{
  "best_global_step": 4000,
  "best_metric": 50.44642474127626,
  "best_model_checkpoint": "/content/drive/MyDrive/Dala/nllb_model/checkpoint-4000",
  "epoch": 3.2389631429728634,
  "eval_steps": 1000,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04050222762251924,
      "grad_norm": 2.6502063274383545,
      "learning_rate": 4.9000000000000005e-06,
      "loss": 5.3214,
      "step": 50
    },
    {
      "epoch": 0.08100445524503848,
      "grad_norm": 1.576899766921997,
      "learning_rate": 9.900000000000002e-06,
      "loss": 4.5892,
      "step": 100
    },
    {
      "epoch": 0.12150668286755771,
      "grad_norm": 1.3972806930541992,
      "learning_rate": 1.49e-05,
      "loss": 3.9326,
      "step": 150
    },
    {
      "epoch": 0.16200891049007696,
      "grad_norm": 1.6953352689743042,
      "learning_rate": 1.9900000000000003e-05,
      "loss": 3.5229,
      "step": 200
    },
    {
      "epoch": 0.2025111381125962,
      "grad_norm": 1.408992886543274,
      "learning_rate": 2.4900000000000002e-05,
      "loss": 3.2072,
      "step": 250
    },
    {
      "epoch": 0.24301336573511542,
      "grad_norm": 1.4607220888137817,
      "learning_rate": 2.9900000000000002e-05,
      "loss": 3.0183,
      "step": 300
    },
    {
      "epoch": 0.28351559335763465,
      "grad_norm": 1.6003291606903076,
      "learning_rate": 3.49e-05,
      "loss": 2.8487,
      "step": 350
    },
    {
      "epoch": 0.3240178209801539,
      "grad_norm": 1.5186729431152344,
      "learning_rate": 3.99e-05,
      "loss": 2.6323,
      "step": 400
    },
    {
      "epoch": 0.3645200486026732,
      "grad_norm": 1.4992791414260864,
      "learning_rate": 4.49e-05,
      "loss": 2.4942,
      "step": 450
    },
    {
      "epoch": 0.4050222762251924,
      "grad_norm": 1.5211806297302246,
      "learning_rate": 4.99e-05,
      "loss": 2.3595,
      "step": 500
    },
    {
      "epoch": 0.44552450384771164,
      "grad_norm": 1.503055453300476,
      "learning_rate": 5.4900000000000006e-05,
      "loss": 2.2573,
      "step": 550
    },
    {
      "epoch": 0.48602673147023084,
      "grad_norm": 1.7007502317428589,
      "learning_rate": 5.99e-05,
      "loss": 2.1864,
      "step": 600
    },
    {
      "epoch": 0.5265289590927501,
      "grad_norm": 1.4285714626312256,
      "learning_rate": 6.49e-05,
      "loss": 2.0554,
      "step": 650
    },
    {
      "epoch": 0.5670311867152693,
      "grad_norm": 1.6091938018798828,
      "learning_rate": 6.99e-05,
      "loss": 2.019,
      "step": 700
    },
    {
      "epoch": 0.6075334143377886,
      "grad_norm": 1.7364023923873901,
      "learning_rate": 7.49e-05,
      "loss": 1.9724,
      "step": 750
    },
    {
      "epoch": 0.6480356419603078,
      "grad_norm": 1.450875163078308,
      "learning_rate": 7.99e-05,
      "loss": 1.9124,
      "step": 800
    },
    {
      "epoch": 0.688537869582827,
      "grad_norm": 1.468355655670166,
      "learning_rate": 8.49e-05,
      "loss": 1.8606,
      "step": 850
    },
    {
      "epoch": 0.7290400972053463,
      "grad_norm": 1.4059218168258667,
      "learning_rate": 8.99e-05,
      "loss": 1.8028,
      "step": 900
    },
    {
      "epoch": 0.7695423248278656,
      "grad_norm": 1.5709952116012573,
      "learning_rate": 9.49e-05,
      "loss": 1.7327,
      "step": 950
    },
    {
      "epoch": 0.8100445524503848,
      "grad_norm": 1.3530505895614624,
      "learning_rate": 9.99e-05,
      "loss": 1.7085,
      "step": 1000
    },
    {
      "epoch": 0.8100445524503848,
      "eval_bleu": 23.576904136717374,
      "eval_chrf": 40.858644433448234,
      "eval_loss": 1.5544426441192627,
      "eval_runtime": 569.3068,
      "eval_samples_per_second": 3.854,
      "eval_steps_per_second": 0.483,
      "step": 1000
    },
    {
      "epoch": 0.850546780072904,
      "grad_norm": 1.5039794445037842,
      "learning_rate": 9.956828193832599e-05,
      "loss": 1.675,
      "step": 1050
    },
    {
      "epoch": 0.8910490076954233,
      "grad_norm": 1.3560293912887573,
      "learning_rate": 9.912775330396477e-05,
      "loss": 1.6383,
      "step": 1100
    },
    {
      "epoch": 0.9315512353179425,
      "grad_norm": 1.483917474746704,
      "learning_rate": 9.868722466960353e-05,
      "loss": 1.5895,
      "step": 1150
    },
    {
      "epoch": 0.9720534629404617,
      "grad_norm": 1.2108521461486816,
      "learning_rate": 9.82466960352423e-05,
      "loss": 1.5732,
      "step": 1200
    },
    {
      "epoch": 1.0121506682867558,
      "grad_norm": 1.229095220565796,
      "learning_rate": 9.780616740088107e-05,
      "loss": 1.5249,
      "step": 1250
    },
    {
      "epoch": 1.052652895909275,
      "grad_norm": 1.3696666955947876,
      "learning_rate": 9.736563876651984e-05,
      "loss": 1.4009,
      "step": 1300
    },
    {
      "epoch": 1.0931551235317942,
      "grad_norm": 1.3702960014343262,
      "learning_rate": 9.692511013215859e-05,
      "loss": 1.4172,
      "step": 1350
    },
    {
      "epoch": 1.1336573511543135,
      "grad_norm": 1.363653540611267,
      "learning_rate": 9.648458149779736e-05,
      "loss": 1.412,
      "step": 1400
    },
    {
      "epoch": 1.1741595787768326,
      "grad_norm": 1.314254641532898,
      "learning_rate": 9.604405286343613e-05,
      "loss": 1.3905,
      "step": 1450
    },
    {
      "epoch": 1.214661806399352,
      "grad_norm": 1.302570104598999,
      "learning_rate": 9.560352422907489e-05,
      "loss": 1.3789,
      "step": 1500
    },
    {
      "epoch": 1.2551640340218713,
      "grad_norm": 1.4111872911453247,
      "learning_rate": 9.516299559471367e-05,
      "loss": 1.4103,
      "step": 1550
    },
    {
      "epoch": 1.2956662616443904,
      "grad_norm": 1.306458592414856,
      "learning_rate": 9.472246696035243e-05,
      "loss": 1.3807,
      "step": 1600
    },
    {
      "epoch": 1.3361684892669097,
      "grad_norm": 1.180014967918396,
      "learning_rate": 9.42819383259912e-05,
      "loss": 1.3666,
      "step": 1650
    },
    {
      "epoch": 1.376670716889429,
      "grad_norm": 1.2211802005767822,
      "learning_rate": 9.384140969162996e-05,
      "loss": 1.3055,
      "step": 1700
    },
    {
      "epoch": 1.417172944511948,
      "grad_norm": 2.443772792816162,
      "learning_rate": 9.340088105726873e-05,
      "loss": 1.3401,
      "step": 1750
    },
    {
      "epoch": 1.4576751721344674,
      "grad_norm": 1.241117238998413,
      "learning_rate": 9.29603524229075e-05,
      "loss": 1.3325,
      "step": 1800
    },
    {
      "epoch": 1.4981773997569867,
      "grad_norm": 1.3065532445907593,
      "learning_rate": 9.251982378854626e-05,
      "loss": 1.3409,
      "step": 1850
    },
    {
      "epoch": 1.5386796273795058,
      "grad_norm": 1.223604679107666,
      "learning_rate": 9.207929515418503e-05,
      "loss": 1.306,
      "step": 1900
    },
    {
      "epoch": 1.5791818550020251,
      "grad_norm": 1.2152003049850464,
      "learning_rate": 9.163876651982379e-05,
      "loss": 1.275,
      "step": 1950
    },
    {
      "epoch": 1.6196840826245444,
      "grad_norm": 1.0613232851028442,
      "learning_rate": 9.119823788546255e-05,
      "loss": 1.3005,
      "step": 2000
    },
    {
      "epoch": 1.6196840826245444,
      "eval_bleu": 30.160628892876375,
      "eval_chrf": 47.67539184781091,
      "eval_loss": 1.2586086988449097,
      "eval_runtime": 510.3397,
      "eval_samples_per_second": 4.299,
      "eval_steps_per_second": 0.539,
      "step": 2000
    },
    {
      "epoch": 1.6601863102470635,
      "grad_norm": 1.5416429042816162,
      "learning_rate": 9.075770925110133e-05,
      "loss": 1.3042,
      "step": 2050
    },
    {
      "epoch": 1.7006885378695829,
      "grad_norm": 1.3111222982406616,
      "learning_rate": 9.031718061674009e-05,
      "loss": 1.3081,
      "step": 2100
    },
    {
      "epoch": 1.7411907654921022,
      "grad_norm": 1.2473504543304443,
      "learning_rate": 8.987665198237886e-05,
      "loss": 1.2599,
      "step": 2150
    },
    {
      "epoch": 1.7816929931146213,
      "grad_norm": 1.267403483390808,
      "learning_rate": 8.943612334801763e-05,
      "loss": 1.2693,
      "step": 2200
    },
    {
      "epoch": 1.8221952207371406,
      "grad_norm": 1.1271162033081055,
      "learning_rate": 8.89955947136564e-05,
      "loss": 1.269,
      "step": 2250
    },
    {
      "epoch": 1.86269744835966,
      "grad_norm": 1.209218144416809,
      "learning_rate": 8.855506607929516e-05,
      "loss": 1.2379,
      "step": 2300
    },
    {
      "epoch": 1.903199675982179,
      "grad_norm": 1.30668044090271,
      "learning_rate": 8.811453744493392e-05,
      "loss": 1.2547,
      "step": 2350
    },
    {
      "epoch": 1.9437019036046983,
      "grad_norm": 1.0480397939682007,
      "learning_rate": 8.767400881057269e-05,
      "loss": 1.2334,
      "step": 2400
    },
    {
      "epoch": 1.9842041312272176,
      "grad_norm": 1.234099268913269,
      "learning_rate": 8.723348017621145e-05,
      "loss": 1.23,
      "step": 2450
    },
    {
      "epoch": 2.0243013365735116,
      "grad_norm": 1.1937974691390991,
      "learning_rate": 8.679295154185022e-05,
      "loss": 1.121,
      "step": 2500
    },
    {
      "epoch": 2.0648035641960307,
      "grad_norm": 1.277874231338501,
      "learning_rate": 8.635242290748899e-05,
      "loss": 1.0507,
      "step": 2550
    },
    {
      "epoch": 2.10530579181855,
      "grad_norm": 1.0902061462402344,
      "learning_rate": 8.591189427312776e-05,
      "loss": 1.0746,
      "step": 2600
    },
    {
      "epoch": 2.1458080194410694,
      "grad_norm": 1.304015040397644,
      "learning_rate": 8.547136563876652e-05,
      "loss": 1.0718,
      "step": 2650
    },
    {
      "epoch": 2.1863102470635885,
      "grad_norm": 1.2821192741394043,
      "learning_rate": 8.50308370044053e-05,
      "loss": 1.0724,
      "step": 2700
    },
    {
      "epoch": 2.2268124746861075,
      "grad_norm": 1.2118302583694458,
      "learning_rate": 8.459030837004406e-05,
      "loss": 1.0497,
      "step": 2750
    },
    {
      "epoch": 2.267314702308627,
      "grad_norm": 1.0659773349761963,
      "learning_rate": 8.414977973568282e-05,
      "loss": 1.0519,
      "step": 2800
    },
    {
      "epoch": 2.307816929931146,
      "grad_norm": 1.0666295289993286,
      "learning_rate": 8.37092511013216e-05,
      "loss": 1.0759,
      "step": 2850
    },
    {
      "epoch": 2.3483191575536653,
      "grad_norm": 1.0976322889328003,
      "learning_rate": 8.326872246696036e-05,
      "loss": 1.0821,
      "step": 2900
    },
    {
      "epoch": 2.388821385176185,
      "grad_norm": 1.4252384901046753,
      "learning_rate": 8.282819383259911e-05,
      "loss": 1.0573,
      "step": 2950
    },
    {
      "epoch": 2.429323612798704,
      "grad_norm": 1.2431281805038452,
      "learning_rate": 8.238766519823788e-05,
      "loss": 1.0453,
      "step": 3000
    },
    {
      "epoch": 2.429323612798704,
      "eval_bleu": 32.711090146730186,
      "eval_chrf": 49.657307772440426,
      "eval_loss": 1.1696988344192505,
      "eval_runtime": 477.0319,
      "eval_samples_per_second": 4.599,
      "eval_steps_per_second": 0.576,
      "step": 3000
    },
    {
      "epoch": 2.469825840421223,
      "grad_norm": 1.0499687194824219,
      "learning_rate": 8.194713656387666e-05,
      "loss": 1.0526,
      "step": 3050
    },
    {
      "epoch": 2.5103280680437425,
      "grad_norm": 1.279381513595581,
      "learning_rate": 8.150660792951542e-05,
      "loss": 1.0419,
      "step": 3100
    },
    {
      "epoch": 2.5508302956662616,
      "grad_norm": 1.122578501701355,
      "learning_rate": 8.106607929515418e-05,
      "loss": 1.0498,
      "step": 3150
    },
    {
      "epoch": 2.5913325232887807,
      "grad_norm": 1.1935076713562012,
      "learning_rate": 8.062555066079296e-05,
      "loss": 1.0293,
      "step": 3200
    },
    {
      "epoch": 2.6318347509113003,
      "grad_norm": 1.2038148641586304,
      "learning_rate": 8.018502202643172e-05,
      "loss": 0.9951,
      "step": 3250
    },
    {
      "epoch": 2.6723369785338194,
      "grad_norm": 1.0469157695770264,
      "learning_rate": 7.974449339207049e-05,
      "loss": 1.0359,
      "step": 3300
    },
    {
      "epoch": 2.7128392061563384,
      "grad_norm": 1.1141313314437866,
      "learning_rate": 7.930396475770926e-05,
      "loss": 1.0082,
      "step": 3350
    },
    {
      "epoch": 2.753341433778858,
      "grad_norm": 1.1356587409973145,
      "learning_rate": 7.886343612334803e-05,
      "loss": 1.0567,
      "step": 3400
    },
    {
      "epoch": 2.793843661401377,
      "grad_norm": 1.2304331064224243,
      "learning_rate": 7.842290748898679e-05,
      "loss": 1.0109,
      "step": 3450
    },
    {
      "epoch": 2.834345889023896,
      "grad_norm": 1.166832685470581,
      "learning_rate": 7.798237885462555e-05,
      "loss": 1.0364,
      "step": 3500
    },
    {
      "epoch": 2.8748481166464157,
      "grad_norm": 1.1052038669586182,
      "learning_rate": 7.754185022026432e-05,
      "loss": 1.0407,
      "step": 3550
    },
    {
      "epoch": 2.915350344268935,
      "grad_norm": 1.0822720527648926,
      "learning_rate": 7.710132158590308e-05,
      "loss": 1.0361,
      "step": 3600
    },
    {
      "epoch": 2.955852571891454,
      "grad_norm": 0.9728385210037231,
      "learning_rate": 7.666079295154185e-05,
      "loss": 1.0417,
      "step": 3650
    },
    {
      "epoch": 2.9963547995139734,
      "grad_norm": 1.2478103637695312,
      "learning_rate": 7.622026431718062e-05,
      "loss": 1.0308,
      "step": 3700
    },
    {
      "epoch": 3.0364520048602675,
      "grad_norm": 1.046961784362793,
      "learning_rate": 7.577973568281939e-05,
      "loss": 0.9011,
      "step": 3750
    },
    {
      "epoch": 3.0769542324827865,
      "grad_norm": 1.0165048837661743,
      "learning_rate": 7.533920704845815e-05,
      "loss": 0.8754,
      "step": 3800
    },
    {
      "epoch": 3.1174564601053056,
      "grad_norm": 1.2605644464492798,
      "learning_rate": 7.489867841409693e-05,
      "loss": 0.8747,
      "step": 3850
    },
    {
      "epoch": 3.157958687727825,
      "grad_norm": 1.2887046337127686,
      "learning_rate": 7.445814977973569e-05,
      "loss": 0.863,
      "step": 3900
    },
    {
      "epoch": 3.1984609153503443,
      "grad_norm": 0.9447133541107178,
      "learning_rate": 7.401762114537445e-05,
      "loss": 0.8687,
      "step": 3950
    },
    {
      "epoch": 3.2389631429728634,
      "grad_norm": 1.0172377824783325,
      "learning_rate": 7.357709251101322e-05,
      "loss": 0.8843,
      "step": 4000
    },
    {
      "epoch": 3.2389631429728634,
      "eval_bleu": 33.64125387109404,
      "eval_chrf": 50.44642474127626,
      "eval_loss": 1.1550147533416748,
      "eval_runtime": 480.3333,
      "eval_samples_per_second": 4.568,
      "eval_steps_per_second": 0.573,
      "step": 4000
    }
  ],
  "logging_steps": 50,
  "max_steps": 12350,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 1000,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.356082635276288e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}