{
  "best_global_step": 5250,
  "best_metric": 0.9456975772765246,
  "best_model_checkpoint": "outputs/modernbert-conll2003-ner/checkpoint-5250",
  "epoch": 3.0,
  "eval_steps": 150,
  "global_step": 5268,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02847380410022779,
      "grad_norm": 9.695123672485352,
      "learning_rate": 4.648956356736243e-06,
      "loss": 1.6757090759277344,
      "step": 50
    },
    {
      "epoch": 0.05694760820045558,
      "grad_norm": 9.274518013000488,
      "learning_rate": 9.392789373814043e-06,
      "loss": 0.5932406234741211,
      "step": 100
    },
    {
      "epoch": 0.08542141230068337,
      "grad_norm": 4.3831586837768555,
      "learning_rate": 1.413662239089184e-05,
      "loss": 0.3480891418457031,
      "step": 150
    },
    {
      "epoch": 0.08542141230068337,
      "eval_accuracy": 0.923675950574119,
      "eval_f1": 0.5273540055346626,
      "eval_loss": 0.2841052711009979,
      "eval_precision": 0.4558949822107717,
      "eval_recall": 0.6253786603837092,
      "eval_runtime": 31.3588,
      "eval_samples_per_second": 103.639,
      "eval_steps_per_second": 12.979,
      "step": 150
    },
    {
      "epoch": 0.11389521640091116,
      "grad_norm": 2.5049362182617188,
      "learning_rate": 1.888045540796964e-05,
      "loss": 0.2505049324035645,
      "step": 200
    },
    {
      "epoch": 0.14236902050113895,
      "grad_norm": 84.79695892333984,
      "learning_rate": 2.362428842504744e-05,
      "loss": 0.19772991180419922,
      "step": 250
    },
    {
      "epoch": 0.17084282460136674,
      "grad_norm": 3.9303884506225586,
      "learning_rate": 2.836812144212524e-05,
      "loss": 0.1767332649230957,
      "step": 300
    },
    {
      "epoch": 0.17084282460136674,
      "eval_accuracy": 0.9527725906207372,
      "eval_f1": 0.7449722200485172,
      "eval_loss": 0.18341314792633057,
      "eval_precision": 0.6962117887962557,
      "eval_recall": 0.8010770784247728,
      "eval_runtime": 29.9644,
      "eval_samples_per_second": 108.462,
      "eval_steps_per_second": 13.583,
      "step": 300
    },
    {
      "epoch": 0.19931662870159453,
      "grad_norm": 3.788025379180908,
      "learning_rate": 3.311195445920304e-05,
      "loss": 0.21780847549438476,
      "step": 350
    },
    {
      "epoch": 0.22779043280182232,
      "grad_norm": 2.6969878673553467,
      "learning_rate": 3.7855787476280835e-05,
      "loss": 0.18941253662109375,
      "step": 400
    },
    {
      "epoch": 0.25626423690205014,
      "grad_norm": 2.7121334075927734,
      "learning_rate": 4.2599620493358636e-05,
      "loss": 0.15380273818969725,
      "step": 450
    },
    {
      "epoch": 0.25626423690205014,
      "eval_accuracy": 0.9678175533394254,
      "eval_f1": 0.8266303477006889,
      "eval_loss": 0.1342710554599762,
      "eval_precision": 0.7888379204892967,
      "eval_recall": 0.8682261864692022,
      "eval_runtime": 29.1886,
      "eval_samples_per_second": 111.345,
      "eval_steps_per_second": 13.944,
      "step": 450
    },
    {
      "epoch": 0.2847380410022779,
      "grad_norm": 5.127914905548096,
      "learning_rate": 4.7343453510436436e-05,
      "loss": 0.16923311233520508,
      "step": 500
    },
    {
      "epoch": 0.3132118451025057,
      "grad_norm": 0.5341155529022217,
      "learning_rate": 4.976798143851508e-05,
      "loss": 0.14784460067749022,
      "step": 550
    },
    {
      "epoch": 0.3416856492027335,
      "grad_norm": 4.445528507232666,
      "learning_rate": 4.924066652604936e-05,
      "loss": 0.12438019752502441,
      "step": 600
    },
    {
      "epoch": 0.3416856492027335,
      "eval_accuracy": 0.9660428834412703,
      "eval_f1": 0.830950650189807,
      "eval_loss": 0.1361789107322693,
      "eval_precision": 0.7988818139462649,
      "eval_recall": 0.865701783911141,
      "eval_runtime": 29.0808,
      "eval_samples_per_second": 111.758,
      "eval_steps_per_second": 13.995,
      "step": 600
    },
    {
      "epoch": 0.3701594533029613,
      "grad_norm": 6.205238342285156,
      "learning_rate": 4.871335161358363e-05,
      "loss": 0.10112229347229004,
      "step": 650
    },
    {
      "epoch": 0.39863325740318906,
      "grad_norm": 1.6438570022583008,
      "learning_rate": 4.818603670111791e-05,
      "loss": 0.105415678024292,
      "step": 700
    },
    {
      "epoch": 0.4271070615034169,
      "grad_norm": 0.5931701064109802,
      "learning_rate": 4.7658721788652186e-05,
      "loss": 0.1022566795349121,
      "step": 750
    },
    {
      "epoch": 0.4271070615034169,
      "eval_accuracy": 0.9678837723654761,
      "eval_f1": 0.8410666231754057,
      "eval_loss": 0.1157219260931015,
      "eval_precision": 0.8158519221642145,
      "eval_recall": 0.8678895994614608,
      "eval_runtime": 29.5173,
      "eval_samples_per_second": 110.105,
      "eval_steps_per_second": 13.789,
      "step": 750
    },
    {
      "epoch": 0.45558086560364464,
      "grad_norm": 10.889516830444336,
      "learning_rate": 4.713140687618646e-05,
      "loss": 0.11253717422485351,
      "step": 800
    },
    {
      "epoch": 0.48405466970387245,
      "grad_norm": 2.053253650665283,
      "learning_rate": 4.660409196372073e-05,
      "loss": 0.1010161018371582,
      "step": 850
    },
    {
      "epoch": 0.5125284738041003,
      "grad_norm": 1.8022727966308594,
      "learning_rate": 4.607677705125501e-05,
      "loss": 0.11372275352478027,
      "step": 900
    },
    {
      "epoch": 0.5125284738041003,
      "eval_accuracy": 0.9745718939965831,
      "eval_f1": 0.8665790338481761,
      "eval_loss": 0.09593585878610611,
      "eval_precision": 0.8465489566613162,
      "eval_recall": 0.8875799394143387,
      "eval_runtime": 29.4183,
      "eval_samples_per_second": 110.476,
      "eval_steps_per_second": 13.835,
      "step": 900
    },
    {
      "epoch": 0.541002277904328,
      "grad_norm": 4.1512250900268555,
      "learning_rate": 4.554946213878929e-05,
      "loss": 0.13236145973205565,
      "step": 950
    },
    {
      "epoch": 0.5694760820045558,
      "grad_norm": 0.27113714814186096,
      "learning_rate": 4.5022147226323564e-05,
      "loss": 0.10169049263000489,
      "step": 1000
    },
    {
      "epoch": 0.5979498861047836,
      "grad_norm": 13.862340927124023,
      "learning_rate": 4.449483231385784e-05,
      "loss": 0.08363029479980469,
      "step": 1050
    },
    {
      "epoch": 0.5979498861047836,
      "eval_accuracy": 0.9776709444157495,
      "eval_f1": 0.8889258120794218,
      "eval_loss": 0.08504071831703186,
      "eval_precision": 0.8777686628383922,
      "eval_recall": 0.9003702457085156,
      "eval_runtime": 29.4995,
      "eval_samples_per_second": 110.171,
      "eval_steps_per_second": 13.797,
      "step": 1050
    },
    {
      "epoch": 0.6264236902050114,
      "grad_norm": 102.99384307861328,
      "learning_rate": 4.396751740139211e-05,
      "loss": 0.060327496528625485,
      "step": 1100
    },
    {
      "epoch": 0.6548974943052391,
      "grad_norm": 0.5466595888137817,
      "learning_rate": 4.344020248892639e-05,
      "loss": 0.09849649429321289,
      "step": 1150
    },
    {
      "epoch": 0.683371298405467,
      "grad_norm": 2.1020138263702393,
      "learning_rate": 4.2912887576460664e-05,
      "loss": 0.09878103256225586,
      "step": 1200
    },
    {
      "epoch": 0.683371298405467,
      "eval_accuracy": 0.979246957235753,
      "eval_f1": 0.8983611984770733,
      "eval_loss": 0.08141425251960754,
      "eval_precision": 0.8838762214983713,
      "eval_recall": 0.9133288455065635,
      "eval_runtime": 30.0152,
      "eval_samples_per_second": 108.278,
      "eval_steps_per_second": 13.56,
      "step": 1200
    },
    {
      "epoch": 0.7118451025056948,
      "grad_norm": 1.834499478340149,
      "learning_rate": 4.238557266399494e-05,
      "loss": 0.07460307598114013,
      "step": 1250
    },
    {
      "epoch": 0.7403189066059226,
      "grad_norm": 2.9035305976867676,
      "learning_rate": 4.185825775152922e-05,
      "loss": 0.0928777027130127,
      "step": 1300
    },
    {
      "epoch": 0.7687927107061503,
      "grad_norm": 2.3792476654052734,
      "learning_rate": 4.133094283906349e-05,
      "loss": 0.0847599983215332,
      "step": 1350
    },
    {
      "epoch": 0.7687927107061503,
      "eval_accuracy": 0.9788099116638193,
      "eval_f1": 0.8985074626865672,
      "eval_loss": 0.08915792405605316,
      "eval_precision": 0.88558352402746,
      "eval_recall": 0.9118142039717267,
      "eval_runtime": 29.7402,
      "eval_samples_per_second": 109.28,
      "eval_steps_per_second": 13.685,
      "step": 1350
    },
    {
      "epoch": 0.7972665148063781,
      "grad_norm": 0.3543700575828552,
      "learning_rate": 4.0803627926597765e-05,
      "loss": 0.07656644344329834,
      "step": 1400
    },
    {
      "epoch": 0.8257403189066059,
      "grad_norm": 2.170388698577881,
      "learning_rate": 4.027631301413204e-05,
      "loss": 0.06705051898956299,
      "step": 1450
    },
    {
      "epoch": 0.8542141230068337,
      "grad_norm": 2.1117851734161377,
      "learning_rate": 3.974899810166632e-05,
      "loss": 0.0851218032836914,
      "step": 1500
    },
    {
      "epoch": 0.8542141230068337,
      "eval_accuracy": 0.9790747877680215,
      "eval_f1": 0.8957797860873892,
      "eval_loss": 0.08403440564870834,
      "eval_precision": 0.8828239908481778,
      "eval_recall": 0.9091215079097947,
      "eval_runtime": 29.7275,
      "eval_samples_per_second": 109.326,
      "eval_steps_per_second": 13.691,
      "step": 1500
    },
    {
      "epoch": 0.8826879271070615,
      "grad_norm": 0.6597453355789185,
      "learning_rate": 3.9221683189200595e-05,
      "loss": 0.09833741188049316,
      "step": 1550
    },
    {
      "epoch": 0.9111617312072893,
      "grad_norm": 3.1620845794677734,
      "learning_rate": 3.8694368276734865e-05,
      "loss": 0.06334121704101563,
      "step": 1600
    },
    {
      "epoch": 0.9396355353075171,
      "grad_norm": 6.5027384757995605,
      "learning_rate": 3.816705336426914e-05,
      "loss": 0.08442768096923828,
      "step": 1650
    },
    {
      "epoch": 0.9396355353075171,
      "eval_accuracy": 0.9838425576436621,
      "eval_f1": 0.9172185430463576,
      "eval_loss": 0.060907382518053055,
      "eval_precision": 0.90257412838058,
      "eval_recall": 0.9323460114439582,
      "eval_runtime": 29.0781,
      "eval_samples_per_second": 111.768,
      "eval_steps_per_second": 13.997,
      "step": 1650
    },
    {
      "epoch": 0.9681093394077449,
      "grad_norm": 0.8580997586250305,
      "learning_rate": 3.763973845180342e-05,
      "loss": 0.062264475822448734,
      "step": 1700
    },
    {
      "epoch": 0.9965831435079726,
      "grad_norm": 1.8275635242462158,
      "learning_rate": 3.7112423539337696e-05,
      "loss": 0.06028622627258301,
      "step": 1750
    },
    {
      "epoch": 1.0250569476082005,
      "grad_norm": 0.7617747187614441,
      "learning_rate": 3.658510862687197e-05,
      "loss": 0.08054101943969727,
      "step": 1800
    },
    {
      "epoch": 1.0250569476082005,
      "eval_accuracy": 0.9795648085607956,
      "eval_f1": 0.9017989767288331,
      "eval_loss": 0.0763646736741066,
      "eval_precision": 0.8847150259067358,
      "eval_recall": 0.9195557051497812,
      "eval_runtime": 29.8561,
      "eval_samples_per_second": 108.856,
      "eval_steps_per_second": 13.632,
      "step": 1800
    },
    {
      "epoch": 1.0535307517084282,
      "grad_norm": 0.37690117955207825,
      "learning_rate": 3.605779371440625e-05,
      "loss": 0.04260779857635498,
      "step": 1850
    },
    {
      "epoch": 1.082004555808656,
      "grad_norm": 0.010842532850801945,
      "learning_rate": 3.553047880194052e-05,
      "loss": 0.04732504367828369,
      "step": 1900
    },
    {
      "epoch": 1.1104783599088839,
      "grad_norm": 0.5702534914016724,
      "learning_rate": 3.5003163889474796e-05,
      "loss": 0.04710710525512695,
      "step": 1950
    },
    {
      "epoch": 1.1104783599088839,
      "eval_accuracy": 0.9815116479266823,
      "eval_f1": 0.911798968214345,
      "eval_loss": 0.07324562221765518,
      "eval_precision": 0.90174456879526,
      "eval_recall": 0.9220801077078424,
      "eval_runtime": 31.9077,
      "eval_samples_per_second": 101.856,
      "eval_steps_per_second": 12.756,
      "step": 1950
    },
    {
      "epoch": 1.1389521640091116,
      "grad_norm": 1.4758163690567017,
      "learning_rate": 3.447584897700907e-05,
      "loss": 0.04873208999633789,
      "step": 2000
    },
    {
      "epoch": 1.1674259681093395,
      "grad_norm": 0.1115245372056961,
      "learning_rate": 3.394853406454335e-05,
      "loss": 0.0670417594909668,
      "step": 2050
    },
    {
      "epoch": 1.1958997722095672,
      "grad_norm": 0.03725523129105568,
      "learning_rate": 3.342121915207762e-05,
      "loss": 0.047343168258666996,
      "step": 2100
    },
    {
      "epoch": 1.1958997722095672,
      "eval_accuracy": 0.9838822890592925,
      "eval_f1": 0.9264925993680359,
      "eval_loss": 0.05988152697682381,
      "eval_precision": 0.915680473372781,
      "eval_recall": 0.9375631100639515,
      "eval_runtime": 29.8404,
      "eval_samples_per_second": 108.913,
      "eval_steps_per_second": 13.639,
      "step": 2100
    },
    {
      "epoch": 1.224373576309795,
      "grad_norm": 0.09743179380893707,
      "learning_rate": 3.2893904239611903e-05,
      "loss": 0.03012749433517456,
      "step": 2150
    },
    {
      "epoch": 1.2528473804100229,
      "grad_norm": 0.12638814747333527,
      "learning_rate": 3.2366589327146174e-05,
      "loss": 0.029717156887054442,
      "step": 2200
    },
    {
      "epoch": 1.2813211845102506,
      "grad_norm": 0.03454405814409256,
      "learning_rate": 3.183927441468045e-05,
      "loss": 0.024935431480407715,
      "step": 2250
    },
    {
      "epoch": 1.2813211845102506,
      "eval_accuracy": 0.9831406359675262,
      "eval_f1": 0.9189278898016762,
      "eval_loss": 0.0698927640914917,
      "eval_precision": 0.9063676542805696,
      "eval_recall": 0.931841130932346,
      "eval_runtime": 30.0056,
      "eval_samples_per_second": 108.313,
      "eval_steps_per_second": 13.564,
      "step": 2250
    },
    {
      "epoch": 1.3097949886104785,
      "grad_norm": 1.2544814348220825,
      "learning_rate": 3.131195950221472e-05,
      "loss": 0.07158967971801758,
      "step": 2300
    },
    {
      "epoch": 1.3382687927107062,
      "grad_norm": 0.010657007806003094,
      "learning_rate": 3.0784644589749004e-05,
      "loss": 0.027039694786071777,
      "step": 2350
    },
    {
      "epoch": 1.366742596810934,
      "grad_norm": 0.13316844403743744,
      "learning_rate": 3.0257329677283274e-05,
      "loss": 0.030671508312225343,
      "step": 2400
    },
    {
      "epoch": 1.366742596810934,
      "eval_accuracy": 0.982650615174752,
      "eval_f1": 0.9172006324373804,
      "eval_loss": 0.06499247997999191,
      "eval_precision": 0.9071604938271605,
      "eval_recall": 0.9274654998317065,
      "eval_runtime": 31.207,
      "eval_samples_per_second": 104.143,
      "eval_steps_per_second": 13.042,
      "step": 2400
    },
    {
      "epoch": 1.3952164009111616,
      "grad_norm": 0.034438204020261765,
      "learning_rate": 2.9730014764817547e-05,
      "loss": 0.04132942199707031,
      "step": 2450
    },
    {
      "epoch": 1.4236902050113895,
      "grad_norm": 0.07990626990795135,
      "learning_rate": 2.9202699852351828e-05,
      "loss": 0.041249847412109374,
      "step": 2500
    },
    {
      "epoch": 1.4521640091116172,
      "grad_norm": 0.997015118598938,
      "learning_rate": 2.86753849398861e-05,
      "loss": 0.03392451286315918,
      "step": 2550
    },
    {
      "epoch": 1.4521640091116172,
      "eval_accuracy": 0.9826373713695419,
      "eval_f1": 0.9231916480238628,
      "eval_loss": 0.07306068390607834,
      "eval_precision": 0.9092541211033132,
      "eval_recall": 0.9375631100639515,
      "eval_runtime": 47.9701,
      "eval_samples_per_second": 67.751,
      "eval_steps_per_second": 8.484,
      "step": 2550
    },
    {
      "epoch": 1.4806378132118452,
      "grad_norm": 0.004969366826117039,
      "learning_rate": 2.8148070027420375e-05,
      "loss": 0.031039011478424073,
      "step": 2600
    },
    {
      "epoch": 1.5091116173120729,
      "grad_norm": 2.834475040435791,
      "learning_rate": 2.7620755114954648e-05,
      "loss": 0.03727980375289917,
      "step": 2650
    },
    {
      "epoch": 1.5375854214123006,
      "grad_norm": 0.2862681746482849,
      "learning_rate": 2.7093440202488928e-05,
      "loss": 0.0418681001663208,
      "step": 2700
    },
    {
      "epoch": 1.5375854214123006,
      "eval_accuracy": 0.9836306567603004,
      "eval_f1": 0.9234719231406328,
      "eval_loss": 0.05933494120836258,
      "eval_precision": 0.9091650358773646,
      "eval_recall": 0.9382362840794345,
      "eval_runtime": 29.3483,
      "eval_samples_per_second": 110.739,
      "eval_steps_per_second": 13.868,
      "step": 2700
    },
    {
      "epoch": 1.5660592255125285,
      "grad_norm": 0.4975425601005554,
      "learning_rate": 2.65661252900232e-05,
      "loss": 0.04010946273803711,
      "step": 2750
    },
    {
      "epoch": 1.5945330296127562,
      "grad_norm": 0.04493750259280205,
      "learning_rate": 2.6038810377557475e-05,
      "loss": 0.04472488403320313,
      "step": 2800
    },
    {
      "epoch": 1.6230068337129842,
      "grad_norm": 2.411900043487549,
      "learning_rate": 2.5511495465091755e-05,
      "loss": 0.03943685531616211,
      "step": 2850
    },
    {
      "epoch": 1.6230068337129842,
      "eval_accuracy": 0.9843855536572768,
      "eval_f1": 0.9303602629170481,
      "eval_loss": 0.055120810866355896,
      "eval_precision": 0.9200263287806484,
      "eval_recall": 0.9409289801413665,
      "eval_runtime": 30.9529,
      "eval_samples_per_second": 104.998,
      "eval_steps_per_second": 13.149,
      "step": 2850
    },
    {
      "epoch": 1.6514806378132119,
      "grad_norm": 0.9190415740013123,
      "learning_rate": 2.498418055262603e-05,
      "loss": 0.03374749422073364,
      "step": 2900
    },
    {
      "epoch": 1.6799544419134396,
      "grad_norm": 0.021526990458369255,
      "learning_rate": 2.4456865640160302e-05,
      "loss": 0.04065659523010254,
      "step": 2950
    },
    {
      "epoch": 1.7084282460136673,
      "grad_norm": 0.052350789308547974,
      "learning_rate": 2.392955072769458e-05,
      "loss": 0.02165827512741089,
      "step": 3000
    },
    {
      "epoch": 1.7084282460136673,
      "eval_accuracy": 0.9858291284251791,
      "eval_f1": 0.9307326831707926,
      "eval_loss": 0.05327470600605011,
      "eval_precision": 0.9220478943022296,
      "eval_recall": 0.9395826321104005,
      "eval_runtime": 29.7224,
      "eval_samples_per_second": 109.345,
      "eval_steps_per_second": 13.693,
      "step": 3000
    },
    {
      "epoch": 1.7369020501138952,
      "grad_norm": 0.7147663831710815,
      "learning_rate": 2.3402235815228856e-05,
      "loss": 0.023224782943725587,
      "step": 3050
    },
    {
      "epoch": 1.7653758542141231,
      "grad_norm": 1.1497533321380615,
      "learning_rate": 2.287492090276313e-05,
      "loss": 0.03412039518356323,
      "step": 3100
    },
    {
      "epoch": 1.7938496583143508,
      "grad_norm": 1.3410314321517944,
      "learning_rate": 2.2347605990297406e-05,
      "loss": 0.03674775123596191,
      "step": 3150
    },
    {
      "epoch": 1.7938496583143508,
      "eval_accuracy": 0.984292847020806,
      "eval_f1": 0.926780508121616,
      "eval_loss": 0.057060979306697845,
      "eval_precision": 0.91753257463302,
      "eval_recall": 0.9362167620329855,
      "eval_runtime": 30.2221,
      "eval_samples_per_second": 107.537,
      "eval_steps_per_second": 13.467,
      "step": 3150
    },
    {
      "epoch": 1.8223234624145785,
      "grad_norm": 0.02126573957502842,
      "learning_rate": 2.182029107783168e-05,
      "loss": 0.0374279522895813,
      "step": 3200
    },
    {
      "epoch": 1.8507972665148062,
      "grad_norm": 0.24427224695682526,
      "learning_rate": 2.1292976165365956e-05,
      "loss": 0.02806992769241333,
      "step": 3250
    },
    {
      "epoch": 1.8792710706150342,
      "grad_norm": 0.07660132646560669,
      "learning_rate": 2.0765661252900233e-05,
      "loss": 0.05080529689788818,
      "step": 3300
    },
    {
      "epoch": 1.8792710706150342,
      "eval_accuracy": 0.9855377647105567,
      "eval_f1": 0.9361275778575603,
      "eval_loss": 0.05377225950360298,
      "eval_precision": 0.9289146644573322,
      "eval_recall": 0.9434533826994278,
      "eval_runtime": 29.7763,
      "eval_samples_per_second": 109.147,
      "eval_steps_per_second": 13.669,
      "step": 3300
    },
    {
      "epoch": 1.907744874715262,
      "grad_norm": 0.6391944885253906,
      "learning_rate": 2.0238346340434507e-05,
      "loss": 0.02515883207321167,
      "step": 3350
    },
    {
      "epoch": 1.9362186788154898,
      "grad_norm": 0.018758827820420265,
      "learning_rate": 1.9711031427968783e-05,
      "loss": 0.02636443853378296,
      "step": 3400
    },
    {
      "epoch": 1.9646924829157175,
      "grad_norm": 0.05743739381432533,
      "learning_rate": 1.918371651550306e-05,
      "loss": 0.022034635543823244,
      "step": 3450
    },
    {
      "epoch": 1.9646924829157175,
      "eval_accuracy": 0.9856702027626578,
      "eval_f1": 0.9401037309687134,
      "eval_loss": 0.058265428990125656,
      "eval_precision": 0.9346307385229541,
      "eval_recall": 0.9456411982497476,
      "eval_runtime": 29.4144,
      "eval_samples_per_second": 110.49,
      "eval_steps_per_second": 13.837,
      "step": 3450
    },
    {
      "epoch": 1.9931662870159452,
      "grad_norm": 0.42121005058288574,
      "learning_rate": 1.8656401603037334e-05,
      "loss": 0.016365668773651122,
      "step": 3500
    },
    {
      "epoch": 2.021640091116173,
      "grad_norm": 0.6595817804336548,
      "learning_rate": 1.812908669057161e-05,
      "loss": 0.019083222150802614,
      "step": 3550
    },
    {
      "epoch": 2.050113895216401,
      "grad_norm": 0.06287907063961029,
      "learning_rate": 1.7601771778105884e-05,
      "loss": 0.010235008001327515,
      "step": 3600
    },
    {
      "epoch": 2.050113895216401,
      "eval_accuracy": 0.9864780748804747,
      "eval_f1": 0.9406255226626526,
      "eval_loss": 0.05760456249117851,
      "eval_precision": 0.9348404255319149,
      "eval_recall": 0.9464826657691013,
      "eval_runtime": 30.1447,
      "eval_samples_per_second": 107.813,
      "eval_steps_per_second": 13.502,
      "step": 3600
    },
    {
      "epoch": 2.078587699316629,
      "grad_norm": 2.4363925457000732,
      "learning_rate": 1.707445686564016e-05,
      "loss": 0.011650030612945556,
      "step": 3650
    },
    {
      "epoch": 2.1070615034168565,
      "grad_norm": 0.015011913143098354,
      "learning_rate": 1.6547141953174438e-05,
      "loss": 0.011951944828033446,
      "step": 3700
    },
    {
      "epoch": 2.135535307517084,
      "grad_norm": 0.10372404754161835,
      "learning_rate": 1.601982704070871e-05,
      "loss": 0.005397456288337707,
      "step": 3750
    },
    {
      "epoch": 2.135535307517084,
      "eval_accuracy": 0.9865575377117354,
      "eval_f1": 0.9432321712231418,
      "eval_loss": 0.05662254989147186,
      "eval_precision": 0.9371988702442267,
      "eval_recall": 0.9493436553349041,
      "eval_runtime": 30.349,
      "eval_samples_per_second": 107.088,
      "eval_steps_per_second": 13.411,
      "step": 3750
    },
    {
      "epoch": 2.164009111617312,
      "grad_norm": 0.01086343452334404,
      "learning_rate": 1.5492512128242988e-05,
      "loss": 0.012290499210357665,
      "step": 3800
    },
    {
      "epoch": 2.19248291571754,
      "grad_norm": 0.5641162991523743,
      "learning_rate": 1.4965197215777263e-05,
      "loss": 0.009672535061836242,
      "step": 3850
    },
    {
      "epoch": 2.2209567198177678,
      "grad_norm": 0.325865238904953,
      "learning_rate": 1.443788230331154e-05,
      "loss": 0.011955568790435791,
      "step": 3900
    },
    {
      "epoch": 2.2209567198177678,
      "eval_accuracy": 0.9867561947898871,
      "eval_f1": 0.9433110367892976,
      "eval_loss": 0.06146744638681412,
      "eval_precision": 0.9373546028580924,
      "eval_recall": 0.9493436553349041,
      "eval_runtime": 29.9391,
      "eval_samples_per_second": 108.554,
      "eval_steps_per_second": 13.594,
      "step": 3900
    },
    {
      "epoch": 2.2494305239179955,
      "grad_norm": 0.021459011361002922,
      "learning_rate": 1.3910567390845813e-05,
      "loss": 0.024674110412597657,
      "step": 3950
    },
    {
      "epoch": 2.277904328018223,
      "grad_norm": 0.930016279220581,
      "learning_rate": 1.338325247838009e-05,
      "loss": 0.01294115662574768,
      "step": 4000
    },
    {
      "epoch": 2.306378132118451,
      "grad_norm": 0.013619116507470608,
      "learning_rate": 1.2855937565914363e-05,
      "loss": 0.013077363967895508,
      "step": 4050
    },
    {
      "epoch": 2.306378132118451,
      "eval_accuracy": 0.9868621452315679,
      "eval_f1": 0.9398132710903635,
      "eval_loss": 0.055142417550086975,
      "eval_precision": 0.9311199207135779,
      "eval_recall": 0.9486704813194211,
      "eval_runtime": 30.2311,
      "eval_samples_per_second": 107.505,
      "eval_steps_per_second": 13.463,
      "step": 4050
    },
    {
      "epoch": 2.334851936218679,
      "grad_norm": 0.2678438723087311,
      "learning_rate": 1.232862265344864e-05,
      "loss": 0.015998460054397583,
      "step": 4100
    },
    {
      "epoch": 2.3633257403189067,
      "grad_norm": 0.016241000965237617,
      "learning_rate": 1.1801307740982915e-05,
      "loss": 0.008549771904945373,
      "step": 4150
    },
    {
      "epoch": 2.3917995444191344,
      "grad_norm": 0.4150155186653137,
      "learning_rate": 1.127399282851719e-05,
      "loss": 0.009619292616844178,
      "step": 4200
    },
    {
      "epoch": 2.3917995444191344,
      "eval_accuracy": 0.9870740461149298,
      "eval_f1": 0.9395124386388218,
      "eval_loss": 0.058261238038539886,
      "eval_precision": 0.9290768471285173,
      "eval_recall": 0.9501851228542578,
      "eval_runtime": 29.2949,
      "eval_samples_per_second": 110.941,
      "eval_steps_per_second": 13.893,
      "step": 4200
    },
    {
      "epoch": 2.420273348519362,
      "grad_norm": 0.017043378204107285,
      "learning_rate": 1.0746677916051466e-05,
      "loss": 0.01029258370399475,
      "step": 4250
    },
    {
      "epoch": 2.44874715261959,
      "grad_norm": 1.0776268243789673,
      "learning_rate": 1.021936300358574e-05,
      "loss": 0.007834370136260987,
      "step": 4300
    },
    {
      "epoch": 2.477220956719818,
      "grad_norm": 0.009231849573552608,
      "learning_rate": 9.692048091120018e-06,
      "loss": 0.008277364373207093,
      "step": 4350
    },
    {
      "epoch": 2.477220956719818,
      "eval_accuracy": 0.9869813394784589,
      "eval_f1": 0.939451986341301,
      "eval_loss": 0.06197329983115196,
      "eval_precision": 0.9299258037922506,
      "eval_recall": 0.9491753618310333,
      "eval_runtime": 30.1358,
      "eval_samples_per_second": 107.845,
      "eval_steps_per_second": 13.506,
      "step": 4350
    },
    {
      "epoch": 2.5056947608200457,
      "grad_norm": 0.030698262155056,
      "learning_rate": 9.164733178654293e-06,
      "loss": 0.017403430938720703,
      "step": 4400
    },
    {
      "epoch": 2.5341685649202734,
      "grad_norm": 0.010924355126917362,
      "learning_rate": 8.637418266188568e-06,
      "loss": 0.011218159198760987,
      "step": 4450
    },
    {
      "epoch": 2.562642369020501,
      "grad_norm": 0.021999873220920563,
      "learning_rate": 8.110103353722843e-06,
      "loss": 0.011795709133148194,
      "step": 4500
    },
    {
      "epoch": 2.562642369020501,
      "eval_accuracy": 0.9869945832836691,
      "eval_f1": 0.9411666527580739,
      "eval_loss": 0.05743277072906494,
      "eval_precision": 0.9334547260387354,
      "eval_recall": 0.9490070683271625,
      "eval_runtime": 29.728,
      "eval_samples_per_second": 109.325,
      "eval_steps_per_second": 13.691,
      "step": 4500
    },
    {
      "epoch": 2.591116173120729,
      "grad_norm": 0.014707831665873528,
      "learning_rate": 7.582788441257119e-06,
      "loss": 0.01068550705909729,
      "step": 4550
    },
    {
      "epoch": 2.619589977220957,
      "grad_norm": 0.0033975038677453995,
      "learning_rate": 7.055473528791394e-06,
      "loss": 0.005157849192619324,
      "step": 4600
    },
    {
      "epoch": 2.6480637813211843,
      "grad_norm": 0.020954320207238197,
      "learning_rate": 6.528158616325671e-06,
      "loss": 0.019545166492462157,
      "step": 4650
    },
    {
      "epoch": 2.6480637813211843,
      "eval_accuracy": 0.9871270213357702,
      "eval_f1": 0.9429023922647329,
      "eval_loss": 0.056006867438554764,
      "eval_precision": 0.9341040462427745,
      "eval_recall": 0.9518680578929654,
      "eval_runtime": 29.1119,
      "eval_samples_per_second": 111.638,
      "eval_steps_per_second": 13.981,
      "step": 4650
    },
    {
      "epoch": 2.6765375854214124,
      "grad_norm": 0.014844182878732681,
      "learning_rate": 6.000843703859946e-06,
      "loss": 0.009334412813186645,
      "step": 4700
    },
    {
      "epoch": 2.70501138952164,
      "grad_norm": 0.011511904187500477,
      "learning_rate": 5.473528791394221e-06,
      "loss": 0.015157132148742676,
      "step": 4750
    },
    {
      "epoch": 2.733485193621868,
      "grad_norm": 0.0025674111675471067,
      "learning_rate": 4.946213878928496e-06,
      "loss": 0.005306068658828735,
      "step": 4800
    },
    {
      "epoch": 2.733485193621868,
      "eval_accuracy": 0.9876435297389646,
      "eval_f1": 0.9446531430002504,
      "eval_loss": 0.05035378783941269,
      "eval_precision": 0.9372204737452376,
      "eval_recall": 0.9522046449007069,
      "eval_runtime": 28.9709,
      "eval_samples_per_second": 112.182,
      "eval_steps_per_second": 14.049,
      "step": 4800
    },
    {
      "epoch": 2.7619589977220955,
      "grad_norm": 0.09001180529594421,
      "learning_rate": 4.418898966462772e-06,
      "loss": 0.007939810752868653,
      "step": 4850
    },
    {
      "epoch": 2.7904328018223232,
      "grad_norm": 0.009678197093307972,
      "learning_rate": 3.891584053997047e-06,
      "loss": 0.008102912306785583,
      "step": 4900
    },
    {
      "epoch": 2.8189066059225514,
      "grad_norm": 0.0023103775456547737,
      "learning_rate": 3.3642691415313225e-06,
      "loss": 0.008059802651405334,
      "step": 4950
    },
    {
      "epoch": 2.8189066059225514,
      "eval_accuracy": 0.9875243354920736,
      "eval_f1": 0.9437771104437771,
      "eval_loss": 0.051680199801921844,
      "eval_precision": 0.9356599404565001,
      "eval_recall": 0.9520363513968361,
      "eval_runtime": 29.0504,
      "eval_samples_per_second": 111.875,
      "eval_steps_per_second": 14.01,
      "step": 4950
    },
    {
      "epoch": 2.847380410022779,
      "grad_norm": 0.011355778202414513,
      "learning_rate": 2.836954229065598e-06,
      "loss": 0.018458174467086794,
      "step": 5000
    },
    {
      "epoch": 2.875854214123007,
      "grad_norm": 0.5875220894813538,
      "learning_rate": 2.3096393165998736e-06,
      "loss": 0.008523634672164916,
      "step": 5050
    },
    {
      "epoch": 2.9043280182232345,
      "grad_norm": 0.04093571752309799,
      "learning_rate": 1.782324404134149e-06,
      "loss": 0.013776706457138061,
      "step": 5100
    },
    {
      "epoch": 2.9043280182232345,
      "eval_accuracy": 0.9878156992066961,
      "eval_f1": 0.945484949832776,
      "eval_loss": 0.049608755856752396,
      "eval_precision": 0.939514788966434,
      "eval_recall": 0.9515314708852238,
      "eval_runtime": 29.3226,
      "eval_samples_per_second": 110.836,
      "eval_steps_per_second": 13.88,
      "step": 5100
    },
    {
      "epoch": 2.932801822323462,
      "grad_norm": 1.3949384689331055,
      "learning_rate": 1.2550094916684245e-06,
      "loss": 0.010595662593841553,
      "step": 5150
    },
    {
      "epoch": 2.9612756264236904,
      "grad_norm": 3.866323232650757,
      "learning_rate": 7.276945792026999e-07,
      "loss": 0.012341983318328857,
      "step": 5200
    },
    {
      "epoch": 2.989749430523918,
      "grad_norm": 0.03131106495857239,
      "learning_rate": 2.0037966673697531e-07,
      "loss": 0.009559857845306396,
      "step": 5250
    },
    {
      "epoch": 2.989749430523918,
      "eval_accuracy": 0.9879613810640073,
      "eval_f1": 0.9456975772765246,
      "eval_loss": 0.048417747020721436,
      "eval_precision": 0.9389515593895156,
      "eval_recall": 0.9525412319084483,
      "eval_runtime": 29.2374,
      "eval_samples_per_second": 111.159,
      "eval_steps_per_second": 13.921,
      "step": 5250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9879746248692174,
      "eval_f1": 0.9456185782307241,
      "eval_loss": 0.04841749370098114,
      "eval_precision": 0.9387958202023553,
      "eval_recall": 0.9525412319084483,
      "eval_runtime": 29.7055,
      "eval_samples_per_second": 109.407,
      "eval_steps_per_second": 13.701,
      "step": 5268
    },
    {
      "epoch": 3.0,
      "step": 5268,
      "total_flos": 4273154449473294.0,
      "train_loss": 0.07650348544686512,
      "train_runtime": 2897.362,
      "train_samples_per_second": 14.538,
      "train_steps_per_second": 1.818
    },
    {
      "epoch": 3.0,
      "step": 5268,
      "test_accuracy": 0.9743404513537723,
      "test_f1": 0.8968703687559934,
      "test_loss": 0.17622677981853485,
      "test_precision": 0.8833934398076593,
      "test_recall": 0.9107648725212465,
      "test_runtime": 30.0194,
      "test_samples_per_second": 115.026,
      "test_steps_per_second": 14.391
    }
  ],
  "logging_steps": 50,
  "max_steps": 5268,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 150,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4273154449473294.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}