46151439299123903,
      "eval_real_f0_5_m0": 0.4191919191919192,
      "eval_real_f0_5_m0.1": 0.4297297297297297,
      "eval_real_f0_5_m0.2": 0.445906432748538,
      "eval_real_f0_5_m0.3": 0.46151439299123903,
      "step": 5000
    },
    {
      "epoch": 0.2088039331068127,
      "grad_norm": 0.20124033093452454,
      "learning_rate": 0.00029835307098370265,
      "loss": 1.2357362060546875,
      "step": 5500
    },
    {
      "epoch": 0.2277861088437957,
      "grad_norm": 0.23115237057209015,
      "learning_rate": 0.0002975754367125453,
      "loss": 1.2375308837890624,
      "step": 6000
    },
    {
      "epoch": 0.24676828458077865,
      "grad_norm": 0.1954943835735321,
      "learning_rate": 0.00029664933247671615,
      "loss": 1.2349964599609375,
      "step": 6500
    },
    {
      "epoch": 0.2657504603177616,
      "grad_norm": 0.17400752007961273,
      "learning_rate": 0.0002955756899940283,
      "loss": 1.2303714599609374,
      "step": 7000
    },
    {
      "epoch": 0.2847326360547446,
      "grad_norm": 0.17888489365577698,
      "learning_rate": 0.00029435558941484435,
      "loss": 1.2309805908203124,
      "step": 7500
    },
    {
      "epoch": 0.2847326360547446,
      "eval_correction_accuracy": 0.9723985228660039,
      "eval_detection_f1": 0.5397412768051171,
      "eval_loss": 1.2546594142913818,
      "eval_runtime": 60.3599,
      "eval_samples_per_second": 371.803,
      "eval_steps_per_second": 2.916,
      "step": 7500
    },
    {
      "epoch": 0.2847326360547446,
      "eval_edit_f0_5": 0.8861485517636936,
      "eval_edit_f0_5_m0": 0.8749653451621846,
      "eval_edit_f0_5_m0.1": 0.8788065728226123,
      "eval_edit_f0_5_m0.2": 0.8814529381764608,
      "eval_edit_f0_5_m0.3": 0.8861485517636936,
      "step": 7500
    },
    {
      "epoch": 0.2847326360547446,
      "eval_real_f0_5": 0.47371638141809297,
      "eval_real_f0_5_m0": 0.4501055966209081,
      "eval_real_f0_5_m0.1": 0.4613259668508288,
      "eval_real_f0_5_m0.2": 0.47164351851851855,
      "eval_real_f0_5_m0.3": 0.47371638141809297,
      "step": 7500
    },
    {
      "epoch": 0.30371481179172755,
      "grad_norm": 0.1810256987810135,
      "learning_rate": 0.0002929902582353787,
      "loss": 1.227394775390625,
      "step": 8000
    },
    {
      "epoch": 0.32269698752871057,
      "grad_norm": 0.18248461186885834,
      "learning_rate": 0.00029148107006276056,
      "loss": 1.22420361328125,
      "step": 8500
    },
    {
      "epoch": 0.34167916326569353,
      "grad_norm": 0.1456211358308792,
      "learning_rate": 0.0002898295432331011,
      "loss": 1.22528125,
      "step": 9000
    },
    {
      "epoch": 0.3606613390026765,
      "grad_norm": 0.1682305634021759,
      "learning_rate": 0.0002880373392839537,
      "loss": 1.2209486083984376,
      "step": 9500
    },
    {
      "epoch": 0.37964351473965946,
      "grad_norm": 0.1455032080411911,
      "learning_rate": 0.00028610626128270495,
      "loss": 1.222905029296875,
      "step": 10000
    },
    {
      "epoch": 0.37964351473965946,
      "eval_correction_accuracy": 0.9729365736386112,
      "eval_detection_f1": 0.5878967328292231,
      "eval_loss": 1.2507761716842651,
      "eval_runtime": 60.7194,
      "eval_samples_per_second": 369.602,
      "eval_steps_per_second": 2.899,
      "step": 10000
    },
    {
      "epoch": 0.37964351473965946,
      "eval_edit_f0_5": 0.8773855767606823,
      "eval_edit_f0_5_m0": 0.8642143753752115,
      "eval_edit_f0_5_m0.1": 0.871961638097338,
      "eval_edit_f0_5_m0.2": 0.8739938939772413,
      "eval_edit_f0_5_m0.3": 0.8773855767606823,
      "step": 10000
    },
    {
      "epoch": 0.37964351473965946,
      "eval_real_f0_5": 0.43992027334851935,
      "eval_real_f0_5_m0": 0.4214697406340058,
      "eval_real_f0_5_m0.1": 0.42792792792792794,
      "eval_real_f0_5_m0.2": 0.42728237791932056,
      "eval_real_f0_5_m0.3": 0.43992027334851935,
      "step": 10000
    },
    {
      "epoch": 0.3986256904766424,
      "grad_norm": 0.20864084362983704,
      "learning_rate": 0.0002840382520125783,
      "loss": 1.22280712890625,
      "step": 10500
    },
    {
      "epoch": 0.4176078662136254,
      "grad_norm": 0.19003528356552124,
      "learning_rate": 0.0002818353920180744,
      "loss": 1.221659423828125,
      "step": 11000
    },
    {
      "epoch": 0.43659004195060835,
      "grad_norm": 0.1750059872865677,
      "learning_rate": 0.0002794998975118153,
      "loss": 1.2191646728515626,
      "step": 11500
    },
    {
      "epoch": 0.4555722176875914,
      "grad_norm": 0.13223567605018616,
      "learning_rate": 0.00027703411814489855,
      "loss": 1.216736328125,
      "step": 12000
    },
    {
      "epoch": 0.47455439342457434,
      "grad_norm": 0.10733171552419662,
      "learning_rate": 0.00027444053464300346,
      "loss": 1.2142664794921876,
      "step": 12500
    },
    {
      "epoch": 0.47455439342457434,
      "eval_correction_accuracy": 0.9733371445263059,
      "eval_detection_f1": 0.5704304192239259,
      "eval_loss": 1.2477121353149414,
      "eval_runtime": 60.6105,
      "eval_samples_per_second": 370.266,
      "eval_steps_per_second": 2.904,
      "step": 12500
    },
    {
      "epoch": 0.47455439342457434,
      "eval_edit_f0_5": 0.8926609218668673,
      "eval_edit_f0_5_m0": 0.8845489496851842,
      "eval_edit_f0_5_m0.1": 0.8883563574851988,
      "eval_edit_f0_5_m0.2": 0.89086859688196,
      "eval_edit_f0_5_m0.3": 0.8926609218668673,
      "step": 12500
    },
    {
      "epoch": 0.47455439342457434,
      "eval_real_f0_5": 0.4787581699346405,
      "eval_real_f0_5_m0": 0.43419633225458465,
      "eval_real_f0_5_m0.1": 0.44776119402985065,
      "eval_real_f0_5_m0.2": 0.46760391198044005,
      "eval_real_f0_5_m0.3": 0.4787581699346405,
      "step": 12500
    },
    {
      "epoch": 0.4935365691615573,
      "grad_norm": 0.1579447090625763,
      "learning_rate": 0.0002717217563106287,
      "loss": 1.21451904296875,
      "step": 13000
    },
    {
      "epoch": 0.5125187448985403,
      "grad_norm": 0.13867036998271942,
      "learning_rate": 0.00026888051840597135,
      "loss": 1.2117508544921876,
      "step": 13500
    },
    {
      "epoch": 0.5315009206355232,
      "grad_norm": 0.11431417614221573,
      "learning_rate": 0.00026591967938908897,
      "loss": 1.2146973876953124,
      "step": 14000
    },
    {
      "epoch": 0.5504830963725063,
      "grad_norm": 0.14605163037776947,
      "learning_rate": 0.00026284221804611323,
      "loss": 1.211010498046875,
      "step": 14500
    },
    {
      "epoch": 0.5694652721094892,
      "grad_norm": 0.13300460577011108,
      "learning_rate": 0.0002596512304924075,
      "loss": 1.2157716064453126,
      "step": 15000
    },
    {
      "epoch": 0.5694652721094892,
      "eval_correction_accuracy": 0.9736186980555992,
      "eval_detection_f1": 0.6009406304526895,
      "eval_loss": 1.2431864738464355,
      "eval_runtime": 60.2876,
      "eval_samples_per_second": 372.249,
      "eval_steps_per_second": 2.919,
      "step": 15000
    },
    {
      "epoch": 0.5694652721094892,
      "eval_edit_f0_5": 0.8888329698596432,
      "eval_edit_f0_5_m0": 0.8778231292517006,
      "eval_edit_f0_5_m0.1": 0.8819654131210541,
      "eval_edit_f0_5_m0.2": 0.8855925167432335,
      "eval_edit_f0_5_m0.3": 0.8888329698596432,
      "step": 15000
    },
    {
      "epoch": 0.5694652721094892,
      "eval_real_f0_5": 0.4537767756482525,
      "eval_real_f0_5_m0": 0.4154693486590038,
      "eval_real_f0_5_m0.1": 0.42251755265797397,
      "eval_real_f0_5_m0.2": 0.4458512931034483,
      "eval_real_f0_5_m0.3": 0.4537767756482525,
      "step": 15000
    },
    {
      "epoch": 0.5884474478464722,
      "grad_norm": 0.16023032367229462,
      "learning_rate": 0.0002563499270576846,
      "loss": 1.21418017578125,
      "step": 15500
    },
    {
      "epoch": 0.6074296235834551,
      "grad_norm": 0.14038224518299103,
      "learning_rate": 0.0002529416290562178,
      "loss": 1.2102242431640624,
      "step": 16000
    },
    {
      "epoch": 0.6264117993204381,
      "grad_norm": 0.19150033593177795,
      "learning_rate": 0.00024942976544539404,
      "loss": 1.21311474609375,
      "step": 16500
    },
    {
      "epoch": 0.6453939750574211,
      "grad_norm": 0.13594546914100647,
      "learning_rate": 0.0002458178693759725,
      "loss": 1.2101019287109376,
      "step": 17000
    },
    {
      "epoch": 0.664376150794404,
      "grad_norm": 0.0971856564283371,
      "learning_rate": 0.00024210957463751702,
      "loss": 1.2085975341796875,
      "step": 17500
    },
    {
      "epoch": 0.664376150794404,
      "eval_correction_accuracy": 0.9739441000853563,
      "eval_detection_f1": 0.6041727242827385,
      "eval_loss": 1.2378469705581665,
      "eval_runtime": 60.7237,
      "eval_samples_per_second": 369.575,
      "eval_steps_per_second": 2.898,
      "step": 17500
    },
    {
      "epoch": 0.664376150794404,
      "eval_edit_f0_5": 0.8950703830407717,
      "eval_edit_f0_5_m0": 0.8823850308221047,
      "eval_edit_f0_5_m0.1": 0.8891276795062545,
      "eval_edit_f0_5_m0.2": 0.891679139641887,
      "eval_edit_f0_5_m0.3": 0.8950703830407717,
      "step": 17500
    },
    {
      "epoch": 0.664376150794404,
      "eval_real_f0_5": 0.45221843003412976,
      "eval_real_f0_5_m0": 0.4207080504364695,
      "eval_real_f0_5_m0.1": 0.4268916155419223,
      "eval_real_f0_5_m0.2": 0.43951612903225806,
      "eval_real_f0_5_m0.3": 0.45221843003412976,
      "step": 17500
    },
    {
      "epoch": 0.6833583265313871,
      "grad_norm": 0.23436909914016724,
      "learning_rate": 0.00023830861200258084,
      "loss": 1.2103984375,
      "step": 18000
    },
    {
      "epoch": 0.70234050226837,
      "grad_norm": 0.10302391648292542,
      "learning_rate": 0.00023441880547331984,
      "loss": 1.209429443359375,
      "step": 18500
    },
    {
      "epoch": 0.721322678005353,
      "grad_norm": 0.16614890098571777,
      "learning_rate": 0.00023044406843431085,
      "loss": 1.210209228515625,
      "step": 19000
    },
    {
      "epoch": 0.7403048537423359,
      "grad_norm": 0.12630845606327057,
      "learning_rate": 0.00022638839971544675,
      "loss": 1.2063096923828125,
      "step": 19500
    },
    {
      "epoch": 0.7592870294793189,
      "grad_norm": 0.16276787221431732,
      "learning_rate": 0.00022225587956886718,
      "loss": 1.20748486328125,
      "step": 20000
    },
    {
      "epoch": 0.7592870294793189,
      "eval_correction_accuracy": 0.974335110021822,
      "eval_detection_f1": 0.6143125770798066,
      "eval_loss": 1.234763264656067,
      "eval_runtime": 60.5996,
      "eval_samples_per_second": 370.332,
      "eval_steps_per_second": 2.904,
      "step": 20000
    },
    {
      "epoch": 0.7592870294793189,
      "eval_edit_f0_5": 0.9024403864410565,
      "eval_edit_f0_5_m0": 0.8842469830618539,
      "eval_edit_f0_5_m0.1": 0.8886223627541842,
      "eval_edit_f0_5_m0.2": 0.8945219837810999,
      "eval_edit_f0_5_m0.3": 0.9024403864410565,
      "step": 20000
    },
    {
      "epoch": 0.7592870294793189,
      "eval_real_f0_5": 0.4630195599022005,
      "eval_real_f0_5_m0": 0.44520547945205474,
      "eval_real_f0_5_m0.1": 0.4467213114754099,
      "eval_real_f0_5_m0.2": 0.4604503464203233,
      "eval_real_f0_5_m0.3": 0.4630195599022005,
      "step": 20000
    }
  ],
  "logging_steps": 500,
  "max_steps": 52682,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 2500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.6519904481356275e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}