{
  "best_metric": 0.0490318201482296,
  "best_model_checkpoint": "/home1/datahome/villien/project_hub/DinoVdeau/models/Kamoulox-large-2024_10_31-batch-size64_freeze_monolabel/checkpoint-387504",
  "epoch": 150.0,
  "eval_steps": 500,
  "global_step": 403650,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18580453363062058,
      "grad_norm": 2.81394624710083,
      "learning_rate": 0.001,
      "loss": 1.1787,
      "step": 500
    },
    {
      "epoch": 0.37160906726124115,
      "grad_norm": 2.4735236167907715,
      "learning_rate": 0.001,
      "loss": 0.98,
      "step": 1000
    },
    {
      "epoch": 0.5574136008918618,
      "grad_norm": 1.8667224645614624,
      "learning_rate": 0.001,
      "loss": 0.9449,
      "step": 1500
    },
    {
      "epoch": 0.7432181345224823,
      "grad_norm": 1.7729698419570923,
      "learning_rate": 0.001,
      "loss": 0.9107,
      "step": 2000
    },
    {
      "epoch": 0.929022668153103,
      "grad_norm": 1.785703182220459,
      "learning_rate": 0.001,
      "loss": 0.9079,
      "step": 2500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7326527412414418,
      "eval_f1_macro": 0.2723318326456157,
      "eval_f1_micro": 0.7326527412414418,
      "eval_loss": 0.8028618097305298,
      "eval_runtime": 507.3076,
      "eval_samples_per_second": 113.148,
      "eval_steps_per_second": 1.768,
      "learning_rate": 0.001,
      "step": 2691
    },
    {
      "epoch": 1.1148272017837235,
      "grad_norm": 1.4361521005630493,
      "learning_rate": 0.001,
      "loss": 0.9035,
      "step": 3000
    },
    {
      "epoch": 1.3006317354143442,
      "grad_norm": 1.2895761728286743,
      "learning_rate": 0.001,
      "loss": 0.8945,
      "step": 3500
    },
    {
      "epoch": 1.4864362690449646,
      "grad_norm": 1.5815945863723755,
      "learning_rate": 0.001,
      "loss": 0.8879,
      "step": 4000
    },
    {
      "epoch": 1.6722408026755853,
      "grad_norm": 1.3825953006744385,
      "learning_rate": 0.001,
      "loss": 0.8901,
      "step": 4500
    },
    {
      "epoch": 1.8580453363062057,
      "grad_norm": 1.0456748008728027,
      "learning_rate": 0.001,
      "loss": 0.8824,
      "step": 5000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7288723192975732,
      "eval_f1_macro": 0.290650504066453,
      "eval_f1_micro": 0.7288723192975732,
      "eval_loss": 0.8038854002952576,
      "eval_runtime": 507.5267,
      "eval_samples_per_second": 113.099,
      "eval_steps_per_second": 1.767,
      "learning_rate": 0.001,
      "step": 5382
    },
    {
      "epoch": 2.0438498699368264,
      "grad_norm": 1.1838769912719727,
      "learning_rate": 0.001,
      "loss": 0.8883,
      "step": 5500
    },
    {
      "epoch": 2.229654403567447,
      "grad_norm": 1.2504063844680786,
      "learning_rate": 0.001,
      "loss": 0.8713,
      "step": 6000
    },
    {
      "epoch": 2.4154589371980677,
      "grad_norm": 1.2367465496063232,
      "learning_rate": 0.001,
      "loss": 0.8738,
      "step": 6500
    },
    {
      "epoch": 2.6012634708286884,
      "grad_norm": 1.2117840051651,
      "learning_rate": 0.001,
      "loss": 0.8664,
      "step": 7000
    },
    {
      "epoch": 2.787068004459309,
      "grad_norm": 0.9993311762809753,
      "learning_rate": 0.001,
      "loss": 0.8667,
      "step": 7500
    },
    {
      "epoch": 2.9728725380899292,
      "grad_norm": 1.0402483940124512,
      "learning_rate": 0.001,
      "loss": 0.8655,
      "step": 8000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7408581732025574,
      "eval_f1_macro": 0.326985383349658,
      "eval_f1_micro": 0.7408581732025574,
      "eval_loss": 0.7705450654029846,
      "eval_runtime": 518.2697,
      "eval_samples_per_second": 110.755,
      "eval_steps_per_second": 1.731,
      "learning_rate": 0.001,
      "step": 8073
    },
    {
      "epoch": 3.15867707172055,
      "grad_norm": 1.2037022113800049,
      "learning_rate": 0.001,
      "loss": 0.8485,
      "step": 8500
    },
    {
      "epoch": 3.3444816053511706,
      "grad_norm": 1.2167097330093384,
      "learning_rate": 0.001,
      "loss": 0.8707,
      "step": 9000
    },
    {
      "epoch": 3.530286138981791,
      "grad_norm": 1.150800108909607,
      "learning_rate": 0.001,
      "loss": 0.8543,
      "step": 9500
    },
    {
      "epoch": 3.716090672612412,
      "grad_norm": 0.8481634855270386,
      "learning_rate": 0.001,
      "loss": 0.848,
      "step": 10000
    },
    {
      "epoch": 3.901895206243032,
      "grad_norm": 0.9482496380805969,
      "learning_rate": 0.001,
      "loss": 0.8514,
      "step": 10500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7417118168673019,
      "eval_f1_macro": 0.310238611675961,
      "eval_f1_micro": 0.7417118168673019,
      "eval_loss": 0.7623223066329956,
      "eval_runtime": 526.8057,
      "eval_samples_per_second": 108.96,
      "eval_steps_per_second": 1.703,
      "learning_rate": 0.001,
      "step": 10764
    },
    {
      "epoch": 4.087699739873653,
      "grad_norm": 0.9820572137832642,
      "learning_rate": 0.001,
      "loss": 0.8542,
      "step": 11000
    },
    {
      "epoch": 4.273504273504273,
      "grad_norm": 1.2227846384048462,
      "learning_rate": 0.001,
      "loss": 0.8536,
      "step": 11500
    },
    {
      "epoch": 4.459308807134894,
      "grad_norm": 0.9903749823570251,
      "learning_rate": 0.001,
      "loss": 0.8473,
      "step": 12000
    },
    {
      "epoch": 4.645113340765515,
      "grad_norm": 0.802663266658783,
      "learning_rate": 0.001,
      "loss": 0.8475,
      "step": 12500
    },
    {
      "epoch": 4.830917874396135,
      "grad_norm": 0.8765626549720764,
      "learning_rate": 0.001,
      "loss": 0.844,
      "step": 13000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7383495061061653,
      "eval_f1_macro": 0.3107947240093857,
      "eval_f1_micro": 0.7383495061061653,
      "eval_loss": 0.7626621127128601,
      "eval_runtime": 531.6556,
      "eval_samples_per_second": 107.967,
      "eval_steps_per_second": 1.687,
      "learning_rate": 0.001,
      "step": 13455
    },
    {
      "epoch": 5.016722408026756,
      "grad_norm": 0.846638560295105,
      "learning_rate": 0.001,
      "loss": 0.8519,
      "step": 13500
    },
    {
      "epoch": 5.202526941657377,
      "grad_norm": 0.9798027276992798,
      "learning_rate": 0.001,
      "loss": 0.8477,
      "step": 14000
    },
    {
      "epoch": 5.388331475287997,
      "grad_norm": 1.171106219291687,
      "learning_rate": 0.001,
      "loss": 0.8377,
      "step": 14500
    },
    {
      "epoch": 5.574136008918618,
      "grad_norm": 5.736990451812744,
      "learning_rate": 0.001,
      "loss": 0.8358,
      "step": 15000
    },
    {
      "epoch": 5.759940542549238,
      "grad_norm": 0.9148433208465576,
      "learning_rate": 0.001,
      "loss": 0.8361,
      "step": 15500
    },
    {
      "epoch": 5.9457450761798585,
      "grad_norm": 0.9974657297134399,
      "learning_rate": 0.001,
      "loss": 0.8437,
      "step": 16000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7447257016428285,
      "eval_f1_macro": 0.34331624846537584,
      "eval_f1_micro": 0.7447257016428285,
      "eval_loss": 0.7451828122138977,
      "eval_runtime": 532.2923,
      "eval_samples_per_second": 107.837,
      "eval_steps_per_second": 1.685,
      "learning_rate": 0.001,
      "step": 16146
    },
    {
      "epoch": 6.131549609810479,
      "grad_norm": 0.7431060671806335,
      "learning_rate": 0.001,
      "loss": 0.8485,
      "step": 16500
    },
    {
      "epoch": 6.3173541434411,
      "grad_norm": 0.899385392665863,
      "learning_rate": 0.001,
      "loss": 0.8323,
      "step": 17000
    },
    {
      "epoch": 6.5031586770717205,
      "grad_norm": 0.8070719838142395,
      "learning_rate": 0.001,
      "loss": 0.8388,
      "step": 17500
    },
    {
      "epoch": 6.688963210702341,
      "grad_norm": 0.7958774566650391,
      "learning_rate": 0.001,
      "loss": 0.8428,
      "step": 18000
    },
    {
      "epoch": 6.874767744332962,
      "grad_norm": 0.9422007203102112,
      "learning_rate": 0.001,
      "loss": 0.8289,
      "step": 18500
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7467291510600861,
      "eval_f1_macro": 0.3283260141196405,
      "eval_f1_micro": 0.7467291510600861,
      "eval_loss": 0.7458378672599792,
      "eval_runtime": 529.9283,
      "eval_samples_per_second": 108.318,
      "eval_steps_per_second": 1.693,
      "learning_rate": 0.001,
      "step": 18837
    },
    {
      "epoch": 7.060572277963582,
      "grad_norm": 0.8445965051651001,
      "learning_rate": 0.001,
      "loss": 0.8383,
      "step": 19000
    },
    {
      "epoch": 7.246376811594203,
      "grad_norm": 0.9642378687858582,
      "learning_rate": 0.001,
      "loss": 0.8319,
      "step": 19500
    },
    {
      "epoch": 7.432181345224824,
      "grad_norm": 0.9426402449607849,
      "learning_rate": 0.001,
      "loss": 0.8365,
      "step": 20000
    },
    {
      "epoch": 7.617985878855444,
      "grad_norm": 1.0661524534225464,
      "learning_rate": 0.001,
      "loss": 0.8348,
      "step": 20500
    },
    {
      "epoch": 7.803790412486064,
      "grad_norm": 0.7548210620880127,
      "learning_rate": 0.001,
      "loss": 0.8436,
      "step": 21000
    },
    {
      "epoch": 7.989594946116685,
      "grad_norm": 1.0589923858642578,
      "learning_rate": 0.001,
      "loss": 0.8402,
      "step": 21500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7458929286946221,
      "eval_f1_macro": 0.3352756381207012,
      "eval_f1_micro": 0.7458929286946221,
      "eval_loss": 0.7398682832717896,
      "eval_runtime": 542.5535,
      "eval_samples_per_second": 105.798,
      "eval_steps_per_second": 1.653,
      "learning_rate": 0.001,
      "step": 21528
    },
    {
      "epoch": 8.175399479747306,
      "grad_norm": 0.9055599570274353,
      "learning_rate": 0.001,
      "loss": 0.8301,
      "step": 22000
    },
    {
      "epoch": 8.361204013377927,
      "grad_norm": 0.6724838614463806,
      "learning_rate": 0.001,
      "loss": 0.8362,
      "step": 22500
    },
    {
      "epoch": 8.547008547008547,
      "grad_norm": 0.9552667140960693,
      "learning_rate": 0.001,
      "loss": 0.837,
      "step": 23000
    },
    {
      "epoch": 8.732813080639168,
      "grad_norm": 0.7529712319374084,
      "learning_rate": 0.001,
      "loss": 0.8335,
      "step": 23500
    },
    {
      "epoch": 8.918617614269788,
      "grad_norm": 0.9149639010429382,
      "learning_rate": 0.001,
      "loss": 0.8274,
      "step": 24000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7455270814097316,
      "eval_f1_macro": 0.329402464136349,
      "eval_f1_micro": 0.7455270814097316,
      "eval_loss": 0.7424591779708862,
      "eval_runtime": 539.3684,
      "eval_samples_per_second": 106.423,
      "eval_steps_per_second": 1.663,
      "learning_rate": 0.001,
      "step": 24219
    },
    {
      "epoch": 9.104422147900408,
      "grad_norm": 0.9954981207847595,
      "learning_rate": 0.001,
      "loss": 0.8204,
      "step": 24500
    },
    {
      "epoch": 9.29022668153103,
      "grad_norm": 0.8292973041534424,
      "learning_rate": 0.001,
      "loss": 0.8373,
      "step": 25000
    },
    {
      "epoch": 9.47603121516165,
      "grad_norm": 0.6125388741493225,
      "learning_rate": 0.001,
      "loss": 0.834,
      "step": 25500
    },
    {
      "epoch": 9.66183574879227,
      "grad_norm": 0.7945879697799683,
      "learning_rate": 0.001,
      "loss": 0.8523,
      "step": 26000
    },
    {
      "epoch": 9.84764028242289,
      "grad_norm": 0.8813033699989319,
      "learning_rate": 0.001,
      "loss": 0.8289,
      "step": 26500
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7474782669291475,
      "eval_f1_macro": 0.31573051576045374,
      "eval_f1_micro": 0.7474782669291475,
      "eval_loss": 0.7364382147789001,
      "eval_runtime": 565.5402,
      "eval_samples_per_second": 101.498,
      "eval_steps_per_second": 1.586,
      "learning_rate": 0.001,
      "step": 26910
    },
    {
      "epoch": 10.033444816053512,
      "grad_norm": 0.7872824668884277,
      "learning_rate": 0.001,
      "loss": 0.8307,
      "step": 27000
    },
    {
      "epoch": 10.219249349684132,
      "grad_norm": 0.9812045097351074,
      "learning_rate": 0.001,
      "loss": 0.8275,
      "step": 27500
    },
    {
      "epoch": 10.405053883314753,
      "grad_norm": 0.9366200566291809,
      "learning_rate": 0.001,
      "loss": 0.8271,
      "step": 28000
    },
    {
      "epoch": 10.590858416945373,
      "grad_norm": 0.8672580122947693,
      "learning_rate": 0.001,
      "loss": 0.8314,
      "step": 28500
    },
    {
      "epoch": 10.776662950575995,
      "grad_norm": 0.8797865509986877,
      "learning_rate": 0.001,
      "loss": 0.845,
      "step": 29000
    },
    {
      "epoch": 10.962467484206615,
      "grad_norm": 0.708544135093689,
      "learning_rate": 0.001,
      "loss": 0.8368,
      "step": 29500
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.7465375167680005,
      "eval_f1_macro": 0.34419509138343996,
      "eval_f1_micro": 0.7465375167680005,
      "eval_loss": 0.7368418574333191,
      "eval_runtime": 540.8249,
      "eval_samples_per_second": 106.136,
      "eval_steps_per_second": 1.659,
      "learning_rate": 0.001,
      "step": 29601
    },
    {
      "epoch": 11.148272017837236,
      "grad_norm": 0.7683461904525757,
      "learning_rate": 0.001,
      "loss": 0.8328,
      "step": 30000
    },
    {
      "epoch": 11.334076551467856,
      "grad_norm": 0.8067987561225891,
      "learning_rate": 0.001,
      "loss": 0.8367,
      "step": 30500
    },
    {
      "epoch": 11.519881085098476,
      "grad_norm": 0.8982560038566589,
      "learning_rate": 0.001,
      "loss": 0.8333,
      "step": 31000
    },
    {
      "epoch": 11.705685618729097,
      "grad_norm": 0.8005286455154419,
      "learning_rate": 0.001,
      "loss": 0.8264,
      "step": 31500
    },
    {
      "epoch": 11.891490152359717,
      "grad_norm": 0.848416805267334,
      "learning_rate": 0.001,
      "loss": 0.8329,
      "step": 32000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7428093587219735,
      "eval_f1_macro": 0.3321199292959056,
      "eval_f1_micro": 0.7428093587219735,
      "eval_loss": 0.7442134022712708,
      "eval_runtime": 602.0094,
      "eval_samples_per_second": 95.349,
      "eval_steps_per_second": 1.49,
      "learning_rate": 0.001,
      "step": 32292
    },
    {
      "epoch": 12.077294685990339,
      "grad_norm": 0.9192249178886414,
      "learning_rate": 0.001,
      "loss": 0.8242,
      "step": 32500
    },
    {
      "epoch": 12.263099219620958,
      "grad_norm": 0.9414801597595215,
      "learning_rate": 0.001,
      "loss": 0.8211,
      "step": 33000
    },
    {
      "epoch": 12.44890375325158,
      "grad_norm": 0.8428712487220764,
      "learning_rate": 0.001,
      "loss": 0.8252,
      "step": 33500
    },
    {
      "epoch": 12.6347082868822,
      "grad_norm": 0.8568095564842224,
      "learning_rate": 0.001,
      "loss": 0.834,
      "step": 34000
    },
    {
      "epoch": 12.820512820512821,
      "grad_norm": 0.552550196647644,
      "learning_rate": 0.001,
      "loss": 0.8359,
      "step": 34500
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.7479312207104406,
      "eval_f1_macro": 0.35283143267744377,
      "eval_f1_micro": 0.7479312207104406,
      "eval_loss": 0.7384127378463745,
      "eval_runtime": 536.6777,
      "eval_samples_per_second": 106.956,
      "eval_steps_per_second": 1.671,
      "learning_rate": 0.001,
      "step": 34983
    },
    {
      "epoch": 13.006317354143441,
      "grad_norm": 0.790766179561615,
      "learning_rate": 0.001,
      "loss": 0.8317,
      "step": 35000
    },
    {
      "epoch": 13.192121887774062,
      "grad_norm": 0.6483836770057678,
      "learning_rate": 0.001,
      "loss": 0.8142,
      "step": 35500
    },
    {
      "epoch": 13.377926421404682,
      "grad_norm": 0.7646375298500061,
      "learning_rate": 0.001,
      "loss": 0.8336,
      "step": 36000
    },
    {
      "epoch": 13.563730955035304,
      "grad_norm": 0.8401319980621338,
      "learning_rate": 0.001,
      "loss": 0.8318,
      "step": 36500
    },
    {
      "epoch": 13.749535488665924,
      "grad_norm": 0.8368563055992126,
      "learning_rate": 0.001,
      "loss": 0.8305,
      "step": 37000
    },
    {
      "epoch": 13.935340022296543,
      "grad_norm": 1.0513092279434204,
      "learning_rate": 0.001,
      "loss": 0.8388,
      "step": 37500
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7463633037751956,
      "eval_f1_macro": 0.33455884660313173,
      "eval_f1_micro": 0.7463633037751956,
      "eval_loss": 0.7464041113853455,
      "eval_runtime": 519.5938,
      "eval_samples_per_second": 110.473,
      "eval_steps_per_second": 1.726,
      "learning_rate": 0.001,
      "step": 37674
    },
    {
      "epoch": 14.121144555927165,
      "grad_norm": 0.9093230962753296,
      "learning_rate": 0.001,
      "loss": 0.8277,
      "step": 38000
    },
    {
      "epoch": 14.306949089557785,
      "grad_norm": 0.7018683552742004,
      "learning_rate": 0.001,
      "loss": 0.8154,
      "step": 38500
    },
    {
      "epoch": 14.492753623188406,
      "grad_norm": 1.060300588607788,
      "learning_rate": 0.001,
      "loss": 0.8342,
      "step": 39000
    },
    {
      "epoch": 14.678558156819026,
      "grad_norm": 0.9249637126922607,
      "learning_rate": 0.001,
      "loss": 0.8382,
      "step": 39500
    },
    {
      "epoch": 14.864362690449648,
      "grad_norm": 0.7552688717842102,
      "learning_rate": 0.001,
      "loss": 0.8306,
      "step": 40000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.7446734377449871,
      "eval_f1_macro": 0.34277495445998946,
      "eval_f1_micro": 0.7446734377449871,
      "eval_loss": 0.7394037842750549,
      "eval_runtime": 526.6425,
      "eval_samples_per_second": 108.994,
      "eval_steps_per_second": 1.703,
      "learning_rate": 0.001,
      "step": 40365
    },
    {
      "epoch": 15.050167224080267,
      "grad_norm": 0.7855104207992554,
      "learning_rate": 0.001,
      "loss": 0.8399,
      "step": 40500
    },
    {
      "epoch": 15.235971757710889,
      "grad_norm": 0.709564745426178,
      "learning_rate": 0.001,
      "loss": 0.8384,
      "step": 41000
    },
    {
      "epoch": 15.421776291341509,
      "grad_norm": 0.7003925442695618,
      "learning_rate": 0.001,
      "loss": 0.8236,
      "step": 41500
    },
    {
      "epoch": 15.607580824972128,
      "grad_norm": 0.7792437672615051,
      "learning_rate": 0.001,
      "loss": 0.8296,
      "step": 42000
    },
    {
      "epoch": 15.79338535860275,
      "grad_norm": 0.8424106240272522,
      "learning_rate": 0.001,
      "loss": 0.8329,
      "step": 42500
    },
    {
      "epoch": 15.97918989223337,
      "grad_norm": 0.8874714374542236,
      "learning_rate": 0.001,
      "loss": 0.8304,
      "step": 43000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.7478789568125991,
      "eval_f1_macro": 0.3506456767847629,
      "eval_f1_micro": 0.7478789568125991,
      "eval_loss": 0.7397111058235168,
      "eval_runtime": 533.4385,
      "eval_samples_per_second": 107.606,
      "eval_steps_per_second": 1.682,
      "learning_rate": 0.001,
      "step": 43056
    },
    {
      "epoch": 16.16499442586399,
      "grad_norm": 0.7367419600486755,
      "learning_rate": 0.0001,
      "loss": 0.8145,
      "step": 43500
    },
    {
      "epoch": 16.35079895949461,
      "grad_norm": 0.9213405251502991,
      "learning_rate": 0.0001,
      "loss": 0.7917,
      "step": 44000
    },
    {
      "epoch": 16.53660349312523,
      "grad_norm": 0.7902110815048218,
      "learning_rate": 0.0001,
      "loss": 0.7909,
      "step": 44500
    },
    {
      "epoch": 16.722408026755854,
      "grad_norm": 0.7533911466598511,
      "learning_rate": 0.0001,
      "loss": 0.8054,
      "step": 45000
    },
    {
      "epoch": 16.908212560386474,
      "grad_norm": 0.6558223962783813,
      "learning_rate": 0.0001,
      "loss": 0.7886,
      "step": 45500
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.7554398007003362,
      "eval_f1_macro": 0.3747287292827094,
      "eval_f1_micro": 0.7554398007003362,
      "eval_loss": 0.7110718488693237,
      "eval_runtime": 519.8124,
      "eval_samples_per_second": 110.426,
      "eval_steps_per_second": 1.726,
      "learning_rate": 0.0001,
      "step": 45747
    },
    {
      "epoch": 17.094017094017094,
      "grad_norm": 0.8171585202217102,
      "learning_rate": 0.0001,
      "loss": 0.7962,
      "step": 46000
    },
    {
      "epoch": 17.279821627647713,
      "grad_norm": 0.9112296104431152,
      "learning_rate": 0.0001,
      "loss": 0.7774,
      "step": 46500
    },
    {
      "epoch": 17.465626161278337,
      "grad_norm": 0.7937678694725037,
      "learning_rate": 0.0001,
      "loss": 0.7802,
      "step": 47000
    },
    {
      "epoch": 17.651430694908957,
      "grad_norm": 0.8527361154556274,
      "learning_rate": 0.0001,
      "loss": 0.7867,
      "step": 47500
    },
    {
      "epoch": 17.837235228539576,
      "grad_norm": 0.8812717199325562,
      "learning_rate": 0.0001,
      "loss": 0.7815,
      "step": 48000
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.7567463981463738,
      "eval_f1_macro": 0.3792648315920964,
      "eval_f1_micro": 0.7567463981463738,
      "eval_loss": 0.7041681408882141,
      "eval_runtime": 506.1969,
      "eval_samples_per_second": 113.397,
      "eval_steps_per_second": 1.772,
      "learning_rate": 0.0001,
      "step": 48438
    },
    {
      "epoch": 18.023039762170196,
      "grad_norm": 0.9349254965782166,
      "learning_rate": 0.0001,
      "loss": 0.7883,
      "step": 48500
    },
    {
      "epoch": 18.208844295800816,
      "grad_norm": 0.8031138777732849,
      "learning_rate": 0.0001,
      "loss": 0.7735,
      "step": 49000
    },
    {
      "epoch": 18.39464882943144,
      "grad_norm": 0.8354169130325317,
      "learning_rate": 0.0001,
      "loss": 0.7861,
      "step": 49500
    },
    {
      "epoch": 18.58045336306206,
      "grad_norm": 0.9157727956771851,
      "learning_rate": 0.0001,
      "loss": 0.7743,
      "step": 50000
    },
    {
      "epoch": 18.76625789669268,
      "grad_norm": 0.8417840600013733,
      "learning_rate": 0.0001,
      "loss": 0.7785,
      "step": 50500
    },
    {
      "epoch": 18.9520624303233,
      "grad_norm": 0.8945490121841431,
      "learning_rate": 0.0001,
      "loss": 0.7682,
      "step": 51000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.7582446298844968,
      "eval_f1_macro": 0.38576725361345504,
      "eval_f1_micro": 0.7582446298844968,
      "eval_loss": 0.7004917860031128,
      "eval_runtime": 517.5109,
      "eval_samples_per_second": 110.917,
      "eval_steps_per_second": 1.733,
      "learning_rate": 0.0001,
      "step": 51129
    },
    {
      "epoch": 19.137866963953922,
      "grad_norm": 1.05229914188385,
      "learning_rate": 0.0001,
      "loss": 0.7689,
      "step": 51500
    },
    {
      "epoch": 19.32367149758454,
      "grad_norm": 0.6953094601631165,
      "learning_rate": 0.0001,
      "loss": 0.7721,
      "step": 52000
    },
    {
      "epoch": 19.50947603121516,
      "grad_norm": 0.7234140634536743,
      "learning_rate": 0.0001,
      "loss": 0.7757,
      "step": 52500
    },
    {
      "epoch": 19.69528056484578,
      "grad_norm": 0.8757944703102112,
      "learning_rate": 0.0001,
      "loss": 0.7688,
      "step": 53000
    },
    {
      "epoch": 19.881085098476404,
      "grad_norm": 1.1684017181396484,
      "learning_rate": 0.0001,
      "loss": 0.7788,
      "step": 53500
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.7602829219003153,
      "eval_f1_macro": 0.39339544323315423,
      "eval_f1_micro": 0.7602829219003153,
      "eval_loss": 0.6942671537399292,
      "eval_runtime": 534.9756,
      "eval_samples_per_second": 107.296,
      "eval_steps_per_second": 1.677,
      "learning_rate": 0.0001,
      "step": 53820
    },
    {
      "epoch": 20.066889632107024,
      "grad_norm": 1.1924428939819336,
      "learning_rate": 0.0001,
      "loss": 0.7598,
      "step": 54000
    },
    {
      "epoch": 20.252694165737644,
      "grad_norm": 0.6945735812187195,
      "learning_rate": 0.0001,
      "loss": 0.7695,
      "step": 54500
    },
    {
      "epoch": 20.438498699368264,
      "grad_norm": 0.8915848135948181,
      "learning_rate": 0.0001,
      "loss": 0.7645,
      "step": 55000
    },
    {
      "epoch": 20.624303232998884,
      "grad_norm": 0.9735463857650757,
      "learning_rate": 0.0001,
      "loss": 0.7747,
      "step": 55500
    },
    {
      "epoch": 20.810107766629507,
      "grad_norm": 1.0891400575637817,
      "learning_rate": 0.0001,
      "loss": 0.774,
      "step": 56000
    },
    {
      "epoch": 20.995912300260127,
      "grad_norm": 0.9328783750534058,
      "learning_rate": 0.0001,
      "loss": 0.7735,
      "step": 56500
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.7592899078413268,
      "eval_f1_macro": 0.3942710537489151,
      "eval_f1_micro": 0.7592899078413268,
      "eval_loss": 0.6919424533843994,
      "eval_runtime": 511.1285,
      "eval_samples_per_second": 112.302,
      "eval_steps_per_second": 1.755,
      "learning_rate": 0.0001,
      "step": 56511
    },
    {
      "epoch": 21.181716833890746,
      "grad_norm": 0.7840440273284912,
      "learning_rate": 0.0001,
      "loss": 0.7777,
      "step": 57000
    },
    {
      "epoch": 21.367521367521366,
      "grad_norm": 0.9738557934761047,
      "learning_rate": 0.0001,
      "loss": 0.7598,
      "step": 57500
    },
    {
      "epoch": 21.55332590115199,
      "grad_norm": 0.8154065608978271,
      "learning_rate": 0.0001,
      "loss": 0.7661,
      "step": 58000
    },
    {
      "epoch": 21.73913043478261,
      "grad_norm": 1.0773000717163086,
      "learning_rate": 0.0001,
      "loss": 0.7691,
      "step": 58500
    },
    {
      "epoch": 21.92493496841323,
      "grad_norm": 1.0894912481307983,
      "learning_rate": 0.0001,
      "loss": 0.7602,
      "step": 59000
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.7606313478859253,
      "eval_f1_macro": 0.3925331634038099,
      "eval_f1_micro": 0.7606313478859253,
      "eval_loss": 0.6903713941574097,
      "eval_runtime": 507.2562,
      "eval_samples_per_second": 113.16,
      "eval_steps_per_second": 1.768,
      "learning_rate": 0.0001,
      "step": 59202
    },
    {
      "epoch": 22.11073950204385,
      "grad_norm": 1.1766314506530762,
      "learning_rate": 0.0001,
      "loss": 0.7665,
      "step": 59500
    },
    {
      "epoch": 22.296544035674472,
      "grad_norm": 1.028473973274231,
      "learning_rate": 0.0001,
      "loss": 0.7638,
      "step": 60000
    },
    {
      "epoch": 22.482348569305092,
      "grad_norm": 0.900393545627594,
      "learning_rate": 0.0001,
      "loss": 0.7633,
      "step": 60500
    },
    {
      "epoch": 22.66815310293571,
      "grad_norm": 0.9336509704589844,
      "learning_rate": 0.0001,
      "loss": 0.7636,
      "step": 61000
    },
    {
      "epoch": 22.85395763656633,
      "grad_norm": 1.0151203870773315,
      "learning_rate": 0.0001,
      "loss": 0.7572,
      "step": 61500
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.7607010330830474,
      "eval_f1_macro": 0.39534246826429575,
      "eval_f1_micro": 0.7607010330830474,
      "eval_loss": 0.6874070167541504,
      "eval_runtime": 507.2598,
      "eval_samples_per_second": 113.159,
      "eval_steps_per_second": 1.768,
      "learning_rate": 0.0001,
      "step": 61893
    },
    {
      "epoch": 23.03976217019695,
      "grad_norm": 1.0469074249267578,
      "learning_rate": 0.0001,
      "loss": 0.7692,
      "step": 62000
    },
    {
      "epoch": 23.225566703827575,
      "grad_norm": 0.9608765840530396,
      "learning_rate": 0.0001,
      "loss": 0.7497,
      "step": 62500
    },
    {
      "epoch": 23.411371237458194,
      "grad_norm": 1.2574784755706787,
      "learning_rate": 0.0001,
      "loss": 0.7566,
      "step": 63000
    },
    {
      "epoch": 23.597175771088814,
      "grad_norm": 1.0878256559371948,
      "learning_rate": 0.0001,
      "loss": 0.7658,
      "step": 63500
    },
    {
      "epoch": 23.782980304719434,
      "grad_norm": 1.1496306657791138,
      "learning_rate": 0.0001,
      "loss": 0.7642,
      "step": 64000
    },
    {
      "epoch": 23.968784838350057,
      "grad_norm": 1.0485353469848633,
      "learning_rate": 0.0001,
      "loss": 0.7593,
      "step": 64500
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.7612236720614624,
      "eval_f1_macro": 0.3933203620961568,
      "eval_f1_micro": 0.7612236720614624,
      "eval_loss": 0.6864963173866272,
      "eval_runtime": 500.8584,
      "eval_samples_per_second": 114.605,
      "eval_steps_per_second": 1.791,
      "learning_rate": 0.0001,
      "step": 64584
    },
    {
      "epoch": 24.154589371980677,
      "grad_norm": 1.0143301486968994,
      "learning_rate": 0.0001,
      "loss": 0.7559,
      "step": 65000
    },
    {
      "epoch": 24.340393905611297,
      "grad_norm": 0.8291743993759155,
      "learning_rate": 0.0001,
      "loss": 0.7643,
      "step": 65500
    },
    {
      "epoch": 24.526198439241917,
      "grad_norm": 1.1790424585342407,
      "learning_rate": 0.0001,
      "loss": 0.7602,
      "step": 66000
    },
    {
      "epoch": 24.71200297287254,
      "grad_norm": 1.1937212944030762,
      "learning_rate": 0.0001,
      "loss": 0.7636,
      "step": 66500
    },
    {
      "epoch": 24.89780750650316,
      "grad_norm": 1.230569839477539,
      "learning_rate": 0.0001,
      "loss": 0.7548,
      "step": 67000
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.7614153063535478,
      "eval_f1_macro": 0.402343965759826,
      "eval_f1_micro": 0.7614153063535478,
      "eval_loss": 0.684335470199585,
      "eval_runtime": 514.3947,
      "eval_samples_per_second": 111.589,
      "eval_steps_per_second": 1.744,
      "learning_rate": 0.0001,
      "step": 67275
    },
    {
      "epoch": 25.08361204013378,
      "grad_norm": 1.0271650552749634,
      "learning_rate": 0.0001,
      "loss": 0.7568,
      "step": 67500
    },
    {
      "epoch": 25.2694165737644,
      "grad_norm": 1.042904019355774,
      "learning_rate": 0.0001,
      "loss": 0.7524,
      "step": 68000
    },
    {
      "epoch": 25.45522110739502,
      "grad_norm": 1.03508460521698,
      "learning_rate": 0.0001,
      "loss": 0.7555,
      "step": 68500
    },
    {
      "epoch": 25.641025641025642,
      "grad_norm": 0.8760964870452881,
      "learning_rate": 0.0001,
      "loss": 0.7605,
      "step": 69000
    },
    {
      "epoch": 25.826830174656262,
      "grad_norm": 1.2160849571228027,
      "learning_rate": 0.0001,
      "loss": 0.7557,
      "step": 69500
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.7629309593909513,
      "eval_f1_macro": 0.4055175650763901,
      "eval_f1_micro": 0.7629309593909513,
      "eval_loss": 0.6830293536186218,
      "eval_runtime": 568.9631,
      "eval_samples_per_second": 100.887,
      "eval_steps_per_second": 1.577,
      "learning_rate": 0.0001,
      "step": 69966
    },
    {
      "epoch": 26.012634708286882,
      "grad_norm": 0.9436312317848206,
      "learning_rate": 0.0001,
      "loss": 0.7523,
      "step": 70000
    },
    {
      "epoch": 26.1984392419175,
      "grad_norm": 1.0304313898086548,
      "learning_rate": 0.0001,
      "loss": 0.7623,
      "step": 70500
    },
    {
      "epoch": 26.384243775548125,
      "grad_norm": 1.1352418661117554,
      "learning_rate": 0.0001,
      "loss": 0.7476,
      "step": 71000
    },
    {
      "epoch": 26.570048309178745,
      "grad_norm": 0.825731098651886,
      "learning_rate": 0.0001,
      "loss": 0.752,
      "step": 71500
    },
    {
      "epoch": 26.755852842809364,
      "grad_norm": 0.9680258631706238,
      "learning_rate": 0.0001,
      "loss": 0.7572,
      "step": 72000
    },
    {
      "epoch": 26.941657376439984,
      "grad_norm": 0.8693468570709229,
      "learning_rate": 0.0001,
      "loss": 0.7534,
      "step": 72500
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.7630703297851954,
      "eval_f1_macro": 0.40742696523740624,
      "eval_f1_micro": 0.7630703297851954,
      "eval_loss": 0.6827249526977539,
      "eval_runtime": 550.472,
      "eval_samples_per_second": 104.276,
      "eval_steps_per_second": 1.63,
      "learning_rate": 0.0001,
      "step": 72657
    },
    {
      "epoch": 27.127461910070604,
      "grad_norm": 1.1599422693252563,
      "learning_rate": 0.0001,
      "loss": 0.7587,
      "step": 73000
    },
    {
      "epoch": 27.313266443701227,
      "grad_norm": 1.182005763053894,
      "learning_rate": 0.0001,
      "loss": 0.7495,
      "step": 73500
    },
    {
      "epoch": 27.499070977331847,
      "grad_norm": 0.935971736907959,
      "learning_rate": 0.0001,
      "loss": 0.7522,
      "step": 74000
    },
    {
      "epoch": 27.684875510962467,
      "grad_norm": 1.0990965366363525,
      "learning_rate": 0.0001,
      "loss": 0.7513,
      "step": 74500
    },
    {
      "epoch": 27.870680044593087,
      "grad_norm": 0.834299623966217,
      "learning_rate": 0.0001,
      "loss": 0.7609,
      "step": 75000
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.7629309593909513,
      "eval_f1_macro": 0.413578373717749,
      "eval_f1_micro": 0.7629309593909513,
      "eval_loss": 0.6805527210235596,
      "eval_runtime": 553.9776,
      "eval_samples_per_second": 103.616,
      "eval_steps_per_second": 1.619,
      "learning_rate": 0.0001,
      "step": 75348
    },
    {
      "epoch": 28.05648457822371,
      "grad_norm": 1.1609550714492798,
      "learning_rate": 0.0001,
      "loss": 0.7471,
      "step": 75500
    },
    {
      "epoch": 28.24228911185433,
      "grad_norm": 0.9144974946975708,
      "learning_rate": 0.0001,
      "loss": 0.7524,
      "step": 76000
    },
    {
      "epoch": 28.42809364548495,
      "grad_norm": 1.08271324634552,
      "learning_rate": 0.0001,
      "loss": 0.7461,
      "step": 76500
    },
    {
      "epoch": 28.61389817911557,
      "grad_norm": 1.0623596906661987,
      "learning_rate": 0.0001,
      "loss": 0.7497,
      "step": 77000
    },
    {
      "epoch": 28.799702712746193,
      "grad_norm": 1.168769359588623,
      "learning_rate": 0.0001,
      "loss": 0.7433,
      "step": 77500
    },
    {
      "epoch": 28.985507246376812,
      "grad_norm": 0.9733015298843384,
      "learning_rate": 0.0001,
      "loss": 0.7537,
      "step": 78000
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.7625825334053413,
      "eval_f1_macro": 0.4138447052049864,
      "eval_f1_micro": 0.7625825334053413,
      "eval_loss": 0.6796479225158691,
      "eval_runtime": 535.8607,
      "eval_samples_per_second": 107.119,
      "eval_steps_per_second": 1.674,
      "learning_rate": 0.0001,
      "step": 78039
    },
    {
      "epoch": 29.171311780007432,
      "grad_norm": 1.3112975358963013,
      "learning_rate": 0.0001,
      "loss": 0.7435,
      "step": 78500
    },
    {
      "epoch": 29.357116313638052,
      "grad_norm": 1.1170014142990112,
      "learning_rate": 0.0001,
      "loss": 0.7524,
      "step": 79000
    },
    {
      "epoch": 29.54292084726867,
      "grad_norm": 0.9579864144325256,
      "learning_rate": 0.0001,
      "loss": 0.7544,
      "step": 79500
    },
    {
      "epoch": 29.728725380899295,
      "grad_norm": 0.9765433073043823,
      "learning_rate": 0.0001,
      "loss": 0.7498,
      "step": 80000
    },
    {
      "epoch": 29.914529914529915,
      "grad_norm": 1.1268893480300903,
      "learning_rate": 0.0001,
      "loss": 0.7533,
      "step": 80500
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.7635929687636104,
      "eval_f1_macro": 0.41283784117218203,
      "eval_f1_micro": 0.7635929687636104,
      "eval_loss": 0.6774595379829407,
      "eval_runtime": 535.3939,
      "eval_samples_per_second": 107.213,
      "eval_steps_per_second": 1.675,
      "learning_rate": 0.0001,
      "step": 80730
    },
    {
      "epoch": 30.100334448160535,
      "grad_norm": 1.268797755241394,
      "learning_rate": 0.0001,
      "loss": 0.7368,
      "step": 81000
    },
    {
      "epoch": 30.286138981791154,
      "grad_norm": 1.029023289680481,
      "learning_rate": 0.0001,
      "loss": 0.7478,
      "step": 81500
    },
    {
      "epoch": 30.471943515421778,
      "grad_norm": 1.3363114595413208,
      "learning_rate": 0.0001,
      "loss": 0.7487,
      "step": 82000
    },
    {
      "epoch": 30.657748049052397,
      "grad_norm": 1.4572941064834595,
      "learning_rate": 0.0001,
      "loss": 0.7467,
      "step": 82500
    },
    {
      "epoch": 30.843552582683017,
      "grad_norm": 1.1603232622146606,
      "learning_rate": 0.0001,
      "loss": 0.7481,
      "step": 83000
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.7643769272312328,
      "eval_f1_macro": 0.4100283180344899,
      "eval_f1_micro": 0.7643769272312328,
      "eval_loss": 0.677918553352356,
      "eval_runtime": 535.5808,
      "eval_samples_per_second": 107.175,
      "eval_steps_per_second": 1.675,
      "learning_rate": 0.0001,
      "step": 83421
    },
    {
      "epoch": 31.029357116313637,
      "grad_norm": 1.3755314350128174,
      "learning_rate": 0.0001,
      "loss": 0.7545,
      "step": 83500
    },
    {
      "epoch": 31.21516164994426,
      "grad_norm": 1.2730791568756104,
      "learning_rate": 0.0001,
      "loss": 0.7523,
      "step": 84000
    },
    {
      "epoch": 31.40096618357488,
      "grad_norm": 1.3255988359451294,
      "learning_rate": 0.0001,
      "loss": 0.7459,
      "step": 84500
    },
    {
      "epoch": 31.5867707172055,
      "grad_norm": 1.246517300605774,
      "learning_rate": 0.0001,
      "loss": 0.7411,
      "step": 85000
    },
    {
      "epoch": 31.77257525083612,
      "grad_norm": 1.3307832479476929,
      "learning_rate": 0.0001,
      "loss": 0.7313,
      "step": 85500
    },
    {
      "epoch": 31.95837978446674,
      "grad_norm": 1.4022419452667236,
      "learning_rate": 0.0001,
      "loss": 0.7523,
      "step": 86000
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.7641504503405864,
      "eval_f1_macro": 0.41093585032897456,
      "eval_f1_micro": 0.7641504503405864,
      "eval_loss": 0.6754601001739502,
      "eval_runtime": 518.6265,
      "eval_samples_per_second": 110.679,
      "eval_steps_per_second": 1.73,
      "learning_rate": 0.0001,
      "step": 86112
    },
    {
      "epoch": 32.14418431809736,
      "grad_norm": 1.3993639945983887,
      "learning_rate": 0.0001,
      "loss": 0.7392,
      "step": 86500
    },
    {
      "epoch": 32.32998885172798,
      "grad_norm": 1.2422406673431396,
      "learning_rate": 0.0001,
      "loss": 0.7434,
      "step": 87000
    },
    {
      "epoch": 32.515793385358606,
      "grad_norm": 1.2804062366485596,
      "learning_rate": 0.0001,
      "loss": 0.7366,
      "step": 87500
    },
    {
      "epoch": 32.70159791898922,
      "grad_norm": 1.6868325471878052,
      "learning_rate": 0.0001,
      "loss": 0.7515,
      "step": 88000
    },
    {
      "epoch": 32.887402452619845,
      "grad_norm": 1.067484974861145,
      "learning_rate": 0.0001,
      "loss": 0.7459,
      "step": 88500
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.7645162976254769,
      "eval_f1_macro": 0.4185852176848548,
      "eval_f1_micro": 0.7645162976254769,
      "eval_loss": 0.6749601364135742,
      "eval_runtime": 515.3075,
      "eval_samples_per_second": 111.392,
      "eval_steps_per_second": 1.741,
      "learning_rate": 0.0001,
      "step": 88803
    },
    {
      "epoch": 33.07320698625046,
      "grad_norm": 1.3152741193771362,
      "learning_rate": 0.0001,
      "loss": 0.7448,
      "step": 89000
    },
    {
      "epoch": 33.259011519881085,
      "grad_norm": 1.0844374895095825,
      "learning_rate": 0.0001,
      "loss": 0.7353,
      "step": 89500
    },
    {
      "epoch": 33.44481605351171,
      "grad_norm": 1.1964083909988403,
      "learning_rate": 0.0001,
      "loss": 0.7439,
      "step": 90000
    },
    {
      "epoch": 33.630620587142324,
      "grad_norm": 1.1734319925308228,
      "learning_rate": 0.0001,
      "loss": 0.7434,
      "step": 90500
    },
    {
      "epoch": 33.81642512077295,
      "grad_norm": 1.2583281993865967,
      "learning_rate": 0.0001,
      "loss": 0.7454,
      "step": 91000
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.7650040940053309,
      "eval_f1_macro": 0.41458520697205553,
      "eval_f1_micro": 0.7650040940053309,
      "eval_loss": 0.6746455430984497,
      "eval_runtime": 513.1282,
      "eval_samples_per_second": 111.865,
      "eval_steps_per_second": 1.748,
      "learning_rate": 0.0001,
      "step": 91494
    },
    {
      "epoch": 34.002229654403564,
      "grad_norm": 1.368170142173767,
      "learning_rate": 0.0001,
      "loss": 0.7475,
      "step": 91500
    },
    {
      "epoch": 34.18803418803419,
      "grad_norm": 1.1858837604522705,
      "learning_rate": 0.0001,
      "loss": 0.7363,
      "step": 92000
    },
    {
      "epoch": 34.37383872166481,
      "grad_norm": 1.5839638710021973,
      "learning_rate": 0.0001,
      "loss": 0.7352,
      "step": 92500
    },
    {
      "epoch": 34.55964325529543,
      "grad_norm": 1.560922622680664,
      "learning_rate": 0.0001,
      "loss": 0.7371,
      "step": 93000
    },
    {
      "epoch": 34.74544778892605,
      "grad_norm": 1.3220325708389282,
      "learning_rate": 0.0001,
      "loss": 0.7442,
      "step": 93500
    },
    {
      "epoch": 34.931252322556674,
      "grad_norm": 1.275272250175476,
      "learning_rate": 0.0001,
      "loss": 0.7426,
      "step": 94000
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.7641852929391474,
      "eval_f1_macro": 0.42549626405712787,
      "eval_f1_micro": 0.7641852929391474,
      "eval_loss": 0.6740487813949585,
      "eval_runtime": 514.2418,
      "eval_samples_per_second": 111.623,
      "eval_steps_per_second": 1.744,
      "learning_rate": 0.0001,
      "step": 94185
    },
    {
      "epoch": 35.11705685618729,
      "grad_norm": 1.3994961977005005,
      "learning_rate": 0.0001,
      "loss": 0.7432,
      "step": 94500
    },
    {
      "epoch": 35.30286138981791,
      "grad_norm": 1.2828725576400757,
      "learning_rate": 0.0001,
      "loss": 0.7443,
      "step": 95000
    },
    {
      "epoch": 35.48866592344853,
      "grad_norm": 1.3517532348632812,
      "learning_rate": 0.0001,
      "loss": 0.7426,
      "step": 95500
    },
    {
      "epoch": 35.67447045707915,
      "grad_norm": 1.2311525344848633,
      "learning_rate": 0.0001,
      "loss": 0.7274,
      "step": 96000
    },
    {
      "epoch": 35.860274990709776,
      "grad_norm": 1.209617257118225,
      "learning_rate": 0.0001,
      "loss": 0.7446,
      "step": 96500
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.7646905106182819,
      "eval_f1_macro": 0.42042185334833837,
      "eval_f1_micro": 0.7646905106182819,
      "eval_loss": 0.6740365624427795,
      "eval_runtime": 526.8929,
      "eval_samples_per_second": 108.942,
      "eval_steps_per_second": 1.702,
      "learning_rate": 0.0001,
      "step": 96876
    },
    {
      "epoch": 36.04607952434039,
      "grad_norm": 1.2096632719039917,
      "learning_rate": 0.0001,
      "loss": 0.738,
      "step": 97000
    },
    {
      "epoch": 36.231884057971016,
      "grad_norm": 1.258972406387329,
      "learning_rate": 0.0001,
      "loss": 0.7421,
      "step": 97500
    },
    {
      "epoch": 36.41768859160163,
      "grad_norm": 1.5985496044158936,
      "learning_rate": 0.0001,
      "loss": 0.7426,
      "step": 98000
    },
    {
      "epoch": 36.603493125232255,
      "grad_norm": 1.261830449104309,
      "learning_rate": 0.0001,
      "loss": 0.7384,
      "step": 98500
    },
    {
      "epoch": 36.78929765886288,
      "grad_norm": 1.5109082460403442,
      "learning_rate": 0.0001,
      "loss": 0.7383,
      "step": 99000
    },
    {
      "epoch": 36.975102192493495,
      "grad_norm": 1.1752451658248901,
      "learning_rate": 0.0001,
      "loss": 0.7431,
      "step": 99500
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.7641678716398669,
      "eval_f1_macro": 0.4194338951085209,
      "eval_f1_micro": 0.7641678716398669,
      "eval_loss": 0.6731483936309814,
      "eval_runtime": 503.5131,
      "eval_samples_per_second": 114.001,
      "eval_steps_per_second": 1.781,
      "learning_rate": 0.0001,
      "step": 99567
    },
    {
      "epoch": 37.16090672612412,
      "grad_norm": 1.4535642862319946,
      "learning_rate": 0.0001,
      "loss": 0.7282,
      "step": 100000
    },
    {
      "epoch": 37.34671125975474,
      "grad_norm": 1.5286682844161987,
      "learning_rate": 0.0001,
      "loss": 0.7332,
      "step": 100500
    },
    {
      "epoch": 37.53251579338536,
      "grad_norm": 1.6406091451644897,
      "learning_rate": 0.0001,
      "loss": 0.7387,
      "step": 101000
    },
    {
      "epoch": 37.71832032701598,
      "grad_norm": 1.0090203285217285,
      "learning_rate": 0.0001,
      "loss": 0.7364,
      "step": 101500
    },
    {
      "epoch": 37.9041248606466,
      "grad_norm": 1.3233065605163574,
      "learning_rate": 0.0001,
      "loss": 0.7468,
      "step": 102000
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.7648821449103674,
      "eval_f1_macro": 0.42433192329853336,
      "eval_f1_micro": 0.7648821449103674,
      "eval_loss": 0.671998918056488,
      "eval_runtime": 536.8447,
      "eval_samples_per_second": 106.923,
      "eval_steps_per_second": 1.671,
      "learning_rate": 0.0001,
      "step": 102258
    },
    {
      "epoch": 38.08992939427722,
      "grad_norm": 1.2303552627563477,
      "learning_rate": 0.0001,
      "loss": 0.7463,
      "step": 102500
    },
    {
      "epoch": 38.275733927907844,
      "grad_norm": 1.200378179550171,
      "learning_rate": 0.0001,
      "loss": 0.7296,
      "step": 103000
    },
    {
      "epoch": 38.46153846153846,
      "grad_norm": 1.3014321327209473,
      "learning_rate": 0.0001,
      "loss": 0.7437,
      "step": 103500
    },
    {
      "epoch": 38.64734299516908,
      "grad_norm": 1.429656982421875,
      "learning_rate": 0.0001,
      "loss": 0.7444,
      "step": 104000
    },
    {
      "epoch": 38.8331475287997,
      "grad_norm": 1.3417949676513672,
      "learning_rate": 0.0001,
      "loss": 0.7307,
      "step": 104500
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.7659274228671974,
      "eval_f1_macro": 0.42040362773805245,
      "eval_f1_micro": 0.7659274228671974,
      "eval_loss": 0.6694707870483398,
      "eval_runtime": 527.8534,
      "eval_samples_per_second": 108.744,
      "eval_steps_per_second": 1.699,
      "learning_rate": 0.0001,
      "step": 104949
    },
    {
      "epoch": 39.01895206243032,
      "grad_norm": 1.656445860862732,
      "learning_rate": 0.0001,
      "loss": 0.7423,
      "step": 105000
    },
    {
      "epoch": 39.204756596060946,
      "grad_norm": 1.33378267288208,
      "learning_rate": 0.0001,
      "loss": 0.7407,
      "step": 105500
    },
    {
      "epoch": 39.39056112969156,
      "grad_norm": 1.814835548400879,
      "learning_rate": 0.0001,
      "loss": 0.7395,
      "step": 106000
    },
    {
      "epoch": 39.576365663322186,
      "grad_norm": 1.5420753955841064,
      "learning_rate": 0.0001,
      "loss": 0.7321,
      "step": 106500
    },
    {
      "epoch": 39.76217019695281,
      "grad_norm": 1.2062960863113403,
      "learning_rate": 0.0001,
      "loss": 0.7302,
      "step": 107000
    },
    {
      "epoch": 39.947974730583425,
      "grad_norm": 1.4515705108642578,
      "learning_rate": 0.0001,
      "loss": 0.7404,
      "step": 107500
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.7665023257434539,
      "eval_f1_macro": 0.42052839725843943,
      "eval_f1_micro": 0.7665023257434539,
      "eval_loss": 0.6693674325942993,
      "eval_runtime": 517.9605,
      "eval_samples_per_second": 110.821,
      "eval_steps_per_second": 1.732,
      "learning_rate": 0.0001,
      "step": 107640
    },
    {
      "epoch": 40.13377926421405,
      "grad_norm": 1.211496114730835,
      "learning_rate": 0.0001,
      "loss": 0.728,
      "step": 108000
    },
    {
      "epoch": 40.319583797844665,
      "grad_norm": 1.0356364250183105,
      "learning_rate": 0.0001,
      "loss": 0.7307,
      "step": 108500
    },
    {
      "epoch": 40.50538833147529,
      "grad_norm": 1.508729338645935,
      "learning_rate": 0.0001,
      "loss": 0.732,
      "step": 109000
    },
    {
      "epoch": 40.69119286510591,
      "grad_norm": 1.315461277961731,
      "learning_rate": 0.0001,
      "loss": 0.7471,
      "step": 109500
    },
    {
      "epoch": 40.87699739873653,
      "grad_norm": 1.2628021240234375,
      "learning_rate": 0.0001,
      "loss": 0.7355,
      "step": 110000
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.7658228950715145,
      "eval_f1_macro": 0.4176095837463332,
      "eval_f1_micro": 0.7658228950715145,
      "eval_loss": 0.6682748198509216,
      "eval_runtime": 491.8206,
      "eval_samples_per_second": 116.711,
      "eval_steps_per_second": 1.824,
      "learning_rate": 0.0001,
      "step": 110331
    },
    {
      "epoch": 41.06280193236715,
      "grad_norm": 1.4534225463867188,
      "learning_rate": 0.0001,
      "loss": 0.7362,
      "step": 110500
    },
    {
      "epoch": 41.24860646599777,
      "grad_norm": 1.305442214012146,
      "learning_rate": 0.0001,
      "loss": 0.7292,
      "step": 111000
    },
    {
      "epoch": 41.43441099962839,
      "grad_norm": 1.2671395540237427,
      "learning_rate": 0.0001,
      "loss": 0.7293,
      "step": 111500
    },
    {
      "epoch": 41.620215533259014,
      "grad_norm": 1.7886228561401367,
      "learning_rate": 0.0001,
      "loss": 0.7379,
      "step": 112000
    },
    {
      "epoch": 41.80602006688963,
      "grad_norm": 1.565897822380066,
      "learning_rate": 0.0001,
      "loss": 0.7388,
      "step": 112500
    },
    {
      "epoch": 41.99182460052025,
      "grad_norm": 1.4743560552597046,
      "learning_rate": 0.0001,
      "loss": 0.7508,
      "step": 113000
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.7665371683420149,
      "eval_f1_macro": 0.4307432846853069,
      "eval_f1_micro": 0.7665371683420149,
      "eval_loss": 0.6682831645011902,
      "eval_runtime": 484.4423,
      "eval_samples_per_second": 118.489,
      "eval_steps_per_second": 1.852,
      "learning_rate": 0.0001,
      "step": 113022
    },
    {
      "epoch": 42.17762913415088,
      "grad_norm": 1.530715823173523,
      "learning_rate": 0.0001,
      "loss": 0.7309,
      "step": 113500
    },
    {
      "epoch": 42.36343366778149,
      "grad_norm": 1.4992977380752563,
      "learning_rate": 0.0001,
      "loss": 0.7282,
      "step": 114000
    },
    {
      "epoch": 42.549238201412116,
      "grad_norm": 1.5467846393585205,
      "learning_rate": 0.0001,
      "loss": 0.7377,
      "step": 114500
    },
    {
      "epoch": 42.73504273504273,
      "grad_norm": 1.5137412548065186,
      "learning_rate": 0.0001,
      "loss": 0.733,
      "step": 115000
    },
    {
      "epoch": 42.920847268673356,
      "grad_norm": 1.6055561304092407,
      "learning_rate": 0.0001,
      "loss": 0.7368,
      "step": 115500
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.7664152192470515,
      "eval_f1_macro": 0.42637174689527435,
      "eval_f1_micro": 0.7664152192470515,
      "eval_loss": 0.6695142984390259,
      "eval_runtime": 488.5528,
      "eval_samples_per_second": 117.492,
      "eval_steps_per_second": 1.836,
      "learning_rate": 0.0001,
      "step": 115713
    },
    {
      "epoch": 43.10665180230398,
      "grad_norm": 1.460276484489441,
      "learning_rate": 0.0001,
      "loss": 0.7371,
      "step": 116000
    },
    {
      "epoch": 43.292456335934595,
      "grad_norm": 1.4303547143936157,
      "learning_rate": 0.0001,
      "loss": 0.7276,
      "step": 116500
    },
    {
      "epoch": 43.47826086956522,
      "grad_norm": 1.5059736967086792,
      "learning_rate": 0.0001,
      "loss": 0.7282,
      "step": 117000
    },
    {
      "epoch": 43.664065403195835,
      "grad_norm": 1.4985263347625732,
      "learning_rate": 0.0001,
      "loss": 0.7347,
      "step": 117500
    },
    {
      "epoch": 43.84986993682646,
      "grad_norm": 1.1422086954116821,
      "learning_rate": 0.0001,
      "loss": 0.7362,
      "step": 118000
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.765840316370795,
      "eval_f1_macro": 0.42609728385101026,
      "eval_f1_micro": 0.765840316370795,
      "eval_loss": 0.669391393661499,
      "eval_runtime": 508.1911,
      "eval_samples_per_second": 112.952,
      "eval_steps_per_second": 1.765,
      "learning_rate": 0.0001,
      "step": 118404
    },
    {
      "epoch": 44.03567447045708,
      "grad_norm": 1.8496322631835938,
      "learning_rate": 0.0001,
      "loss": 0.7336,
      "step": 118500
    },
    {
      "epoch": 44.2214790040877,
      "grad_norm": 1.4898383617401123,
      "learning_rate": 0.0001,
      "loss": 0.7361,
      "step": 119000
    },
    {
      "epoch": 44.40728353771832,
      "grad_norm": 1.341386079788208,
      "learning_rate": 0.0001,
      "loss": 0.733,
      "step": 119500
    },
    {
      "epoch": 44.593088071348944,
      "grad_norm": 1.747814416885376,
      "learning_rate": 0.0001,
      "loss": 0.7256,
      "step": 120000
    },
    {
      "epoch": 44.77889260497956,
      "grad_norm": 1.5329481363296509,
      "learning_rate": 0.0001,
      "loss": 0.7203,
      "step": 120500
    },
    {
      "epoch": 44.964697138610184,
      "grad_norm": 1.4953057765960693,
      "learning_rate": 0.0001,
      "loss": 0.7287,
      "step": 121000
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.7649518301074895,
      "eval_f1_macro": 0.4307801257532618,
      "eval_f1_micro": 0.7649518301074895,
      "eval_loss": 0.6695447564125061,
      "eval_runtime": 486.6366,
      "eval_samples_per_second": 117.955,
      "eval_steps_per_second": 1.843,
      "learning_rate": 0.0001,
      "step": 121095
    },
    {
      "epoch": 45.1505016722408,
      "grad_norm": 1.5713474750518799,
      "learning_rate": 0.0001,
      "loss": 0.7282,
      "step": 121500
    },
    {
      "epoch": 45.33630620587142,
      "grad_norm": 1.133755087852478,
      "learning_rate": 0.0001,
      "loss": 0.7333,
      "step": 122000
    },
    {
      "epoch": 45.52211073950205,
      "grad_norm": 1.4431047439575195,
      "learning_rate": 0.0001,
      "loss": 0.7271,
      "step": 122500
    },
    {
      "epoch": 45.70791527313266,
      "grad_norm": 1.942878007888794,
      "learning_rate": 0.0001,
      "loss": 0.7298,
      "step": 123000
    },
    {
      "epoch": 45.893719806763286,
      "grad_norm": 1.6559330224990845,
      "learning_rate": 0.0001,
      "loss": 0.7374,
      "step": 123500
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.7673908120067595,
      "eval_f1_macro": 0.4362286100546047,
      "eval_f1_micro": 0.7673908120067595,
      "eval_loss": 0.6653340458869934,
      "eval_runtime": 483.687,
      "eval_samples_per_second": 118.674,
      "eval_steps_per_second": 1.855,
      "learning_rate": 0.0001,
      "step": 123786
    },
    {
      "epoch": 46.0795243403939,
      "grad_norm": 1.346835732460022,
      "learning_rate": 0.0001,
      "loss": 0.7369,
      "step": 124000
    },
    {
      "epoch": 46.265328874024526,
      "grad_norm": 1.3699783086776733,
      "learning_rate": 0.0001,
      "loss": 0.7344,
      "step": 124500
    },
    {
      "epoch": 46.45113340765515,
      "grad_norm": 1.4523323774337769,
      "learning_rate": 0.0001,
      "loss": 0.7322,
      "step": 125000
    },
    {
      "epoch": 46.636937941285765,
      "grad_norm": 1.5713220834732056,
      "learning_rate": 0.0001,
      "loss": 0.7276,
      "step": 125500
    },
    {
      "epoch": 46.82274247491639,
      "grad_norm": 1.4702354669570923,
      "learning_rate": 0.0001,
      "loss": 0.7321,
      "step": 126000
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.7674430759046009,
      "eval_f1_macro": 0.43279477858934173,
      "eval_f1_micro": 0.7674430759046009,
      "eval_loss": 0.6659862995147705,
      "eval_runtime": 487.7507,
      "eval_samples_per_second": 117.685,
      "eval_steps_per_second": 1.839,
      "learning_rate": 0.0001,
      "step": 126477
    },
    {
      "epoch": 47.00854700854701,
      "grad_norm": 1.5675249099731445,
      "learning_rate": 0.0001,
      "loss": 0.7255,
      "step": 126500
    },
    {
      "epoch": 47.19435154217763,
      "grad_norm": 1.4069844484329224,
      "learning_rate": 0.0001,
      "loss": 0.7272,
      "step": 127000
    },
    {
      "epoch": 47.38015607580825,
      "grad_norm": 1.2489932775497437,
      "learning_rate": 0.0001,
      "loss": 0.7273,
      "step": 127500
    },
    {
      "epoch": 47.56596060943887,
      "grad_norm": 1.8463833332061768,
      "learning_rate": 0.0001,
      "loss": 0.74,
      "step": 128000
    },
    {
      "epoch": 47.75176514306949,
      "grad_norm": 1.3903892040252686,
      "learning_rate": 0.0001,
      "loss": 0.7391,
      "step": 128500
    },
    {
      "epoch": 47.937569676700114,
      "grad_norm": 1.2857866287231445,
      "learning_rate": 0.0001,
      "loss": 0.7352,
      "step": 129000
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.7668855943276249,
      "eval_f1_macro": 0.43077383038275147,
      "eval_f1_micro": 0.7668855943276249,
      "eval_loss": 0.665557861328125,
      "eval_runtime": 486.8422,
      "eval_samples_per_second": 117.905,
      "eval_steps_per_second": 1.842,
      "learning_rate": 0.0001,
      "step": 129168
    },
    {
      "epoch": 48.12337421033073,
      "grad_norm": 1.7764297723770142,
      "learning_rate": 0.0001,
      "loss": 0.7203,
      "step": 129500
    },
    {
      "epoch": 48.309178743961354,
      "grad_norm": 1.469336748123169,
      "learning_rate": 0.0001,
      "loss": 0.7292,
      "step": 130000
    },
    {
      "epoch": 48.49498327759197,
      "grad_norm": 1.4002306461334229,
      "learning_rate": 0.0001,
      "loss": 0.7279,
      "step": 130500
    },
    {
      "epoch": 48.680787811222594,
      "grad_norm": 1.6152621507644653,
      "learning_rate": 0.0001,
      "loss": 0.7386,
      "step": 131000
    },
    {
      "epoch": 48.86659234485322,
      "grad_norm": 1.5064826011657715,
      "learning_rate": 0.0001,
      "loss": 0.7373,
      "step": 131500
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.7666242748384174,
      "eval_f1_macro": 0.4241681801855841,
      "eval_f1_micro": 0.7666242748384174,
      "eval_loss": 0.6672787666320801,
      "eval_runtime": 484.1155,
      "eval_samples_per_second": 118.569,
      "eval_steps_per_second": 1.853,
      "learning_rate": 0.0001,
      "step": 131859
    },
    {
      "epoch": 49.05239687848383,
      "grad_norm": 1.329869031906128,
      "learning_rate": 0.0001,
      "loss": 0.7295,
      "step": 132000
    },
    {
      "epoch": 49.238201412114456,
      "grad_norm": 1.749746322631836,
      "learning_rate": 0.0001,
      "loss": 0.7301,
      "step": 132500
    },
    {
      "epoch": 49.42400594574507,
      "grad_norm": 2.059324026107788,
      "learning_rate": 0.0001,
      "loss": 0.7214,
      "step": 133000
    },
    {
      "epoch": 49.609810479375696,
      "grad_norm": 1.6400606632232666,
      "learning_rate": 0.0001,
      "loss": 0.7285,
      "step": 133500
    },
    {
      "epoch": 49.79561501300632,
      "grad_norm": 1.6473073959350586,
      "learning_rate": 0.0001,
      "loss": 0.7283,
      "step": 134000
    },
    {
      "epoch": 49.981419546636936,
      "grad_norm": 1.7654767036437988,
      "learning_rate": 0.0001,
      "loss": 0.7307,
      "step": 134500
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.7662584275535269,
      "eval_f1_macro": 0.43151276516162357,
      "eval_f1_micro": 0.7662584275535269,
      "eval_loss": 0.6661437749862671,
      "eval_runtime": 486.5547,
      "eval_samples_per_second": 117.974,
      "eval_steps_per_second": 1.844,
      "learning_rate": 0.0001,
      "step": 134550
    },
    {
      "epoch": 50.16722408026756,
      "grad_norm": 1.484832525253296,
      "learning_rate": 0.0001,
      "loss": 0.731,
      "step": 135000
    },
    {
      "epoch": 50.35302861389818,
      "grad_norm": 1.4548848867416382,
      "learning_rate": 0.0001,
      "loss": 0.7262,
      "step": 135500
    },
    {
      "epoch": 50.5388331475288,
      "grad_norm": 1.7120981216430664,
      "learning_rate": 0.0001,
      "loss": 0.7345,
      "step": 136000
    },
    {
      "epoch": 50.72463768115942,
      "grad_norm": 1.789556860923767,
      "learning_rate": 0.0001,
      "loss": 0.7253,
      "step": 136500
    },
    {
      "epoch": 50.91044221479004,
      "grad_norm": 2.5206713676452637,
      "learning_rate": 0.0001,
      "loss": 0.7235,
      "step": 137000
    },
    {
      "epoch": 51.0,
      "eval_accuracy": 0.7667288026341005,
      "eval_f1_macro": 0.4307690989303114,
      "eval_f1_micro": 0.7667288026341005,
      "eval_loss": 0.6638755798339844,
      "eval_runtime": 505.3338,
      "eval_samples_per_second": 113.59,
      "eval_steps_per_second": 1.775,
      "learning_rate": 0.0001,
      "step": 137241
    },
    {
      "epoch": 51.09624674842066,
      "grad_norm": 1.9092504978179932,
      "learning_rate": 0.0001,
      "loss": 0.7266,
      "step": 137500
    },
    {
      "epoch": 51.282051282051285,
      "grad_norm": 1.6534066200256348,
      "learning_rate": 0.0001,
      "loss": 0.7266,
      "step": 138000
    },
    {
      "epoch": 51.4678558156819,
      "grad_norm": 1.7287977933883667,
      "learning_rate": 0.0001,
      "loss": 0.7304,
      "step": 138500
    },
    {
      "epoch": 51.653660349312524,
      "grad_norm": 2.113309860229492,
      "learning_rate": 0.0001,
      "loss": 0.7356,
      "step": 139000
    },
    {
      "epoch": 51.83946488294314,
      "grad_norm": 1.4403417110443115,
      "learning_rate": 0.0001,
      "loss": 0.7295,
      "step": 139500
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.7679134509851745,
      "eval_f1_macro": 0.4427799679621408,
      "eval_f1_micro": 0.7679134509851745,
      "eval_loss": 0.6654694676399231,
      "eval_runtime": 502.5773,
      "eval_samples_per_second": 114.213,
      "eval_steps_per_second": 1.785,
      "learning_rate": 0.0001,
      "step": 139932
    },
    {
      "epoch": 52.025269416573764,
      "grad_norm": 1.7017192840576172,
      "learning_rate": 0.0001,
      "loss": 0.712,
      "step": 140000
    },
    {
      "epoch": 52.21107395020439,
      "grad_norm": 1.3907127380371094,
      "learning_rate": 0.0001,
      "loss": 0.7329,
      "step": 140500
    },
    {
      "epoch": 52.396878483835,
      "grad_norm": 1.791027307510376,
      "learning_rate": 0.0001,
      "loss": 0.7276,
      "step": 141000
    },
    {
      "epoch": 52.58268301746563,
      "grad_norm": 1.384203553199768,
      "learning_rate": 0.0001,
      "loss": 0.7126,
      "step": 141500
    },
    {
      "epoch": 52.76848755109625,
      "grad_norm": 2.309049367904663,
      "learning_rate": 0.0001,
      "loss": 0.7284,
      "step": 142000
    },
    {
      "epoch": 52.954292084726866,
      "grad_norm": 1.4892642498016357,
      "learning_rate": 0.0001,
      "loss": 0.7267,
      "step": 142500
    },
    {
      "epoch": 53.0,
      "eval_accuracy": 0.767234020313235,
      "eval_f1_macro": 0.4341825403324277,
      "eval_f1_micro": 0.767234020313235,
      "eval_loss": 0.6643231511116028,
      "eval_runtime": 484.553,
      "eval_samples_per_second": 118.462,
      "eval_steps_per_second": 1.851,
      "learning_rate": 0.0001,
      "step": 142623
    },
    {
      "epoch": 53.14009661835749,
      "grad_norm": 1.9168522357940674,
      "learning_rate": 0.0001,
      "loss": 0.7301,
      "step": 143000
    },
    {
      "epoch": 53.325901151988106,
      "grad_norm": 2.0074169635772705,
      "learning_rate": 0.0001,
      "loss": 0.7194,
      "step": 143500
    },
    {
      "epoch": 53.51170568561873,
      "grad_norm": 1.7252275943756104,
      "learning_rate": 0.0001,
      "loss": 0.7248,
      "step": 144000
    },
    {
      "epoch": 53.69751021924935,
      "grad_norm": 1.5249263048171997,
      "learning_rate": 0.0001,
      "loss": 0.7318,
      "step": 144500
    },
    {
      "epoch": 53.88331475287997,
      "grad_norm": 1.5287233591079712,
      "learning_rate": 0.0001,
      "loss": 0.724,
      "step": 145000
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.7663455340499294,
      "eval_f1_macro": 0.4459616186399035,
      "eval_f1_micro": 0.7663455340499294,
      "eval_loss": 0.667382001876831,
      "eval_runtime": 485.0681,
      "eval_samples_per_second": 118.336,
      "eval_steps_per_second": 1.849,
      "learning_rate": 0.0001,
      "step": 145314
    },
    {
      "epoch": 54.06911928651059,
      "grad_norm": 2.074103593826294,
      "learning_rate": 0.0001,
      "loss": 0.7333,
      "step": 145500
    },
    {
      "epoch": 54.25492382014121,
      "grad_norm": 1.8293678760528564,
      "learning_rate": 0.0001,
      "loss": 0.7265,
      "step": 146000
    },
    {
      "epoch": 54.44072835377183,
      "grad_norm": 1.985289454460144,
      "learning_rate": 0.0001,
      "loss": 0.7241,
      "step": 146500
    },
    {
      "epoch": 54.626532887402455,
      "grad_norm": 1.5796940326690674,
      "learning_rate": 0.0001,
      "loss": 0.7188,
      "step": 147000
    },
    {
      "epoch": 54.81233742103307,
      "grad_norm": 2.1631064414978027,
      "learning_rate": 0.0001,
      "loss": 0.7321,
      "step": 147500
    },
    {
      "epoch": 54.998141954663694,
      "grad_norm": 1.892568588256836,
      "learning_rate": 0.0001,
      "loss": 0.734,
      "step": 148000
    },
    {
      "epoch": 55.0,
      "eval_accuracy": 0.7684709325621505,
      "eval_f1_macro": 0.4389385481332223,
      "eval_f1_micro": 0.7684709325621505,
      "eval_loss": 0.6627440452575684,
      "eval_runtime": 477.8566,
      "eval_samples_per_second": 120.122,
      "eval_steps_per_second": 1.877,
      "learning_rate": 0.0001,
      "step": 148005
    },
    {
      "epoch": 55.18394648829432,
      "grad_norm": 1.820019006729126,
      "learning_rate": 0.0001,
      "loss": 0.7167,
      "step": 148500
    },
    {
      "epoch": 55.369751021924934,
      "grad_norm": 1.9633277654647827,
      "learning_rate": 0.0001,
      "loss": 0.7347,
      "step": 149000
    },
    {
      "epoch": 55.55555555555556,
      "grad_norm": 1.589463710784912,
      "learning_rate": 0.0001,
      "loss": 0.7285,
      "step": 149500
    },
    {
      "epoch": 55.74136008918617,
      "grad_norm": 1.771761417388916,
      "learning_rate": 0.0001,
      "loss": 0.73,
      "step": 150000
    },
    {
      "epoch": 55.9271646228168,
      "grad_norm": 1.93351411819458,
      "learning_rate": 0.0001,
      "loss": 0.7285,
      "step": 150500
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.767094649918991,
      "eval_f1_macro": 0.43857797557707,
      "eval_f1_micro": 0.767094649918991,
      "eval_loss": 0.6627209186553955,
      "eval_runtime": 479.9805,
      "eval_samples_per_second": 119.59,
      "eval_steps_per_second": 1.869,
      "learning_rate": 0.0001,
      "step": 150696
    },
    {
      "epoch": 56.11296915644742,
      "grad_norm": 1.5868228673934937,
      "learning_rate": 0.0001,
      "loss": 0.724,
      "step": 151000
    },
    {
      "epoch": 56.298773690078036,
      "grad_norm": 1.6344318389892578,
      "learning_rate": 0.0001,
      "loss": 0.7218,
      "step": 151500
    },
    {
      "epoch": 56.48457822370866,
      "grad_norm": 1.6172913312911987,
      "learning_rate": 0.0001,
      "loss": 0.7248,
      "step": 152000
    },
    {
      "epoch": 56.670382757339276,
      "grad_norm": 1.5940032005310059,
      "learning_rate": 0.0001,
      "loss": 0.7336,
      "step": 152500
    },
    {
      "epoch": 56.8561872909699,
      "grad_norm": 1.6969693899154663,
      "learning_rate": 0.0001,
      "loss": 0.729,
      "step": 153000
    },
    {
      "epoch": 57.0,
      "eval_accuracy": 0.7669204369261859,
      "eval_f1_macro": 0.43847119749006624,
      "eval_f1_micro": 0.7669204369261859,
      "eval_loss": 0.6640397310256958,
      "eval_runtime": 484.1334,
      "eval_samples_per_second": 118.564,
      "eval_steps_per_second": 1.853,
      "learning_rate": 0.0001,
      "step": 153387
    },
    {
      "epoch": 57.04199182460052,
      "grad_norm": 1.940199851989746,
      "learning_rate": 0.0001,
      "loss": 0.7136,
      "step": 153500
    },
    {
      "epoch": 57.22779635823114,
      "grad_norm": 2.240196466445923,
      "learning_rate": 0.0001,
      "loss": 0.7264,
      "step": 154000
    },
    {
      "epoch": 57.41360089186176,
      "grad_norm": 1.6355632543563843,
      "learning_rate": 0.0001,
      "loss": 0.7226,
      "step": 154500
    },
    {
      "epoch": 57.599405425492385,
      "grad_norm": 1.930558681488037,
      "learning_rate": 0.0001,
      "loss": 0.7322,
      "step": 155000
    },
    {
      "epoch": 57.785209959123,
      "grad_norm": 2.192221164703369,
      "learning_rate": 0.0001,
      "loss": 0.7322,
      "step": 155500
    },
    {
      "epoch": 57.971014492753625,
      "grad_norm": 1.9269661903381348,
      "learning_rate": 0.0001,
      "loss": 0.7179,
      "step": 156000
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.7672862842110765,
      "eval_f1_macro": 0.43760916892251167,
      "eval_f1_micro": 0.7672862842110765,
      "eval_loss": 0.6627684235572815,
      "eval_runtime": 479.7239,
      "eval_samples_per_second": 119.654,
      "eval_steps_per_second": 1.87,
      "learning_rate": 0.0001,
      "step": 156078
    },
    {
      "epoch": 58.15681902638424,
      "grad_norm": 2.0105526447296143,
      "learning_rate": 0.0001,
      "loss": 0.7321,
      "step": 156500
    },
    {
      "epoch": 58.342623560014864,
      "grad_norm": 1.7992244958877563,
      "learning_rate": 0.0001,
      "loss": 0.7319,
      "step": 157000
    },
    {
      "epoch": 58.52842809364549,
      "grad_norm": 2.2000529766082764,
      "learning_rate": 0.0001,
      "loss": 0.7215,
      "step": 157500
    },
    {
      "epoch": 58.714232627276104,
      "grad_norm": 2.374115228652954,
      "learning_rate": 0.0001,
      "loss": 0.7189,
      "step": 158000
    },
    {
      "epoch": 58.90003716090673,
      "grad_norm": 2.0600476264953613,
      "learning_rate": 0.0001,
      "loss": 0.7257,
      "step": 158500
    },
    {
      "epoch": 59.0,
      "eval_accuracy": 0.7679134509851745,
      "eval_f1_macro": 0.439932052501894,
      "eval_f1_micro": 0.7679134509851745,
      "eval_loss": 0.6614954471588135,
      "eval_runtime": 476.1801,
      "eval_samples_per_second": 120.545,
      "eval_steps_per_second": 1.884,
      "learning_rate": 0.0001,
      "step": 158769
    },
    {
      "epoch": 59.08584169453734,
      "grad_norm": 1.518589735031128,
      "learning_rate": 0.0001,
      "loss": 0.7106,
      "step": 159000
    },
    {
      "epoch": 59.27164622816797,
      "grad_norm": 1.8264387845993042,
      "learning_rate": 0.0001,
      "loss": 0.7262,
      "step": 159500
    },
    {
      "epoch": 59.45745076179859,
      "grad_norm": 2.5099358558654785,
      "learning_rate": 0.0001,
      "loss": 0.7217,
      "step": 160000
    },
    {
      "epoch": 59.643255295429206,
      "grad_norm": 2.0228655338287354,
      "learning_rate": 0.0001,
      "loss": 0.7234,
      "step": 160500
    },
    {
      "epoch": 59.82905982905983,
      "grad_norm": 1.8197671175003052,
      "learning_rate": 0.0001,
      "loss": 0.7297,
      "step": 161000
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.766990122123308,
      "eval_f1_macro": 0.44188579219640917,
      "eval_f1_micro": 0.766990122123308,
      "eval_loss": 0.6633245944976807,
      "eval_runtime": 478.7046,
      "eval_samples_per_second": 119.909,
      "eval_steps_per_second": 1.874,
      "learning_rate": 0.0001,
      "step": 161460
    },
    {
      "epoch": 60.01486436269045,
      "grad_norm": 1.6832709312438965,
      "learning_rate": 0.0001,
      "loss": 0.7319,
      "step": 161500
    },
    {
      "epoch": 60.20066889632107,
      "grad_norm": 2.1717429161071777,
      "learning_rate": 0.0001,
      "loss": 0.7281,
      "step": 162000
    },
    {
      "epoch": 60.38647342995169,
      "grad_norm": 2.8050575256347656,
      "learning_rate": 0.0001,
      "loss": 0.7216,
      "step": 162500
    },
    {
      "epoch": 60.57227796358231,
      "grad_norm": 2.1643614768981934,
      "learning_rate": 0.0001,
      "loss": 0.7272,
      "step": 163000
    },
    {
      "epoch": 60.75808249721293,
      "grad_norm": 2.2237448692321777,
      "learning_rate": 0.0001,
      "loss": 0.7238,
      "step": 163500
    },
    {
      "epoch": 60.943887030843555,
      "grad_norm": 2.3279480934143066,
      "learning_rate": 0.0001,
      "loss": 0.7297,
      "step": 164000
    },
    {
      "epoch": 61.0,
      "eval_accuracy": 0.7685754603578335,
      "eval_f1_macro": 0.4370683373238057,
      "eval_f1_micro": 0.7685754603578335,
      "eval_loss": 0.6611309051513672,
      "eval_runtime": 490.7606,
      "eval_samples_per_second": 116.963,
      "eval_steps_per_second": 1.828,
      "learning_rate": 0.0001,
      "step": 164151
    },
    {
      "epoch": 61.12969156447417,
      "grad_norm": 1.1877753734588623,
      "learning_rate": 0.0001,
      "loss": 0.721,
      "step": 164500
    },
    {
      "epoch": 61.315496098104795,
      "grad_norm": 1.6631951332092285,
      "learning_rate": 0.0001,
      "loss": 0.7222,
      "step": 165000
    },
    {
      "epoch": 61.50130063173541,
      "grad_norm": 1.9209131002426147,
      "learning_rate": 0.0001,
      "loss": 0.7283,
      "step": 165500
    },
    {
      "epoch": 61.687105165366034,
      "grad_norm": 2.088667392730713,
      "learning_rate": 0.0001,
      "loss": 0.7217,
      "step": 166000
    },
    {
      "epoch": 61.87290969899666,
      "grad_norm": 2.503835916519165,
      "learning_rate": 0.0001,
      "loss": 0.7262,
      "step": 166500
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.7684360899635895,
      "eval_f1_macro": 0.45352172583851785,
      "eval_f1_micro": 0.7684360899635895,
      "eval_loss": 0.660831093788147,
      "eval_runtime": 497.2138,
      "eval_samples_per_second": 115.445,
      "eval_steps_per_second": 1.804,
      "learning_rate": 0.0001,
      "step": 166842
    },
    {
      "epoch": 62.058714232627274,
      "grad_norm": 1.5878970623016357,
      "learning_rate": 0.0001,
      "loss": 0.7166,
      "step": 167000
    },
    {
      "epoch": 62.2445187662579,
      "grad_norm": 1.8134915828704834,
      "learning_rate": 0.0001,
      "loss": 0.7234,
      "step": 167500
    },
    {
      "epoch": 62.43032329988852,
      "grad_norm": 1.9176697731018066,
      "learning_rate": 0.0001,
      "loss": 0.7246,
      "step": 168000
    },
    {
      "epoch": 62.61612783351914,
      "grad_norm": 1.6024200916290283,
      "learning_rate": 0.0001,
      "loss": 0.7322,
      "step": 168500
    },
    {
      "epoch": 62.80193236714976,
      "grad_norm": 1.7893162965774536,
      "learning_rate": 0.0001,
      "loss": 0.7227,
      "step": 169000
    },
    {
      "epoch": 62.98773690078038,
      "grad_norm": 1.9036942720413208,
      "learning_rate": 0.0001,
      "loss": 0.7204,
      "step": 169500
    },
    {
      "epoch": 63.0,
      "eval_accuracy": 0.767791501890211,
      "eval_f1_macro": 0.44611945476580944,
      "eval_f1_micro": 0.767791501890211,
      "eval_loss": 0.6621896028518677,
      "eval_runtime": 505.1543,
      "eval_samples_per_second": 113.631,
      "eval_steps_per_second": 1.776,
      "learning_rate": 0.0001,
      "step": 169533
    },
    {
      "epoch": 63.173541434411,
      "grad_norm": 2.1570420265197754,
      "learning_rate": 0.0001,
      "loss": 0.7229,
      "step": 170000
    },
    {
      "epoch": 63.35934596804162,
      "grad_norm": 1.816713571548462,
      "learning_rate": 0.0001,
      "loss": 0.7209,
      "step": 170500
    },
    {
      "epoch": 63.54515050167224,
      "grad_norm": 1.9060814380645752,
      "learning_rate": 0.0001,
      "loss": 0.7279,
      "step": 171000
    },
    {
      "epoch": 63.73095503530286,
      "grad_norm": 1.8209328651428223,
      "learning_rate": 0.0001,
      "loss": 0.726,
      "step": 171500
    },
    {
      "epoch": 63.91675956893348,
      "grad_norm": 2.5898454189300537,
      "learning_rate": 0.0001,
      "loss": 0.7296,
      "step": 172000
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.767547603700284,
      "eval_f1_macro": 0.4439334258360575,
      "eval_f1_micro": 0.767547603700284,
      "eval_loss": 0.6610415577888489,
      "eval_runtime": 518.2702,
      "eval_samples_per_second": 110.755,
      "eval_steps_per_second": 1.731,
      "learning_rate": 0.0001,
      "step": 172224
    },
    {
      "epoch": 64.1025641025641,
      "grad_norm": 2.03393816947937,
      "learning_rate": 0.0001,
      "loss": 0.7116,
      "step": 172500
    },
    {
      "epoch": 64.28836863619472,
      "grad_norm": 2.2850654125213623,
      "learning_rate": 0.0001,
      "loss": 0.7204,
      "step": 173000
    },
    {
      "epoch": 64.47417316982535,
      "grad_norm": 2.1166341304779053,
      "learning_rate": 0.0001,
      "loss": 0.7139,
      "step": 173500
    },
    {
      "epoch": 64.65997770345597,
      "grad_norm": 2.2916886806488037,
      "learning_rate": 0.0001,
      "loss": 0.7235,
      "step": 174000
    },
    {
      "epoch": 64.84578223708658,
      "grad_norm": 2.0275001525878906,
      "learning_rate": 0.0001,
      "loss": 0.7253,
      "step": 174500
    },
    {
      "epoch": 65.0,
      "eval_accuracy": 0.7680354000801379,
      "eval_f1_macro": 0.434573899819693,
      "eval_f1_micro": 0.7680354000801379,
      "eval_loss": 0.6589834690093994,
      "eval_runtime": 514.9526,
      "eval_samples_per_second": 111.469,
      "eval_steps_per_second": 1.742,
      "learning_rate": 0.0001,
      "step": 174915
    },
    {
      "epoch": 65.03158677071721,
      "grad_norm": 1.5037002563476562,
      "learning_rate": 0.0001,
      "loss": 0.7336,
      "step": 175000
    },
    {
      "epoch": 65.21739130434783,
      "grad_norm": 1.769515037536621,
      "learning_rate": 0.0001,
      "loss": 0.7181,
      "step": 175500
    },
    {
      "epoch": 65.40319583797844,
      "grad_norm": 2.0847535133361816,
      "learning_rate": 0.0001,
      "loss": 0.7236,
      "step": 176000
    },
    {
      "epoch": 65.58900037160906,
      "grad_norm": 2.0402798652648926,
      "learning_rate": 0.0001,
      "loss": 0.7196,
      "step": 176500
    },
    {
      "epoch": 65.77480490523969,
      "grad_norm": 2.2478229999542236,
      "learning_rate": 0.0001,
      "loss": 0.7277,
      "step": 177000
    },
    {
      "epoch": 65.9606094388703,
      "grad_norm": 2.4031426906585693,
      "learning_rate": 0.0001,
      "loss": 0.723,
      "step": 177500
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.76845351126287,
      "eval_f1_macro": 0.4397010901020239,
      "eval_f1_micro": 0.76845351126287,
      "eval_loss": 0.6599727272987366,
      "eval_runtime": 512.5743,
      "eval_samples_per_second": 111.986,
      "eval_steps_per_second": 1.75,
      "learning_rate": 0.0001,
      "step": 177606
    },
    {
      "epoch": 66.14641397250092,
      "grad_norm": 1.965077519416809,
      "learning_rate": 0.0001,
      "loss": 0.7167,
      "step": 178000
    },
    {
      "epoch": 66.33221850613155,
      "grad_norm": 2.626265525817871,
      "learning_rate": 0.0001,
      "loss": 0.7315,
      "step": 178500
    },
    {
      "epoch": 66.51802303976217,
      "grad_norm": 2.360399007797241,
      "learning_rate": 0.0001,
      "loss": 0.7101,
      "step": 179000
    },
    {
      "epoch": 66.70382757339279,
      "grad_norm": 2.442690849304199,
      "learning_rate": 0.0001,
      "loss": 0.7336,
      "step": 179500
    },
    {
      "epoch": 66.88963210702342,
      "grad_norm": 1.652727723121643,
      "learning_rate": 0.0001,
      "loss": 0.7257,
      "step": 180000
    },
    {
      "epoch": 67.0,
      "eval_accuracy": 0.769045835438407,
      "eval_f1_macro": 0.44838573955584105,
      "eval_f1_micro": 0.769045835438407,
      "eval_loss": 0.6572328209877014,
      "eval_runtime": 509.8483,
      "eval_samples_per_second": 112.584,
      "eval_steps_per_second": 1.759,
      "learning_rate": 0.0001,
      "step": 180297
    },
    {
      "epoch": 67.07543664065403,
      "grad_norm": 2.3557684421539307,
      "learning_rate": 0.0001,
      "loss": 0.7238,
      "step": 180500
    },
    {
      "epoch": 67.26124117428465,
      "grad_norm": 2.1195363998413086,
      "learning_rate": 0.0001,
      "loss": 0.718,
      "step": 181000
    },
    {
      "epoch": 67.44704570791528,
      "grad_norm": 2.189373731613159,
      "learning_rate": 0.0001,
      "loss": 0.7148,
      "step": 181500
    },
    {
      "epoch": 67.6328502415459,
      "grad_norm": 2.0466713905334473,
      "learning_rate": 0.0001,
      "loss": 0.7295,
      "step": 182000
    },
    {
      "epoch": 67.81865477517651,
      "grad_norm": 2.2755849361419678,
      "learning_rate": 0.0001,
      "loss": 0.7257,
      "step": 182500
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.7686799881535165,
      "eval_f1_macro": 0.4442341389313245,
      "eval_f1_micro": 0.7686799881535165,
      "eval_loss": 0.658860445022583,
      "eval_runtime": 508.6729,
      "eval_samples_per_second": 112.845,
      "eval_steps_per_second": 1.763,
      "learning_rate": 0.0001,
      "step": 182988
    },
    {
      "epoch": 68.00445930880713,
      "grad_norm": 2.214465618133545,
      "learning_rate": 0.0001,
      "loss": 0.7245,
      "step": 183000
    },
    {
      "epoch": 68.19026384243776,
      "grad_norm": 2.169665813446045,
      "learning_rate": 0.0001,
      "loss": 0.7131,
      "step": 183500
    },
    {
      "epoch": 68.37606837606837,
      "grad_norm": 2.195838451385498,
      "learning_rate": 0.0001,
      "loss": 0.7217,
      "step": 184000
    },
    {
      "epoch": 68.56187290969899,
      "grad_norm": 2.631565809249878,
      "learning_rate": 0.0001,
      "loss": 0.714,
      "step": 184500
    },
    {
      "epoch": 68.74767744332962,
      "grad_norm": 2.0778391361236572,
      "learning_rate": 0.0001,
      "loss": 0.7235,
      "step": 185000
    },
    {
      "epoch": 68.93348197696024,
      "grad_norm": 2.478853464126587,
      "learning_rate": 0.0001,
      "loss": 0.7299,
      "step": 185500
    },
    {
      "epoch": 69.0,
      "eval_accuracy": 0.7688542011463215,
      "eval_f1_macro": 0.43926108990057783,
      "eval_f1_micro": 0.7688542011463215,
      "eval_loss": 0.659292995929718,
      "eval_runtime": 515.951,
      "eval_samples_per_second": 111.253,
      "eval_steps_per_second": 1.739,
      "learning_rate": 0.0001,
      "step": 185679
    },
    {
      "epoch": 69.11928651059085,
      "grad_norm": 2.1239874362945557,
      "learning_rate": 0.0001,
      "loss": 0.7327,
      "step": 186000
    },
    {
      "epoch": 69.30509104422148,
      "grad_norm": 2.156639814376831,
      "learning_rate": 0.0001,
      "loss": 0.7178,
      "step": 186500
    },
    {
      "epoch": 69.4908955778521,
      "grad_norm": 2.094212532043457,
      "learning_rate": 0.0001,
      "loss": 0.7205,
      "step": 187000
    },
    {
      "epoch": 69.67670011148272,
      "grad_norm": 2.293548822402954,
      "learning_rate": 0.0001,
      "loss": 0.7275,
      "step": 187500
    },
    {
      "epoch": 69.86250464511335,
      "grad_norm": 2.42189359664917,
      "learning_rate": 0.0001,
      "loss": 0.7289,
      "step": 188000
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.7679134509851745,
      "eval_f1_macro": 0.4357439201261406,
      "eval_f1_micro": 0.7679134509851745,
      "eval_loss": 0.658970832824707,
      "eval_runtime": 507.1382,
      "eval_samples_per_second": 113.186,
      "eval_steps_per_second": 1.769,
      "learning_rate": 0.0001,
      "step": 188370
    },
    {
      "epoch": 70.04830917874396,
      "grad_norm": 2.229328155517578,
      "learning_rate": 0.0001,
      "loss": 0.7243,
      "step": 188500
    },
    {
      "epoch": 70.23411371237458,
      "grad_norm": 2.238511323928833,
      "learning_rate": 0.0001,
      "loss": 0.7194,
      "step": 189000
    },
    {
      "epoch": 70.4199182460052,
      "grad_norm": 2.1684393882751465,
      "learning_rate": 0.0001,
      "loss": 0.7217,
      "step": 189500
    },
    {
      "epoch": 70.60572277963583,
      "grad_norm": 2.2486910820007324,
      "learning_rate": 0.0001,
      "loss": 0.7249,
      "step": 190000
    },
    {
      "epoch": 70.79152731326644,
      "grad_norm": 2.3368637561798096,
      "learning_rate": 0.0001,
      "loss": 0.7169,
      "step": 190500
    },
    {
      "epoch": 70.97733184689706,
      "grad_norm": 2.2436201572418213,
      "learning_rate": 0.0001,
      "loss": 0.7179,
      "step": 191000
    },
    {
      "epoch": 71.0,
      "eval_accuracy": 0.768244455671504,
      "eval_f1_macro": 0.4432082950234043,
      "eval_f1_micro": 0.768244455671504,
      "eval_loss": 0.6567061543464661,
      "eval_runtime": 510.955,
      "eval_samples_per_second": 112.341,
      "eval_steps_per_second": 1.756,
      "learning_rate": 0.0001,
      "step": 191061
    },
    {
      "epoch": 71.16313638052769,
      "grad_norm": 2.196202039718628,
      "learning_rate": 0.0001,
      "loss": 0.7218,
      "step": 191500
    },
    {
      "epoch": 71.3489409141583,
      "grad_norm": 1.8094837665557861,
      "learning_rate": 0.0001,
      "loss": 0.7271,
      "step": 192000
    },
    {
      "epoch": 71.53474544778892,
      "grad_norm": 2.3732166290283203,
      "learning_rate": 0.0001,
      "loss": 0.7135,
      "step": 192500
    },
    {
      "epoch": 71.72054998141955,
      "grad_norm": 2.5137126445770264,
      "learning_rate": 0.0001,
      "loss": 0.7195,
      "step": 193000
    },
    {
      "epoch": 71.90635451505017,
      "grad_norm": 1.83713960647583,
      "learning_rate": 0.0001,
      "loss": 0.7292,
      "step": 193500
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.7681225065765405,
      "eval_f1_macro": 0.4369170898714745,
      "eval_f1_micro": 0.7681225065765405,
      "eval_loss": 0.65887850522995,
      "eval_runtime": 522.7505,
      "eval_samples_per_second": 109.806,
      "eval_steps_per_second": 1.716,
      "learning_rate": 0.0001,
      "step": 193752
    },
    {
      "epoch": 72.09215904868078,
      "grad_norm": 2.539991855621338,
      "learning_rate": 0.0001,
      "loss": 0.7165,
      "step": 194000
    },
    {
      "epoch": 72.27796358231141,
      "grad_norm": 2.3056955337524414,
      "learning_rate": 0.0001,
      "loss": 0.7267,
      "step": 194500
    },
    {
      "epoch": 72.46376811594203,
      "grad_norm": 2.263315200805664,
      "learning_rate": 0.0001,
      "loss": 0.7303,
      "step": 195000
    },
    {
      "epoch": 72.64957264957265,
      "grad_norm": 2.132422685623169,
      "learning_rate": 0.0001,
      "loss": 0.7237,
      "step": 195500
    },
    {
      "epoch": 72.83537718320326,
      "grad_norm": 2.034406900405884,
      "learning_rate": 0.0001,
      "loss": 0.7139,
      "step": 196000
    },
    {
      "epoch": 73.0,
      "eval_accuracy": 0.7675998675981255,
      "eval_f1_macro": 0.4436833314710758,
      "eval_f1_micro": 0.7675998675981255,
      "eval_loss": 0.6611541509628296,
      "eval_runtime": 506.2664,
      "eval_samples_per_second": 113.381,
      "eval_steps_per_second": 1.772,
      "learning_rate": 0.0001,
      "step": 196443
    },
    {
      "epoch": 73.0211817168339,
      "grad_norm": 1.860060691833496,
      "learning_rate": 0.0001,
      "loss": 0.7214,
      "step": 196500
    },
    {
      "epoch": 73.20698625046451,
      "grad_norm": 2.3336594104766846,
      "learning_rate": 0.0001,
      "loss": 0.7223,
      "step": 197000
    },
    {
      "epoch": 73.39279078409513,
      "grad_norm": 2.637392282485962,
      "learning_rate": 0.0001,
      "loss": 0.7284,
      "step": 197500
    },
    {
      "epoch": 73.57859531772576,
      "grad_norm": 1.9666019678115845,
      "learning_rate": 0.0001,
      "loss": 0.7141,
      "step": 198000
    },
    {
      "epoch": 73.76439985135637,
      "grad_norm": 2.4116933345794678,
      "learning_rate": 0.0001,
      "loss": 0.7158,
      "step": 198500
    },
    {
      "epoch": 73.95020438498699,
      "grad_norm": 2.568808078765869,
      "learning_rate": 0.0001,
      "loss": 0.7307,
      "step": 199000
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.768488353861431,
      "eval_f1_macro": 0.44906341810873124,
      "eval_f1_micro": 0.768488353861431,
      "eval_loss": 0.6570971012115479,
      "eval_runtime": 519.8036,
      "eval_samples_per_second": 110.428,
      "eval_steps_per_second": 1.726,
      "learning_rate": 0.0001,
      "step": 199134
    },
    {
      "epoch": 74.13600891861762,
      "grad_norm": 2.4878387451171875,
      "learning_rate": 0.0001,
      "loss": 0.7172,
      "step": 199500
    },
    {
      "epoch": 74.32181345224824,
      "grad_norm": 2.029940605163574,
      "learning_rate": 0.0001,
      "loss": 0.713,
      "step": 200000
    },
    {
      "epoch": 74.50761798587885,
      "grad_norm": 1.8782607316970825,
      "learning_rate": 0.0001,
      "loss": 0.7148,
      "step": 200500
    },
    {
      "epoch": 74.69342251950948,
      "grad_norm": 2.0745551586151123,
      "learning_rate": 0.0001,
      "loss": 0.7249,
      "step": 201000
    },
    {
      "epoch": 74.8792270531401,
      "grad_norm": 2.2313883304595947,
      "learning_rate": 0.0001,
      "loss": 0.7238,
      "step": 201500
    },
    {
      "epoch": 75.0,
      "eval_accuracy": 0.768174770474382,
      "eval_f1_macro": 0.44439364748025234,
      "eval_f1_micro": 0.768174770474382,
      "eval_loss": 0.6557245254516602,
      "eval_runtime": 535.3584,
      "eval_samples_per_second": 107.22,
      "eval_steps_per_second": 1.676,
      "learning_rate": 0.0001,
      "step": 201825
    },
    {
      "epoch": 75.06503158677071,
      "grad_norm": 2.656343936920166,
      "learning_rate": 0.0001,
      "loss": 0.7334,
      "step": 202000
    },
    {
      "epoch": 75.25083612040133,
      "grad_norm": 2.0540146827697754,
      "learning_rate": 0.0001,
      "loss": 0.7186,
      "step": 202500
    },
    {
      "epoch": 75.43664065403196,
      "grad_norm": 1.8651381731033325,
      "learning_rate": 0.0001,
      "loss": 0.723,
      "step": 203000
    },
    {
      "epoch": 75.62244518766258,
      "grad_norm": 1.9370218515396118,
      "learning_rate": 0.0001,
      "loss": 0.7211,
      "step": 203500
    },
    {
      "epoch": 75.8082497212932,
      "grad_norm": 1.9121958017349243,
      "learning_rate": 0.0001,
      "loss": 0.7205,
      "step": 204000
    },
    {
      "epoch": 75.99405425492382,
      "grad_norm": 2.61309552192688,
      "learning_rate": 0.0001,
      "loss": 0.7257,
      "step": 204500
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.7683489834671869,
      "eval_f1_macro": 0.4478607285233603,
      "eval_f1_micro": 0.7683489834671869,
      "eval_loss": 0.658838152885437,
      "eval_runtime": 499.8427,
      "eval_samples_per_second": 114.838,
      "eval_steps_per_second": 1.795,
      "learning_rate": 0.0001,
      "step": 204516
    },
    {
      "epoch": 76.17985878855444,
      "grad_norm": 2.235535144805908,
      "learning_rate": 0.0001,
      "loss": 0.7205,
      "step": 205000
    },
    {
      "epoch": 76.36566332218506,
      "grad_norm": 1.883389949798584,
      "learning_rate": 0.0001,
      "loss": 0.7299,
      "step": 205500
    },
    {
      "epoch": 76.55146785581569,
      "grad_norm": 2.6287364959716797,
      "learning_rate": 0.0001,
      "loss": 0.7118,
      "step": 206000
    },
    {
      "epoch": 76.7372723894463,
      "grad_norm": 2.0496041774749756,
      "learning_rate": 0.0001,
      "loss": 0.72,
      "step": 206500
    },
    {
      "epoch": 76.92307692307692,
      "grad_norm": 2.1624155044555664,
      "learning_rate": 0.0001,
      "loss": 0.7252,
      "step": 207000
    },
    {
      "epoch": 77.0,
      "eval_accuracy": 0.7686277242556749,
      "eval_f1_macro": 0.44887316801028765,
      "eval_f1_micro": 0.7686277242556749,
      "eval_loss": 0.6572225093841553,
      "eval_runtime": 492.6745,
      "eval_samples_per_second": 116.509,
      "eval_steps_per_second": 1.821,
      "learning_rate": 0.0001,
      "step": 207207
    },
    {
      "epoch": 77.10888145670755,
      "grad_norm": 2.232707977294922,
      "learning_rate": 0.0001,
      "loss": 0.7144,
      "step": 207500
    },
    {
      "epoch": 77.29468599033817,
      "grad_norm": 2.15569806098938,
      "learning_rate": 0.0001,
      "loss": 0.7182,
      "step": 208000
    },
    {
      "epoch": 77.48049052396878,
      "grad_norm": 2.2068002223968506,
      "learning_rate": 0.0001,
      "loss": 0.7128,
      "step": 208500
    },
    {
      "epoch": 77.6662950575994,
      "grad_norm": 2.9470760822296143,
      "learning_rate": 0.0001,
      "loss": 0.7247,
      "step": 209000
    },
    {
      "epoch": 77.85209959123003,
      "grad_norm": 2.5690219402313232,
      "learning_rate": 0.0001,
      "loss": 0.7231,
      "step": 209500
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.768767094649919,
      "eval_f1_macro": 0.4439839848601264,
      "eval_f1_micro": 0.768767094649919,
      "eval_loss": 0.6562930941581726,
      "eval_runtime": 495.5045,
      "eval_samples_per_second": 115.844,
      "eval_steps_per_second": 1.81,
      "learning_rate": 0.0001,
      "step": 209898
    },
    {
      "epoch": 78.03790412486065,
      "grad_norm": 1.879889726638794,
      "learning_rate": 0.0001,
      "loss": 0.7258,
      "step": 210000
    },
    {
      "epoch": 78.22370865849126,
      "grad_norm": 1.8929120302200317,
      "learning_rate": 0.0001,
      "loss": 0.7188,
      "step": 210500
    },
    {
      "epoch": 78.40951319212189,
      "grad_norm": 2.2578604221343994,
      "learning_rate": 0.0001,
      "loss": 0.7239,
      "step": 211000
    },
    {
      "epoch": 78.59531772575251,
      "grad_norm": 2.1204187870025635,
      "learning_rate": 0.0001,
      "loss": 0.7255,
      "step": 211500
    },
    {
      "epoch": 78.78112225938312,
      "grad_norm": 1.968096375465393,
      "learning_rate": 0.0001,
      "loss": 0.7206,
      "step": 212000
    },
    {
      "epoch": 78.96692679301376,
      "grad_norm": 2.7695274353027344,
      "learning_rate": 0.0001,
      "loss": 0.7207,
      "step": 212500
    },
    {
      "epoch": 79.0,
      "eval_accuracy": 0.76810508527726,
      "eval_f1_macro": 0.4379298766193662,
      "eval_f1_micro": 0.76810508527726,
      "eval_loss": 0.6564787030220032,
      "eval_runtime": 498.2393,
      "eval_samples_per_second": 115.208,
      "eval_steps_per_second": 1.8,
      "learning_rate": 0.0001,
      "step": 212589
    },
    {
      "epoch": 79.15273132664437,
      "grad_norm": 2.0273005962371826,
      "learning_rate": 0.0001,
      "loss": 0.7177,
      "step": 213000
    },
    {
      "epoch": 79.33853586027499,
      "grad_norm": 2.40535306930542,
      "learning_rate": 0.0001,
      "loss": 0.7288,
      "step": 213500
    },
    {
      "epoch": 79.52434039390562,
      "grad_norm": 2.329434871673584,
      "learning_rate": 0.0001,
      "loss": 0.7139,
      "step": 214000
    },
    {
      "epoch": 79.71014492753623,
      "grad_norm": 2.750331401824951,
      "learning_rate": 0.0001,
      "loss": 0.7244,
      "step": 214500
    },
    {
      "epoch": 79.89594946116685,
      "grad_norm": 2.482513427734375,
      "learning_rate": 0.0001,
      "loss": 0.7179,
      "step": 215000
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.768383826065748,
      "eval_f1_macro": 0.4460529321244195,
      "eval_f1_micro": 0.768383826065748,
      "eval_loss": 0.661143958568573,
      "eval_runtime": 505.2043,
      "eval_samples_per_second": 113.619,
      "eval_steps_per_second": 1.776,
      "learning_rate": 0.0001,
      "step": 215280
    },
    {
      "epoch": 80.08175399479747,
      "grad_norm": 2.6754283905029297,
      "learning_rate": 0.0001,
      "loss": 0.7247,
      "step": 215500
    },
    {
      "epoch": 80.2675585284281,
      "grad_norm": 3.016185760498047,
      "learning_rate": 0.0001,
      "loss": 0.7194,
      "step": 216000
    },
    {
      "epoch": 80.45336306205871,
      "grad_norm": 2.425431489944458,
      "learning_rate": 0.0001,
      "loss": 0.7197,
      "step": 216500
    },
    {
      "epoch": 80.63916759568933,
      "grad_norm": 2.2331862449645996,
      "learning_rate": 0.0001,
      "loss": 0.7217,
      "step": 217000
    },
    {
      "epoch": 80.82497212931996,
      "grad_norm": 2.3233511447906494,
      "learning_rate": 0.0001,
      "loss": 0.7275,
      "step": 217500
    },
    {
      "epoch": 81.0,
      "eval_accuracy": 0.768941307642724,
      "eval_f1_macro": 0.44750591322776384,
      "eval_f1_micro": 0.768941307642724,
      "eval_loss": 0.660437285900116,
      "eval_runtime": 497.6996,
      "eval_samples_per_second": 115.333,
      "eval_steps_per_second": 1.802,
      "learning_rate": 0.0001,
      "step": 217971
    },
    {
      "epoch": 81.01077666295058,
      "grad_norm": 2.5221219062805176,
      "learning_rate": 1e-05,
      "loss": 0.7201,
      "step": 218000
    },
    {
      "epoch": 81.19658119658119,
      "grad_norm": 2.2925214767456055,
      "learning_rate": 1e-05,
      "loss": 0.7095,
      "step": 218500
    },
    {
      "epoch": 81.38238573021182,
      "grad_norm": 2.733447313308716,
      "learning_rate": 1e-05,
      "loss": 0.7125,
      "step": 219000
    },
    {
      "epoch": 81.56819026384244,
      "grad_norm": 2.3250811100006104,
      "learning_rate": 1e-05,
      "loss": 0.7193,
      "step": 219500
    },
    {
      "epoch": 81.75399479747306,
      "grad_norm": 2.4225757122039795,
      "learning_rate": 1e-05,
      "loss": 0.7081,
      "step": 220000
    },
    {
      "epoch": 81.93979933110369,
      "grad_norm": 2.1880016326904297,
      "learning_rate": 1e-05,
      "loss": 0.7101,
      "step": 220500
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.7705614884758105,
      "eval_f1_macro": 0.4526720456188751,
      "eval_f1_micro": 0.7705614884758105,
      "eval_loss": 0.6531779766082764,
      "eval_runtime": 493.1117,
      "eval_samples_per_second": 116.406,
      "eval_steps_per_second": 1.819,
      "learning_rate": 1e-05,
      "step": 220662
    },
    {
      "epoch": 82.1256038647343,
      "grad_norm": 2.038079261779785,
      "learning_rate": 1e-05,
      "loss": 0.7052,
      "step": 221000
    },
    {
      "epoch": 82.31140839836492,
      "grad_norm": 2.8099365234375,
      "learning_rate": 1e-05,
      "loss": 0.7054,
      "step": 221500
    },
    {
      "epoch": 82.49721293199553,
      "grad_norm": 2.683091878890991,
      "learning_rate": 1e-05,
      "loss": 0.7013,
      "step": 222000
    },
    {
      "epoch": 82.68301746562616,
      "grad_norm": 2.0887763500213623,
      "learning_rate": 1e-05,
      "loss": 0.7024,
      "step": 222500
    },
    {
      "epoch": 82.86882199925678,
      "grad_norm": 2.346670627593994,
      "learning_rate": 1e-05,
      "loss": 0.7063,
      "step": 223000
    },
    {
      "epoch": 83.0,
      "eval_accuracy": 0.77019564119092,
      "eval_f1_macro": 0.4489367718812771,
      "eval_f1_micro": 0.77019564119092,
      "eval_loss": 0.6532895565032959,
      "eval_runtime": 492.0028,
      "eval_samples_per_second": 116.668,
      "eval_steps_per_second": 1.823,
      "learning_rate": 1e-05,
      "step": 223353
    },
    {
      "epoch": 83.0546265328874,
      "grad_norm": 2.1046745777130127,
      "learning_rate": 1e-05,
      "loss": 0.7073,
      "step": 223500
    },
    {
      "epoch": 83.24043106651803,
      "grad_norm": 1.7634518146514893,
      "learning_rate": 1e-05,
      "loss": 0.7081,
      "step": 224000
    },
    {
      "epoch": 83.42623560014864,
      "grad_norm": 2.451301097869873,
      "learning_rate": 1e-05,
      "loss": 0.7151,
      "step": 224500
    },
    {
      "epoch": 83.61204013377926,
      "grad_norm": 2.347801923751831,
      "learning_rate": 1e-05,
      "loss": 0.6973,
      "step": 225000
    },
    {
      "epoch": 83.79784466740989,
      "grad_norm": 2.34899640083313,
      "learning_rate": 1e-05,
      "loss": 0.7013,
      "step": 225500
    },
    {
      "epoch": 83.9836492010405,
      "grad_norm": 2.269644260406494,
      "learning_rate": 1e-05,
      "loss": 0.7067,
      "step": 226000
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.7705266458772495,
      "eval_f1_macro": 0.45139096558153424,
      "eval_f1_micro": 0.7705266458772495,
      "eval_loss": 0.6505005359649658,
      "eval_runtime": 486.6669,
      "eval_samples_per_second": 117.947,
      "eval_steps_per_second": 1.843,
      "learning_rate": 1e-05,
      "step": 226044
    },
    {
      "epoch": 84.16945373467112,
      "grad_norm": 2.6176042556762695,
      "learning_rate": 1e-05,
      "loss": 0.6912,
      "step": 226500
    },
    {
      "epoch": 84.35525826830175,
      "grad_norm": 2.330493211746216,
      "learning_rate": 1e-05,
      "loss": 0.7045,
      "step": 227000
    },
    {
      "epoch": 84.54106280193237,
      "grad_norm": 2.351470470428467,
      "learning_rate": 1e-05,
      "loss": 0.7049,
      "step": 227500
    },
    {
      "epoch": 84.72686733556299,
      "grad_norm": 2.444443702697754,
      "learning_rate": 1e-05,
      "loss": 0.7122,
      "step": 228000
    },
    {
      "epoch": 84.9126718691936,
      "grad_norm": 2.2119200229644775,
      "learning_rate": 1e-05,
      "loss": 0.707,
      "step": 228500
    },
    {
      "epoch": 85.0,
      "eval_accuracy": 0.7708053866657375,
      "eval_f1_macro": 0.4565625671001629,
      "eval_f1_micro": 0.7708053866657375,
      "eval_loss": 0.6501905918121338,
      "eval_runtime": 485.2188,
      "eval_samples_per_second": 118.299,
      "eval_steps_per_second": 1.849,
      "learning_rate": 1e-05,
      "step": 228735
    },
    {
      "epoch": 85.09847640282423,
      "grad_norm": 2.385796546936035,
      "learning_rate": 1e-05,
      "loss": 0.6966,
      "step": 229000
    },
    {
      "epoch": 85.28428093645485,
      "grad_norm": 2.420588254928589,
      "learning_rate": 1e-05,
      "loss": 0.705,
      "step": 229500
    },
    {
      "epoch": 85.47008547008546,
      "grad_norm": 2.1321732997894287,
      "learning_rate": 1e-05,
      "loss": 0.6976,
      "step": 230000
    },
    {
      "epoch": 85.6558900037161,
      "grad_norm": 2.215148448944092,
      "learning_rate": 1e-05,
      "loss": 0.6998,
      "step": 230500
    },
    {
      "epoch": 85.84169453734671,
      "grad_norm": 2.643437385559082,
      "learning_rate": 1e-05,
      "loss": 0.6944,
      "step": 231000
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.7707182801693351,
      "eval_f1_macro": 0.4559124472677571,
      "eval_f1_micro": 0.7707182801693351,
      "eval_loss": 0.6507149338722229,
      "eval_runtime": 493.7179,
      "eval_samples_per_second": 116.263,
      "eval_steps_per_second": 1.817,
      "learning_rate": 1e-05,
      "step": 231426
    },
    {
      "epoch": 86.02749907097733,
      "grad_norm": 2.8356385231018066,
      "learning_rate": 1e-05,
      "loss": 0.6966,
      "step": 231500
    },
    {
      "epoch": 86.21330360460796,
      "grad_norm": 2.272819757461548,
      "learning_rate": 1e-05,
      "loss": 0.7011,
      "step": 232000
    },
    {
      "epoch": 86.39910813823857,
      "grad_norm": 2.5435869693756104,
      "learning_rate": 1e-05,
      "loss": 0.7019,
      "step": 232500
    },
    {
      "epoch": 86.58491267186919,
      "grad_norm": 2.345691204071045,
      "learning_rate": 1e-05,
      "loss": 0.6964,
      "step": 233000
    },
    {
      "epoch": 86.77071720549982,
      "grad_norm": 1.9666670560836792,
      "learning_rate": 1e-05,
      "loss": 0.6989,
      "step": 233500
    },
    {
      "epoch": 86.95652173913044,
      "grad_norm": 2.101868152618408,
      "learning_rate": 1e-05,
      "loss": 0.6958,
      "step": 234000
    },
    {
      "epoch": 87.0,
      "eval_accuracy": 0.770927335760701,
      "eval_f1_macro": 0.45838476700319497,
      "eval_f1_micro": 0.770927335760701,
      "eval_loss": 0.6484472751617432,
      "eval_runtime": 497.0614,
      "eval_samples_per_second": 115.481,
      "eval_steps_per_second": 1.805,
      "learning_rate": 1e-05,
      "step": 234117
    },
    {
      "epoch": 87.14232627276105,
      "grad_norm": 3.2775235176086426,
      "learning_rate": 1e-05,
      "loss": 0.6957,
      "step": 234500
    },
    {
      "epoch": 87.32813080639167,
      "grad_norm": 2.4724538326263428,
      "learning_rate": 1e-05,
      "loss": 0.6963,
      "step": 235000
    },
    {
      "epoch": 87.5139353400223,
      "grad_norm": 2.9358458518981934,
      "learning_rate": 1e-05,
      "loss": 0.6987,
      "step": 235500
    },
    {
      "epoch": 87.69973987365292,
      "grad_norm": 2.5557122230529785,
      "learning_rate": 1e-05,
      "loss": 0.708,
      "step": 236000
    },
    {
      "epoch": 87.88554440728353,
      "grad_norm": 2.3544118404388428,
      "learning_rate": 1e-05,
      "loss": 0.6967,
      "step": 236500
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.77054406717653,
      "eval_f1_macro": 0.4569340367642959,
      "eval_f1_micro": 0.77054406717653,
      "eval_loss": 0.6496042013168335,
      "eval_runtime": 501.5965,
      "eval_samples_per_second": 114.437,
      "eval_steps_per_second": 1.788,
      "learning_rate": 1e-05,
      "step": 236808
    },
    {
      "epoch": 88.07134894091416,
      "grad_norm": 2.06174898147583,
      "learning_rate": 1e-05,
      "loss": 0.6987,
      "step": 237000
    },
    {
      "epoch": 88.25715347454478,
      "grad_norm": 1.922559380531311,
      "learning_rate": 1e-05,
      "loss": 0.7057,
      "step": 237500
    },
    {
      "epoch": 88.4429580081754,
      "grad_norm": 2.603978157043457,
      "learning_rate": 1e-05,
      "loss": 0.7019,
      "step": 238000
    },
    {
      "epoch": 88.62876254180603,
      "grad_norm": 1.859614372253418,
      "learning_rate": 1e-05,
      "loss": 0.6921,
      "step": 238500
    },
    {
      "epoch": 88.81456707543664,
      "grad_norm": 2.208031177520752,
      "learning_rate": 1e-05,
      "loss": 0.698,
      "step": 239000
    },
    {
      "epoch": 89.0,
      "eval_accuracy": 0.7713977108412745,
      "eval_f1_macro": 0.45601319436503734,
      "eval_f1_micro": 0.7713977108412745,
      "eval_loss": 0.6486304402351379,
      "eval_runtime": 503.0751,
      "eval_samples_per_second": 114.1,
      "eval_steps_per_second": 1.783,
      "learning_rate": 1e-05,
      "step": 239499
    },
    {
      "epoch": 89.00037160906726,
      "grad_norm": 2.6880383491516113,
      "learning_rate": 1e-05,
      "loss": 0.6938,
      "step": 239500
    },
    {
      "epoch": 89.18617614269789,
      "grad_norm": 2.5231385231018066,
      "learning_rate": 1e-05,
      "loss": 0.7063,
      "step": 240000
    },
    {
      "epoch": 89.3719806763285,
      "grad_norm": 2.500443935394287,
      "learning_rate": 1e-05,
      "loss": 0.6931,
      "step": 240500
    },
    {
      "epoch": 89.55778520995912,
      "grad_norm": 2.0993025302886963,
      "learning_rate": 1e-05,
      "loss": 0.7048,
      "step": 241000
    },
    {
      "epoch": 89.74358974358974,
      "grad_norm": 2.4644758701324463,
      "learning_rate": 1e-05,
      "loss": 0.7037,
      "step": 241500
    },
    {
      "epoch": 89.92939427722037,
      "grad_norm": 2.391343116760254,
      "learning_rate": 1e-05,
      "loss": 0.6966,
      "step": 242000
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.7711886552499085,
      "eval_f1_macro": 0.45742417795749246,
      "eval_f1_micro": 0.7711886552499085,
      "eval_loss": 0.6490767598152161,
      "eval_runtime": 499.9247,
      "eval_samples_per_second": 114.819,
      "eval_steps_per_second": 1.794,
      "learning_rate": 1e-05,
      "step": 242190
    },
    {
      "epoch": 90.11519881085098,
      "grad_norm": 2.728985071182251,
      "learning_rate": 1e-05,
      "loss": 0.6981,
      "step": 242500
    },
    {
      "epoch": 90.3010033444816,
      "grad_norm": 2.9933202266693115,
      "learning_rate": 1e-05,
      "loss": 0.6972,
      "step": 243000
    },
    {
      "epoch": 90.48680787811223,
      "grad_norm": 2.6327061653137207,
      "learning_rate": 1e-05,
      "loss": 0.6901,
      "step": 243500
    },
    {
      "epoch": 90.67261241174285,
      "grad_norm": 2.532104969024658,
      "learning_rate": 1e-05,
      "loss": 0.6879,
      "step": 244000
    },
    {
      "epoch": 90.85841694537346,
      "grad_norm": 2.649080753326416,
      "learning_rate": 1e-05,
      "loss": 0.7017,
      "step": 244500
    },
    {
      "epoch": 91.0,
      "eval_accuracy": 0.7704221180815666,
      "eval_f1_macro": 0.45182909085509243,
      "eval_f1_micro": 0.7704221180815666,
      "eval_loss": 0.6481940746307373,
      "eval_runtime": 494.0364,
      "eval_samples_per_second": 116.188,
      "eval_steps_per_second": 1.816,
      "learning_rate": 1e-05,
      "step": 244881
    },
    {
      "epoch": 91.0442214790041,
      "grad_norm": 2.7153379917144775,
      "learning_rate": 1e-05,
      "loss": 0.6977,
      "step": 245000
    },
    {
      "epoch": 91.23002601263471,
      "grad_norm": 2.7882251739501953,
      "learning_rate": 1e-05,
      "loss": 0.6979,
      "step": 245500
    },
    {
      "epoch": 91.41583054626533,
      "grad_norm": 2.632366418838501,
      "learning_rate": 1e-05,
      "loss": 0.6984,
      "step": 246000
    },
    {
      "epoch": 91.60163507989596,
      "grad_norm": 2.7723135948181152,
      "learning_rate": 1e-05,
      "loss": 0.6972,
      "step": 246500
    },
    {
      "epoch": 91.78743961352657,
      "grad_norm": 2.384683132171631,
      "learning_rate": 1e-05,
      "loss": 0.6989,
      "step": 247000
    },
    {
      "epoch": 91.97324414715719,
      "grad_norm": 2.210071086883545,
      "learning_rate": 1e-05,
      "loss": 0.7,
      "step": 247500
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.7715545025347991,
      "eval_f1_macro": 0.45503456574154166,
      "eval_f1_micro": 0.7715545025347991,
      "eval_loss": 0.6477252244949341,
      "eval_runtime": 495.0222,
      "eval_samples_per_second": 115.956,
      "eval_steps_per_second": 1.812,
      "learning_rate": 1e-05,
      "step": 247572
    },
    {
      "epoch": 92.1590486807878,
      "grad_norm": 2.723433494567871,
      "learning_rate": 1e-05,
      "loss": 0.6915,
      "step": 248000
    },
    {
      "epoch": 92.34485321441844,
      "grad_norm": 2.203484535217285,
      "learning_rate": 1e-05,
      "loss": 0.6938,
      "step": 248500
    },
    {
      "epoch": 92.53065774804905,
      "grad_norm": 2.4244284629821777,
      "learning_rate": 1e-05,
      "loss": 0.6994,
      "step": 249000
    },
    {
      "epoch": 92.71646228167967,
      "grad_norm": 2.495445728302002,
      "learning_rate": 1e-05,
      "loss": 0.7034,
      "step": 249500
    },
    {
      "epoch": 92.9022668153103,
      "grad_norm": 2.1715991497039795,
      "learning_rate": 1e-05,
      "loss": 0.7,
      "step": 250000
    },
    {
      "epoch": 93.0,
      "eval_accuracy": 0.771206076549189,
      "eval_f1_macro": 0.4518083431267937,
      "eval_f1_micro": 0.771206076549189,
      "eval_loss": 0.6489835381507874,
      "eval_runtime": 487.2388,
      "eval_samples_per_second": 117.809,
      "eval_steps_per_second": 1.841,
      "learning_rate": 1e-05,
      "step": 250263
    },
    {
      "epoch": 93.08807134894091,
      "grad_norm": 2.252865791320801,
      "learning_rate": 1e-05,
      "loss": 0.6916,
      "step": 250500
    },
    {
      "epoch": 93.27387588257153,
      "grad_norm": 2.484537363052368,
      "learning_rate": 1e-05,
      "loss": 0.6953,
      "step": 251000
    },
    {
      "epoch": 93.45968041620216,
      "grad_norm": 1.7372355461120605,
      "learning_rate": 1e-05,
      "loss": 0.6918,
      "step": 251500
    },
    {
      "epoch": 93.64548494983278,
      "grad_norm": 2.283757448196411,
      "learning_rate": 1e-05,
      "loss": 0.7009,
      "step": 252000
    },
    {
      "epoch": 93.8312894834634,
      "grad_norm": 2.009822130203247,
      "learning_rate": 1e-05,
      "loss": 0.7049,
      "step": 252500
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.770753122767896,
      "eval_f1_macro": 0.45105746851856937,
      "eval_f1_micro": 0.770753122767896,
      "eval_loss": 0.6485304832458496,
      "eval_runtime": 500.9843,
      "eval_samples_per_second": 114.576,
      "eval_steps_per_second": 1.79,
      "learning_rate": 1e-05,
      "step": 252954
    },
    {
      "epoch": 94.01709401709402,
      "grad_norm": 2.0452888011932373,
      "learning_rate": 1e-05,
      "loss": 0.6917,
      "step": 253000
    },
    {
      "epoch": 94.20289855072464,
      "grad_norm": 2.5866355895996094,
      "learning_rate": 1e-05,
      "loss": 0.7064,
      "step": 253500
    },
    {
      "epoch": 94.38870308435526,
      "grad_norm": 2.3897805213928223,
      "learning_rate": 1e-05,
      "loss": 0.6944,
      "step": 254000
    },
    {
      "epoch": 94.57450761798587,
      "grad_norm": 2.8837811946868896,
      "learning_rate": 1e-05,
      "loss": 0.6925,
      "step": 254500
    },
    {
      "epoch": 94.7603121516165,
      "grad_norm": 2.44002628326416,
      "learning_rate": 1e-05,
      "loss": 0.6989,
      "step": 255000
    },
    {
      "epoch": 94.94611668524712,
      "grad_norm": 3.1637446880340576,
      "learning_rate": 1e-05,
      "loss": 0.6949,
      "step": 255500
    },
    {
      "epoch": 95.0,
      "eval_accuracy": 0.771659030330482,
      "eval_f1_macro": 0.45671492126995916,
      "eval_f1_micro": 0.771659030330482,
      "eval_loss": 0.647895336151123,
      "eval_runtime": 490.9837,
      "eval_samples_per_second": 116.91,
      "eval_steps_per_second": 1.827,
      "learning_rate": 1e-05,
      "step": 255645
    },
    {
      "epoch": 95.13192121887774,
      "grad_norm": 2.9006567001342773,
      "learning_rate": 1e-05,
      "loss": 0.6916,
      "step": 256000
    },
    {
      "epoch": 95.31772575250837,
      "grad_norm": 2.7343852519989014,
      "learning_rate": 1e-05,
      "loss": 0.6968,
      "step": 256500
    },
    {
      "epoch": 95.50353028613898,
      "grad_norm": 2.1861000061035156,
      "learning_rate": 1e-05,
      "loss": 0.6921,
      "step": 257000
    },
    {
      "epoch": 95.6893348197696,
      "grad_norm": 2.3142426013946533,
      "learning_rate": 1e-05,
      "loss": 0.7016,
      "step": 257500
    },
    {
      "epoch": 95.87513935340023,
      "grad_norm": 3.253568649291992,
      "learning_rate": 1e-05,
      "loss": 0.6998,
      "step": 258000
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.7715022386369575,
      "eval_f1_macro": 0.4596959338122155,
      "eval_f1_micro": 0.7715022386369575,
      "eval_loss": 0.6472702622413635,
      "eval_runtime": 484.103,
      "eval_samples_per_second": 118.572,
      "eval_steps_per_second": 1.853,
      "learning_rate": 1e-05,
      "step": 258336
    },
    {
      "epoch": 96.06094388703085,
      "grad_norm": 2.4987857341766357,
      "learning_rate": 1e-05,
      "loss": 0.6887,
      "step": 258500
    },
    {
      "epoch": 96.24674842066146,
      "grad_norm": 3.320939064025879,
      "learning_rate": 1e-05,
      "loss": 0.6984,
      "step": 259000
    },
    {
      "epoch": 96.43255295429208,
      "grad_norm": 2.3504180908203125,
      "learning_rate": 1e-05,
      "loss": 0.6984,
      "step": 259500
    },
    {
      "epoch": 96.61835748792271,
      "grad_norm": 2.383805513381958,
      "learning_rate": 1e-05,
      "loss": 0.6963,
      "step": 260000
    },
    {
      "epoch": 96.80416202155332,
      "grad_norm": 2.4699482917785645,
      "learning_rate": 1e-05,
      "loss": 0.6847,
      "step": 260500
    },
    {
      "epoch": 96.98996655518394,
      "grad_norm": 2.491687297821045,
      "learning_rate": 1e-05,
      "loss": 0.6968,
      "step": 261000
    },
    {
      "epoch": 97.0,
      "eval_accuracy": 0.7713977108412745,
      "eval_f1_macro": 0.4625401473178366,
      "eval_f1_micro": 0.7713977108412745,
      "eval_loss": 0.6460831165313721,
      "eval_runtime": 491.1356,
      "eval_samples_per_second": 116.874,
      "eval_steps_per_second": 1.826,
      "learning_rate": 1e-05,
      "step": 261027
    },
    {
      "epoch": 97.17577108881457,
      "grad_norm": 1.9844508171081543,
      "learning_rate": 1e-05,
      "loss": 0.7004,
      "step": 261500
    },
    {
      "epoch": 97.36157562244519,
      "grad_norm": 2.3233275413513184,
      "learning_rate": 1e-05,
      "loss": 0.6867,
      "step": 262000
    },
    {
      "epoch": 97.5473801560758,
      "grad_norm": 2.318446397781372,
      "learning_rate": 1e-05,
      "loss": 0.6992,
      "step": 262500
    },
    {
      "epoch": 97.73318468970643,
      "grad_norm": 2.4279417991638184,
      "learning_rate": 1e-05,
      "loss": 0.6933,
      "step": 263000
    },
    {
      "epoch": 97.91898922333705,
      "grad_norm": 2.6880953311920166,
      "learning_rate": 1e-05,
      "loss": 0.7055,
      "step": 263500
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.7722165119074581,
      "eval_f1_macro": 0.45892155771298887,
      "eval_f1_micro": 0.7722165119074581,
      "eval_loss": 0.6463102698326111,
      "eval_runtime": 504.464,
      "eval_samples_per_second": 113.786,
      "eval_steps_per_second": 1.778,
      "learning_rate": 1e-05,
      "step": 263718
    },
    {
      "epoch": 98.10479375696767,
      "grad_norm": 2.778778553009033,
      "learning_rate": 1e-05,
      "loss": 0.6994,
      "step": 264000
    },
    {
      "epoch": 98.2905982905983,
      "grad_norm": 2.2124552726745605,
      "learning_rate": 1e-05,
      "loss": 0.6842,
      "step": 264500
    },
    {
      "epoch": 98.47640282422891,
      "grad_norm": 3.063041925430298,
      "learning_rate": 1e-05,
      "loss": 0.696,
      "step": 265000
    },
    {
      "epoch": 98.66220735785953,
      "grad_norm": 2.1842856407165527,
      "learning_rate": 1e-05,
      "loss": 0.6972,
      "step": 265500
    },
    {
      "epoch": 98.84801189149015,
      "grad_norm": 3.16538143157959,
      "learning_rate": 1e-05,
      "loss": 0.6931,
      "step": 266000
    },
    {
      "epoch": 99.0,
      "eval_accuracy": 0.77089249316214,
      "eval_f1_macro": 0.4549241891767946,
      "eval_f1_micro": 0.77089249316214,
      "eval_loss": 0.646852433681488,
      "eval_runtime": 498.8103,
      "eval_samples_per_second": 115.076,
      "eval_steps_per_second": 1.798,
      "learning_rate": 1e-05,
      "step": 266409
    },
    {
      "epoch": 99.03381642512078,
      "grad_norm": 2.8139493465423584,
      "learning_rate": 1e-05,
      "loss": 0.6938,
      "step": 266500
    },
    {
      "epoch": 99.21962095875139,
      "grad_norm": 2.5012643337249756,
      "learning_rate": 1e-05,
      "loss": 0.6978,
      "step": 267000
    },
    {
      "epoch": 99.40542549238201,
      "grad_norm": 2.0635502338409424,
      "learning_rate": 1e-05,
      "loss": 0.702,
      "step": 267500
    },
    {
      "epoch": 99.59123002601264,
      "grad_norm": 1.9686059951782227,
      "learning_rate": 1e-05,
      "loss": 0.6867,
      "step": 268000
    },
    {
      "epoch": 99.77703455964325,
      "grad_norm": 2.32633900642395,
      "learning_rate": 1e-05,
      "loss": 0.6916,
      "step": 268500
    },
    {
      "epoch": 99.96283909327387,
      "grad_norm": 2.2013280391693115,
      "learning_rate": 1e-05,
      "loss": 0.6872,
      "step": 269000
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.7723384610024215,
      "eval_f1_macro": 0.45970146016594643,
      "eval_f1_micro": 0.7723384610024215,
      "eval_loss": 0.6456441879272461,
      "eval_runtime": 523.4938,
      "eval_samples_per_second": 109.65,
      "eval_steps_per_second": 1.713,
      "learning_rate": 1e-05,
      "step": 269100
    },
    {
      "epoch": 100.1486436269045,
      "grad_norm": 2.9276328086853027,
      "learning_rate": 1e-05,
      "loss": 0.6936,
      "step": 269500
    },
    {
      "epoch": 100.33444816053512,
      "grad_norm": 2.9882853031158447,
      "learning_rate": 1e-05,
      "loss": 0.6916,
      "step": 270000
    },
    {
      "epoch": 100.52025269416573,
      "grad_norm": 2.50858998298645,
      "learning_rate": 1e-05,
      "loss": 0.7004,
      "step": 270500
    },
    {
      "epoch": 100.70605722779636,
      "grad_norm": 2.6219215393066406,
      "learning_rate": 1e-05,
      "loss": 0.7033,
      "step": 271000
    },
    {
      "epoch": 100.89186176142698,
      "grad_norm": 2.8792617321014404,
      "learning_rate": 1e-05,
      "loss": 0.6822,
      "step": 271500
    },
    {
      "epoch": 101.0,
      "eval_accuracy": 0.7717461368268845,
      "eval_f1_macro": 0.4573593202819609,
      "eval_f1_micro": 0.7717461368268845,
      "eval_loss": 0.6469387412071228,
      "eval_runtime": 517.8794,
      "eval_samples_per_second": 110.839,
      "eval_steps_per_second": 1.732,
      "learning_rate": 1e-05,
      "step": 271791
    },
    {
      "epoch": 101.0776662950576,
      "grad_norm": 2.3658642768859863,
      "learning_rate": 1e-05,
      "loss": 0.707,
      "step": 272000
    },
    {
      "epoch": 101.26347082868821,
      "grad_norm": 3.0172219276428223,
      "learning_rate": 1e-05,
      "loss": 0.6982,
      "step": 272500
    },
    {
      "epoch": 101.44927536231884,
      "grad_norm": 3.0194907188415527,
      "learning_rate": 1e-05,
      "loss": 0.6899,
      "step": 273000
    },
    {
      "epoch": 101.63507989594946,
      "grad_norm": 2.587480306625366,
      "learning_rate": 1e-05,
      "loss": 0.6917,
      "step": 273500
    },
    {
      "epoch": 101.82088442958008,
      "grad_norm": 2.457369089126587,
      "learning_rate": 1e-05,
      "loss": 0.6875,
      "step": 274000
    },
    {
      "epoch": 102.0,
      "eval_accuracy": 0.7717809794254455,
      "eval_f1_macro": 0.4593480600391769,
      "eval_f1_micro": 0.7717809794254455,
      "eval_loss": 0.646738588809967,
      "eval_runtime": 515.0733,
      "eval_samples_per_second": 111.442,
      "eval_steps_per_second": 1.741,
      "learning_rate": 1e-05,
      "step": 274482
    },
    {
      "epoch": 102.0066889632107,
      "grad_norm": 3.1219208240509033,
      "learning_rate": 1e-05,
      "loss": 0.6994,
      "step": 274500
    },
    {
      "epoch": 102.19249349684132,
      "grad_norm": 2.158254623413086,
      "learning_rate": 1e-05,
      "loss": 0.6944,
      "step": 275000
    },
    {
      "epoch": 102.37829803047194,
      "grad_norm": 2.474193811416626,
      "learning_rate": 1e-05,
      "loss": 0.6939,
      "step": 275500
    },
    {
      "epoch": 102.56410256410257,
      "grad_norm": 2.223621129989624,
      "learning_rate": 1e-05,
      "loss": 0.7028,
      "step": 276000
    },
    {
      "epoch": 102.74990709773319,
      "grad_norm": 2.8731536865234375,
      "learning_rate": 1e-05,
      "loss": 0.6972,
      "step": 276500
    },
    {
      "epoch": 102.9357116313638,
      "grad_norm": 1.9848276376724243,
      "learning_rate": 1e-05,
      "loss": 0.6983,
      "step": 277000
    },
    {
      "epoch": 103.0,
      "eval_accuracy": 0.7723036184038605,
      "eval_f1_macro": 0.4576617106244536,
      "eval_f1_micro": 0.7723036184038605,
      "eval_loss": 0.6467755436897278,
      "eval_runtime": 512.6943,
      "eval_samples_per_second": 111.959,
      "eval_steps_per_second": 1.75,
      "learning_rate": 1e-05,
      "step": 277173
    },
    {
      "epoch": 103.12151616499443,
      "grad_norm": 2.9539742469787598,
      "learning_rate": 1e-05,
      "loss": 0.6854,
      "step": 277500
    },
    {
      "epoch": 103.30732069862505,
      "grad_norm": 3.1808767318725586,
      "learning_rate": 1e-05,
      "loss": 0.6903,
      "step": 278000
    },
    {
      "epoch": 103.49312523225566,
      "grad_norm": 2.223482131958008,
      "learning_rate": 1e-05,
      "loss": 0.6962,
      "step": 278500
    },
    {
      "epoch": 103.67892976588628,
      "grad_norm": 2.4267632961273193,
      "learning_rate": 1e-05,
      "loss": 0.6889,
      "step": 279000
    },
    {
      "epoch": 103.86473429951691,
      "grad_norm": 2.554532527923584,
      "learning_rate": 1e-05,
      "loss": 0.6902,
      "step": 279500
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.7725475165937876,
      "eval_f1_macro": 0.4578893476938316,
      "eval_f1_micro": 0.7725475165937876,
      "eval_loss": 0.6456966400146484,
      "eval_runtime": 510.0416,
      "eval_samples_per_second": 112.542,
      "eval_steps_per_second": 1.759,
      "learning_rate": 1e-05,
      "step": 279864
    },
    {
      "epoch": 104.05053883314753,
      "grad_norm": 2.5456807613372803,
      "learning_rate": 1e-05,
      "loss": 0.6956,
      "step": 280000
    },
    {
      "epoch": 104.23634336677814,
      "grad_norm": 2.9953627586364746,
      "learning_rate": 1e-05,
      "loss": 0.7025,
      "step": 280500
    },
    {
      "epoch": 104.42214790040877,
      "grad_norm": 2.574535369873047,
      "learning_rate": 1e-05,
      "loss": 0.6825,
      "step": 281000
    },
    {
      "epoch": 104.60795243403939,
      "grad_norm": 3.287419557571411,
      "learning_rate": 1e-05,
      "loss": 0.6953,
      "step": 281500
    },
    {
      "epoch": 104.79375696767,
      "grad_norm": 2.3188724517822266,
      "learning_rate": 1e-05,
      "loss": 0.6979,
      "step": 282000
    },
    {
      "epoch": 104.97956150130064,
      "grad_norm": 2.4571897983551025,
      "learning_rate": 1e-05,
      "loss": 0.6876,
      "step": 282500
    },
    {
      "epoch": 105.0,
      "eval_accuracy": 0.7719377711189701,
      "eval_f1_macro": 0.45556962818046953,
      "eval_f1_micro": 0.7719377711189701,
      "eval_loss": 0.6455578804016113,
      "eval_runtime": 507.9867,
      "eval_samples_per_second": 112.997,
      "eval_steps_per_second": 1.766,
      "learning_rate": 1e-05,
      "step": 282555
    },
    {
      "epoch": 105.16536603493125,
      "grad_norm": 2.7945613861083984,
      "learning_rate": 1e-05,
      "loss": 0.6903,
      "step": 283000
    },
    {
      "epoch": 105.35117056856187,
      "grad_norm": 2.2913684844970703,
      "learning_rate": 1e-05,
      "loss": 0.7094,
      "step": 283500
    },
    {
      "epoch": 105.5369751021925,
      "grad_norm": 2.745809555053711,
      "learning_rate": 1e-05,
      "loss": 0.685,
      "step": 284000
    },
    {
      "epoch": 105.72277963582312,
      "grad_norm": 2.8689208030700684,
      "learning_rate": 1e-05,
      "loss": 0.6873,
      "step": 284500
    },
    {
      "epoch": 105.90858416945373,
      "grad_norm": 2.0228753089904785,
      "learning_rate": 1e-05,
      "loss": 0.6849,
      "step": 285000
    },
    {
      "epoch": 106.0,
      "eval_accuracy": 0.7723384610024215,
      "eval_f1_macro": 0.4644160307431161,
      "eval_f1_micro": 0.7723384610024215,
      "eval_loss": 0.6444206237792969,
      "eval_runtime": 515.3095,
      "eval_samples_per_second": 111.391,
      "eval_steps_per_second": 1.741,
      "learning_rate": 1e-05,
      "step": 285246
    },
    {
      "epoch": 106.09438870308435,
      "grad_norm": 0.45027607679367065,
      "learning_rate": 1e-05,
      "loss": 0.6656,
      "step": 285500
    },
    {
      "epoch": 106.28019323671498,
      "grad_norm": 0.3893894553184509,
      "learning_rate": 1e-05,
      "loss": 0.5644,
      "step": 286000
    },
    {
      "epoch": 106.4659977703456,
      "grad_norm": 0.3509347438812256,
      "learning_rate": 1e-05,
      "loss": 0.4686,
      "step": 286500
    },
    {
      "epoch": 106.65180230397621,
      "grad_norm": 0.32040271162986755,
      "learning_rate": 1e-05,
      "loss": 0.3969,
      "step": 287000
    },
    {
      "epoch": 106.83760683760684,
      "grad_norm": 0.4328668415546417,
      "learning_rate": 1e-05,
      "loss": 0.3411,
      "step": 287500
    },
    {
      "epoch": 107.0,
      "eval_accuracy": 0.04797825821849794,
      "eval_f1_macro": 0.360721302464235,
      "eval_f1_micro": 0.4910938804941607,
      "eval_loss": 0.26552170515060425,
      "eval_runtime": 519.2635,
      "eval_samples_per_second": 110.543,
      "eval_steps_per_second": 1.727,
      "learning_rate": 1e-05,
      "step": 287937
    },
    {
      "epoch": 107.02341137123746,
      "grad_norm": 0.26703259348869324,
      "learning_rate": 1e-05,
      "loss": 0.2913,
      "step": 288000
    },
    {
      "epoch": 107.20921590486807,
      "grad_norm": 0.2284982055425644,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 288500
    },
    {
      "epoch": 107.3950204384987,
      "grad_norm": 1.4908051490783691,
      "learning_rate": 1e-05,
      "loss": 0.2169,
      "step": 289000
    },
    {
      "epoch": 107.58082497212932,
      "grad_norm": 0.18897071480751038,
      "learning_rate": 1e-05,
      "loss": 0.1873,
      "step": 289500
    },
    {
      "epoch": 107.76662950575994,
      "grad_norm": 0.17562171816825867,
      "learning_rate": 1e-05,
      "loss": 0.1619,
      "step": 290000
    },
    {
      "epoch": 107.95243403939057,
      "grad_norm": 0.19028235971927643,
      "learning_rate": 1e-05,
      "loss": 0.141,
      "step": 290500
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.44983536872179924,
      "eval_f1_macro": 0.22462065038139475,
      "eval_f1_micro": 0.6693680656054029,
      "eval_loss": 0.1419014185667038,
      "eval_runtime": 516.6186,
      "eval_samples_per_second": 111.109,
      "eval_steps_per_second": 1.736,
      "learning_rate": 1e-05,
      "step": 290628
    },
    {
      "epoch": 108.13823857302118,
      "grad_norm": 0.1382223218679428,
      "learning_rate": 1e-05,
      "loss": 0.1234,
      "step": 291000
    },
    {
      "epoch": 108.3240431066518,
      "grad_norm": 0.2637607753276825,
      "learning_rate": 1e-05,
      "loss": 0.109,
      "step": 291500
    },
    {
      "epoch": 108.50984764028242,
      "grad_norm": 0.16210788488388062,
      "learning_rate": 1e-05,
      "loss": 0.0977,
      "step": 292000
    },
    {
      "epoch": 108.69565217391305,
      "grad_norm": 0.1515241414308548,
      "learning_rate": 1e-05,
      "loss": 0.088,
      "step": 292500
    },
    {
      "epoch": 108.88145670754366,
      "grad_norm": 0.15725761651992798,
      "learning_rate": 1e-05,
      "loss": 0.0809,
      "step": 293000
    },
    {
      "epoch": 109.0,
      "eval_accuracy": 0.6714691381683245,
      "eval_f1_macro": 0.2136927959803109,
      "eval_f1_micro": 0.7449736568518617,
      "eval_loss": 0.07755623757839203,
      "eval_runtime": 503.5448,
      "eval_samples_per_second": 113.994,
      "eval_steps_per_second": 1.781,
      "learning_rate": 1e-05,
      "step": 293319
    },
    {
      "epoch": 109.06726124117428,
      "grad_norm": 0.13999715447425842,
      "learning_rate": 1e-05,
      "loss": 0.076,
      "step": 293500
    },
    {
      "epoch": 109.25306577480491,
      "grad_norm": 0.12521325051784515,
      "learning_rate": 1e-05,
      "loss": 0.0714,
      "step": 294000
    },
    {
      "epoch": 109.43887030843553,
      "grad_norm": 0.16435325145721436,
      "learning_rate": 1e-05,
      "loss": 0.0688,
      "step": 294500
    },
    {
      "epoch": 109.62467484206614,
      "grad_norm": 0.1280326545238495,
      "learning_rate": 1e-05,
      "loss": 0.066,
      "step": 295000
    },
    {
      "epoch": 109.81047937569677,
      "grad_norm": 0.1201464980840683,
      "learning_rate": 1e-05,
      "loss": 0.0643,
      "step": 295500
    },
    {
      "epoch": 109.99628390932739,
      "grad_norm": 0.09229467064142227,
      "learning_rate": 1e-05,
      "loss": 0.0621,
      "step": 296000
    },
    {
      "epoch": 110.0,
      "eval_accuracy": 0.6837163115625163,
      "eval_f1_macro": 0.26414762709864326,
      "eval_f1_micro": 0.7489470111853911,
      "eval_loss": 0.05802077427506447,
      "eval_runtime": 515.3576,
      "eval_samples_per_second": 111.381,
      "eval_steps_per_second": 1.741,
      "learning_rate": 1e-05,
      "step": 296010
    },
    {
      "epoch": 110.182088442958,
      "grad_norm": 0.12133761495351791,
      "learning_rate": 1e-05,
      "loss": 0.0613,
      "step": 296500
    },
    {
      "epoch": 110.36789297658864,
      "grad_norm": 0.11678178608417511,
      "learning_rate": 1e-05,
      "loss": 0.0601,
      "step": 297000
    },
    {
      "epoch": 110.55369751021925,
      "grad_norm": 0.11925112456083298,
      "learning_rate": 1e-05,
      "loss": 0.0596,
      "step": 297500
    },
    {
      "epoch": 110.73950204384987,
      "grad_norm": 0.16815921664237976,
      "learning_rate": 1e-05,
      "loss": 0.0589,
      "step": 298000
    },
    {
      "epoch": 110.92530657748048,
      "grad_norm": 0.11897558718919754,
      "learning_rate": 1e-05,
      "loss": 0.0582,
      "step": 298500
    },
    {
      "epoch": 111.0,
      "eval_accuracy": 0.6935245030574381,
      "eval_f1_macro": 0.331603552491686,
      "eval_f1_micro": 0.7547299175391458,
      "eval_loss": 0.053473543375730515,
      "eval_runtime": 517.7576,
      "eval_samples_per_second": 110.865,
      "eval_steps_per_second": 1.732,
      "learning_rate": 1e-05,
      "step": 298701
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 0.20368880033493042,
      "learning_rate": 1e-05,
      "loss": 0.0588,
      "step": 299000
    },
    {
      "epoch": 111.29691564474173,
      "grad_norm": 0.1355835497379303,
      "learning_rate": 1e-05,
      "loss": 0.0577,
      "step": 299500
    },
    {
      "epoch": 111.48272017837235,
      "grad_norm": 0.15652874112129211,
      "learning_rate": 1e-05,
      "loss": 0.0569,
      "step": 300000
    },
    {
      "epoch": 111.66852471200298,
      "grad_norm": 0.1435033082962036,
      "learning_rate": 1e-05,
      "loss": 0.0575,
      "step": 300500
    },
    {
      "epoch": 111.8543292456336,
      "grad_norm": 0.13060764968395233,
      "learning_rate": 1e-05,
      "loss": 0.0568,
      "step": 301000
    },
    {
      "epoch": 112.0,
      "eval_accuracy": 0.6998135920976987,
      "eval_f1_macro": 0.3483537603081725,
      "eval_f1_micro": 0.7585280588776449,
      "eval_loss": 0.05167479068040848,
      "eval_runtime": 513.0638,
      "eval_samples_per_second": 111.879,
      "eval_steps_per_second": 1.748,
      "learning_rate": 1e-05,
      "step": 301392
    },
    {
      "epoch": 112.04013377926421,
      "grad_norm": 0.12529785931110382,
      "learning_rate": 1e-05,
      "loss": 0.0571,
      "step": 301500
    },
    {
      "epoch": 112.22593831289484,
      "grad_norm": 0.16798335313796997,
      "learning_rate": 1e-05,
      "loss": 0.0568,
      "step": 302000
    },
    {
      "epoch": 112.41174284652546,
      "grad_norm": 0.1425975114107132,
      "learning_rate": 1e-05,
      "loss": 0.0557,
      "step": 302500
    },
    {
      "epoch": 112.59754738015607,
      "grad_norm": 0.14742553234100342,
      "learning_rate": 1e-05,
      "loss": 0.0553,
      "step": 303000
    },
    {
      "epoch": 112.7833519137867,
      "grad_norm": 0.1510126292705536,
      "learning_rate": 1e-05,
      "loss": 0.056,
      "step": 303500
    },
    {
      "epoch": 112.96915644741732,
      "grad_norm": 0.14474214613437653,
      "learning_rate": 1e-05,
      "loss": 0.0557,
      "step": 304000
    },
    {
      "epoch": 113.0,
      "eval_accuracy": 0.7042734447135067,
      "eval_f1_macro": 0.3378893620669972,
      "eval_f1_micro": 0.7611001027447527,
      "eval_loss": 0.05106380954384804,
      "eval_runtime": 521.661,
      "eval_samples_per_second": 110.035,
      "eval_steps_per_second": 1.72,
      "learning_rate": 1e-05,
      "step": 304083
    },
    {
      "epoch": 113.15496098104794,
      "grad_norm": 0.12822121381759644,
      "learning_rate": 1e-05,
      "loss": 0.056,
      "step": 304500
    },
    {
      "epoch": 113.34076551467855,
      "grad_norm": 0.14637312293052673,
      "learning_rate": 1e-05,
      "loss": 0.0556,
      "step": 305000
    },
    {
      "epoch": 113.52657004830918,
      "grad_norm": 0.1259402185678482,
      "learning_rate": 1e-05,
      "loss": 0.0554,
      "step": 305500
    },
    {
      "epoch": 113.7123745819398,
      "grad_norm": 0.13158832490444183,
      "learning_rate": 1e-05,
      "loss": 0.0556,
      "step": 306000
    },
    {
      "epoch": 113.89817911557041,
      "grad_norm": 0.12858903408050537,
      "learning_rate": 1e-05,
      "loss": 0.0552,
      "step": 306500
    },
    {
      "epoch": 114.0,
      "eval_accuracy": 0.7053535652688978,
      "eval_f1_macro": 0.35703913789456687,
      "eval_f1_micro": 0.7622047244094489,
      "eval_loss": 0.05065497010946274,
      "eval_runtime": 534.7718,
      "eval_samples_per_second": 107.337,
      "eval_steps_per_second": 1.677,
      "learning_rate": 1e-05,
      "step": 306774
    },
    {
      "epoch": 114.08398364920104,
      "grad_norm": 0.16015706956386566,
      "learning_rate": 1e-05,
      "loss": 0.0548,
      "step": 307000
    },
    {
      "epoch": 114.26978818283166,
      "grad_norm": 0.13509486615657806,
      "learning_rate": 1e-05,
      "loss": 0.0551,
      "step": 307500
    },
    {
      "epoch": 114.45559271646228,
      "grad_norm": 0.14582909643650055,
      "learning_rate": 1e-05,
      "loss": 0.0547,
      "step": 308000
    },
    {
      "epoch": 114.64139725009291,
      "grad_norm": 0.14752507209777832,
      "learning_rate": 1e-05,
      "loss": 0.0552,
      "step": 308500
    },
    {
      "epoch": 114.82720178372352,
      "grad_norm": 0.1321249157190323,
      "learning_rate": 1e-05,
      "loss": 0.0555,
      "step": 309000
    },
    {
      "epoch": 115.0,
      "eval_accuracy": 0.7117820247034023,
      "eval_f1_macro": 0.36428903036337407,
      "eval_f1_micro": 0.7647240545893983,
      "eval_loss": 0.05039990693330765,
      "eval_runtime": 609.7033,
      "eval_samples_per_second": 94.146,
      "eval_steps_per_second": 1.471,
      "learning_rate": 1e-05,
      "step": 309465
    },
    {
      "epoch": 115.01300631735414,
      "grad_norm": 0.1596326380968094,
      "learning_rate": 1e-05,
      "loss": 0.0546,
      "step": 309500
    },
    {
      "epoch": 115.19881085098477,
      "grad_norm": 0.19358591735363007,
      "learning_rate": 1e-05,
      "loss": 0.0547,
      "step": 310000
    },
    {
      "epoch": 115.38461538461539,
      "grad_norm": 0.1181555911898613,
      "learning_rate": 1e-05,
      "loss": 0.0551,
      "step": 310500
    },
    {
      "epoch": 115.570419918246,
      "grad_norm": 0.16747964918613434,
      "learning_rate": 1e-05,
      "loss": 0.0551,
      "step": 311000
    },
    {
      "epoch": 115.75622445187662,
      "grad_norm": 0.1441546380519867,
      "learning_rate": 1e-05,
      "loss": 0.0544,
      "step": 311500
    },
    {
      "epoch": 115.94202898550725,
      "grad_norm": 0.1473866105079651,
      "learning_rate": 1e-05,
      "loss": 0.0546,
      "step": 312000
    },
    {
      "epoch": 116.0,
      "eval_accuracy": 0.7101966864688769,
      "eval_f1_macro": 0.3622993891059384,
      "eval_f1_micro": 0.7647289615591668,
      "eval_loss": 0.05015714839100838,
      "eval_runtime": 523.6323,
      "eval_samples_per_second": 109.621,
      "eval_steps_per_second": 1.713,
      "learning_rate": 1e-05,
      "step": 312156
    },
    {
      "epoch": 116.12783351913787,
      "grad_norm": 0.18043090403079987,
      "learning_rate": 1e-05,
      "loss": 0.0548,
      "step": 312500
    },
    {
      "epoch": 116.31363805276848,
      "grad_norm": 0.1623302400112152,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 313000
    },
    {
      "epoch": 116.49944258639911,
      "grad_norm": 0.1569686233997345,
      "learning_rate": 1e-05,
      "loss": 0.055,
      "step": 313500
    },
    {
      "epoch": 116.68524712002973,
      "grad_norm": 0.17043337225914001,
      "learning_rate": 1e-05,
      "loss": 0.0541,
      "step": 314000
    },
    {
      "epoch": 116.87105165366034,
      "grad_norm": 0.15649768710136414,
      "learning_rate": 1e-05,
      "loss": 0.0545,
      "step": 314500
    },
    {
      "epoch": 117.0,
      "eval_accuracy": 0.712914409156635,
      "eval_f1_macro": 0.36544151863175506,
      "eval_f1_micro": 0.7657223847509677,
      "eval_loss": 0.050175271928310394,
      "eval_runtime": 598.1573,
      "eval_samples_per_second": 95.963,
      "eval_steps_per_second": 1.5,
      "learning_rate": 1e-05,
      "step": 314847
    },
    {
      "epoch": 117.05685618729098,
      "grad_norm": 0.1493714600801468,
      "learning_rate": 1e-05,
      "loss": 0.0549,
      "step": 315000
    },
    {
      "epoch": 117.24266072092159,
      "grad_norm": 0.1313825249671936,
      "learning_rate": 1e-05,
      "loss": 0.0547,
      "step": 315500
    },
    {
      "epoch": 117.42846525455221,
      "grad_norm": 0.13993392884731293,
      "learning_rate": 1e-05,
      "loss": 0.0541,
      "step": 316000
    },
    {
      "epoch": 117.61426978818284,
      "grad_norm": 0.13758248090744019,
      "learning_rate": 1e-05,
      "loss": 0.0546,
      "step": 316500
    },
    {
      "epoch": 117.80007432181345,
      "grad_norm": 0.16916200518608093,
      "learning_rate": 1e-05,
      "loss": 0.0545,
      "step": 317000
    },
    {
      "epoch": 117.98587885544407,
      "grad_norm": 0.1452600359916687,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 317500
    },
    {
      "epoch": 118.0,
      "eval_accuracy": 0.7141687427048309,
      "eval_f1_macro": 0.3524192831401073,
      "eval_f1_micro": 0.7654782537680462,
      "eval_loss": 0.050468478351831436,
      "eval_runtime": 549.4659,
      "eval_samples_per_second": 104.467,
      "eval_steps_per_second": 1.632,
      "learning_rate": 1e-05,
      "step": 317538
    },
    {
      "epoch": 118.17168338907469,
      "grad_norm": 0.12589971721172333,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 318000
    },
    {
      "epoch": 118.35748792270532,
      "grad_norm": 0.17545440793037415,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 318500
    },
    {
      "epoch": 118.54329245633593,
      "grad_norm": 0.15301626920700073,
      "learning_rate": 1e-05,
      "loss": 0.0541,
      "step": 319000
    },
    {
      "epoch": 118.72909698996655,
      "grad_norm": 0.1384187638759613,
      "learning_rate": 1e-05,
      "loss": 0.0545,
      "step": 319500
    },
    {
      "epoch": 118.91490152359718,
      "grad_norm": 0.16759108006954193,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 320000
    },
    {
      "epoch": 119.0,
      "eval_accuracy": 0.7127053535652689,
      "eval_f1_macro": 0.34416444697858145,
      "eval_f1_micro": 0.7658673932788375,
      "eval_loss": 0.049900032579898834,
      "eval_runtime": 491.987,
      "eval_samples_per_second": 116.672,
      "eval_steps_per_second": 1.823,
      "learning_rate": 1e-05,
      "step": 320229
    },
    {
      "epoch": 119.1007060572278,
      "grad_norm": 0.15850204229354858,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 320500
    },
    {
      "epoch": 119.28651059085841,
      "grad_norm": 0.18434040248394012,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 321000
    },
    {
      "epoch": 119.47231512448904,
      "grad_norm": 0.15472249686717987,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 321500
    },
    {
      "epoch": 119.65811965811966,
      "grad_norm": 0.1760583370923996,
      "learning_rate": 1e-05,
      "loss": 0.0544,
      "step": 322000
    },
    {
      "epoch": 119.84392419175028,
      "grad_norm": 0.1788097620010376,
      "learning_rate": 1e-05,
      "loss": 0.0541,
      "step": 322500
    },
    {
      "epoch": 120.0,
      "eval_accuracy": 0.7130886221494399,
      "eval_f1_macro": 0.35077501544817247,
      "eval_f1_micro": 0.7657258505633957,
      "eval_loss": 0.049903545528650284,
      "eval_runtime": 513.1716,
      "eval_samples_per_second": 111.855,
      "eval_steps_per_second": 1.748,
      "learning_rate": 1e-05,
      "step": 322920
    },
    {
      "epoch": 120.0297287253809,
      "grad_norm": 0.1507652848958969,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 323000
    },
    {
      "epoch": 120.21553325901152,
      "grad_norm": 0.1735050082206726,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 323500
    },
    {
      "epoch": 120.40133779264214,
      "grad_norm": 0.14820708334445953,
      "learning_rate": 1e-05,
      "loss": 0.0543,
      "step": 324000
    },
    {
      "epoch": 120.58714232627275,
      "grad_norm": 0.14484427869319916,
      "learning_rate": 1e-05,
      "loss": 0.0537,
      "step": 324500
    },
    {
      "epoch": 120.77294685990339,
      "grad_norm": 0.15814656019210815,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 325000
    },
    {
      "epoch": 120.958751393534,
      "grad_norm": 0.19824689626693726,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 325500
    },
    {
      "epoch": 121.0,
      "eval_accuracy": 0.7140816362084285,
      "eval_f1_macro": 0.3627670615797559,
      "eval_f1_micro": 0.7665756914119359,
      "eval_loss": 0.04957958310842514,
      "eval_runtime": 509.187,
      "eval_samples_per_second": 112.731,
      "eval_steps_per_second": 1.762,
      "learning_rate": 1e-05,
      "step": 325611
    },
    {
      "epoch": 121.14455592716462,
      "grad_norm": 0.16547606885433197,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 326000
    },
    {
      "epoch": 121.33036046079525,
      "grad_norm": 0.2055787444114685,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 326500
    },
    {
      "epoch": 121.51616499442586,
      "grad_norm": 0.2001011222600937,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 327000
    },
    {
      "epoch": 121.70196952805648,
      "grad_norm": 0.17856118083000183,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 327500
    },
    {
      "epoch": 121.88777406168711,
      "grad_norm": 0.14508357644081116,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 328000
    },
    {
      "epoch": 122.0,
      "eval_accuracy": 0.7163812477134545,
      "eval_f1_macro": 0.35293584502475456,
      "eval_f1_micro": 0.7672263726699065,
      "eval_loss": 0.04973344877362251,
      "eval_runtime": 495.422,
      "eval_samples_per_second": 115.863,
      "eval_steps_per_second": 1.811,
      "learning_rate": 1e-05,
      "step": 328302
    },
    {
      "epoch": 122.07357859531773,
      "grad_norm": 0.1540980488061905,
      "learning_rate": 1e-05,
      "loss": 0.0543,
      "step": 328500
    },
    {
      "epoch": 122.25938312894834,
      "grad_norm": 0.18639309704303741,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 329000
    },
    {
      "epoch": 122.44518766257897,
      "grad_norm": 0.15830326080322266,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 329500
    },
    {
      "epoch": 122.63099219620959,
      "grad_norm": 0.17811599373817444,
      "learning_rate": 1e-05,
      "loss": 0.0537,
      "step": 330000
    },
    {
      "epoch": 122.8167967298402,
      "grad_norm": 0.1285717487335205,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 330500
    },
    {
      "epoch": 123.0,
      "eval_accuracy": 0.7153533910559049,
      "eval_f1_macro": 0.36741171960145996,
      "eval_f1_micro": 0.7661930650098223,
      "eval_loss": 0.04949206858873367,
      "eval_runtime": 496.6111,
      "eval_samples_per_second": 115.585,
      "eval_steps_per_second": 1.806,
      "learning_rate": 1e-05,
      "step": 330993
    },
    {
      "epoch": 123.00260126347082,
      "grad_norm": 0.17617733776569366,
      "learning_rate": 1e-05,
      "loss": 0.0537,
      "step": 331000
    },
    {
      "epoch": 123.18840579710145,
      "grad_norm": 0.20897239446640015,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 331500
    },
    {
      "epoch": 123.37421033073207,
      "grad_norm": 0.1522960364818573,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 332000
    },
    {
      "epoch": 123.56001486436269,
      "grad_norm": 0.18495243787765503,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 332500
    },
    {
      "epoch": 123.74581939799332,
      "grad_norm": 0.15397348999977112,
      "learning_rate": 1e-05,
      "loss": 0.0541,
      "step": 333000
    },
    {
      "epoch": 123.93162393162393,
      "grad_norm": 0.16448819637298584,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 333500
    },
    {
      "epoch": 124.0,
      "eval_accuracy": 0.7160676643264055,
      "eval_f1_macro": 0.36413807211107735,
      "eval_f1_micro": 0.7673595994775795,
      "eval_loss": 0.049613192677497864,
      "eval_runtime": 491.9647,
      "eval_samples_per_second": 116.677,
      "eval_steps_per_second": 1.823,
      "learning_rate": 1e-05,
      "step": 333684
    },
    {
      "epoch": 124.11742846525455,
      "grad_norm": 0.14923396706581116,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 334000
    },
    {
      "epoch": 124.30323299888518,
      "grad_norm": 0.1714375615119934,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 334500
    },
    {
      "epoch": 124.4890375325158,
      "grad_norm": 0.17104235291481018,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 335000
    },
    {
      "epoch": 124.67484206614641,
      "grad_norm": 0.16875581443309784,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 335500
    },
    {
      "epoch": 124.86064659977704,
      "grad_norm": 0.14356301724910736,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 336000
    },
    {
      "epoch": 125.0,
      "eval_accuracy": 0.7124440340760614,
      "eval_f1_macro": 0.3509397162428964,
      "eval_f1_micro": 0.7658070643240676,
      "eval_loss": 0.04959910735487938,
      "eval_runtime": 487.4427,
      "eval_samples_per_second": 117.759,
      "eval_steps_per_second": 1.84,
      "learning_rate": 1e-05,
      "step": 336375
    },
    {
      "epoch": 125.04645113340766,
      "grad_norm": 0.15746818482875824,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 336500
    },
    {
      "epoch": 125.23225566703827,
      "grad_norm": 0.1541147232055664,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 337000
    },
    {
      "epoch": 125.41806020066889,
      "grad_norm": 0.15555234253406525,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 337500
    },
    {
      "epoch": 125.60386473429952,
      "grad_norm": 0.17335093021392822,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 338000
    },
    {
      "epoch": 125.78966926793014,
      "grad_norm": 0.18515528738498688,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 338500
    },
    {
      "epoch": 125.97547380156075,
      "grad_norm": 0.17530472576618195,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 339000
    },
    {
      "epoch": 126.0,
      "eval_accuracy": 0.7152662845595025,
      "eval_f1_macro": 0.37424111866414195,
      "eval_f1_micro": 0.7660751240774316,
      "eval_loss": 0.0494619682431221,
      "eval_runtime": 502.6945,
      "eval_samples_per_second": 114.187,
      "eval_steps_per_second": 1.784,
      "learning_rate": 1e-05,
      "step": 339066
    },
    {
      "epoch": 126.16127833519138,
      "grad_norm": 0.1517336517572403,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 339500
    },
    {
      "epoch": 126.347082868822,
      "grad_norm": 0.1815425604581833,
      "learning_rate": 1e-05,
      "loss": 0.0542,
      "step": 340000
    },
    {
      "epoch": 126.53288740245262,
      "grad_norm": 0.17021115124225616,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 340500
    },
    {
      "epoch": 126.71869193608325,
      "grad_norm": 0.19511722028255463,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 341000
    },
    {
      "epoch": 126.90449646971386,
      "grad_norm": 0.22142985463142395,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 341500
    },
    {
      "epoch": 127.0,
      "eval_accuracy": 0.7149178585738925,
      "eval_f1_macro": 0.35735030768195364,
      "eval_f1_micro": 0.766314294299216,
      "eval_loss": 0.049399666488170624,
      "eval_runtime": 486.665,
      "eval_samples_per_second": 117.948,
      "eval_steps_per_second": 1.843,
      "learning_rate": 1e-05,
      "step": 341757
    },
    {
      "epoch": 127.09030100334448,
      "grad_norm": 0.1855282485485077,
      "learning_rate": 1e-05,
      "loss": 0.0527,
      "step": 342000
    },
    {
      "epoch": 127.27610553697511,
      "grad_norm": 0.14693668484687805,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 342500
    },
    {
      "epoch": 127.46191007060573,
      "grad_norm": 0.19144034385681152,
      "learning_rate": 1e-05,
      "loss": 0.0537,
      "step": 343000
    },
    {
      "epoch": 127.64771460423634,
      "grad_norm": 0.16127558052539825,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 343500
    },
    {
      "epoch": 127.83351913786696,
      "grad_norm": 0.20746751129627228,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 344000
    },
    {
      "epoch": 128.0,
      "eval_accuracy": 0.714412640894758,
      "eval_f1_macro": 0.36010176970077795,
      "eval_f1_micro": 0.7664776721721585,
      "eval_loss": 0.04938925430178642,
      "eval_runtime": 486.3797,
      "eval_samples_per_second": 118.017,
      "eval_steps_per_second": 1.844,
      "learning_rate": 1e-05,
      "step": 344448
    },
    {
      "epoch": 128.01932367149757,
      "grad_norm": 0.13175125420093536,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 344500
    },
    {
      "epoch": 128.2051282051282,
      "grad_norm": 0.1861082762479782,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 345000
    },
    {
      "epoch": 128.39093273875883,
      "grad_norm": 0.16761593520641327,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 345500
    },
    {
      "epoch": 128.57673727238944,
      "grad_norm": 0.18102224171161652,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 346000
    },
    {
      "epoch": 128.76254180602007,
      "grad_norm": 0.166373148560524,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 346500
    },
    {
      "epoch": 128.9483463396507,
      "grad_norm": 0.2147304117679596,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 347000
    },
    {
      "epoch": 129.0,
      "eval_accuracy": 0.717931743349419,
      "eval_f1_macro": 0.3641550142658243,
      "eval_f1_micro": 0.7674323253122921,
      "eval_loss": 0.049368634819984436,
      "eval_runtime": 497.4863,
      "eval_samples_per_second": 115.382,
      "eval_steps_per_second": 1.803,
      "learning_rate": 1e-05,
      "step": 347139
    },
    {
      "epoch": 129.1341508732813,
      "grad_norm": 0.15825386345386505,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 347500
    },
    {
      "epoch": 129.31995540691193,
      "grad_norm": 0.17780201137065887,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 348000
    },
    {
      "epoch": 129.50575994054256,
      "grad_norm": 0.1677912026643753,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 348500
    },
    {
      "epoch": 129.69156447417316,
      "grad_norm": 0.18808604776859283,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 349000
    },
    {
      "epoch": 129.8773690078038,
      "grad_norm": 0.1602342426776886,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 349500
    },
    {
      "epoch": 130.0,
      "eval_accuracy": 0.717687845159492,
      "eval_f1_macro": 0.3602711408206009,
      "eval_f1_micro": 0.7667705923765463,
      "eval_loss": 0.049409620463848114,
      "eval_runtime": 497.5454,
      "eval_samples_per_second": 115.368,
      "eval_steps_per_second": 1.803,
      "learning_rate": 1e-05,
      "step": 349830
    },
    {
      "epoch": 130.06317354143442,
      "grad_norm": 0.23750899732112885,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 350000
    },
    {
      "epoch": 130.24897807506503,
      "grad_norm": 0.16858229041099548,
      "learning_rate": 1e-05,
      "loss": 0.0529,
      "step": 350500
    },
    {
      "epoch": 130.43478260869566,
      "grad_norm": 0.19223752617835999,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 351000
    },
    {
      "epoch": 130.6205871423263,
      "grad_norm": 0.1930113583803177,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 351500
    },
    {
      "epoch": 130.8063916759569,
      "grad_norm": 0.20461726188659668,
      "learning_rate": 1e-05,
      "loss": 0.054,
      "step": 352000
    },
    {
      "epoch": 130.99219620958752,
      "grad_norm": 0.20236244797706604,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 352500
    },
    {
      "epoch": 131.0,
      "eval_accuracy": 0.718210484137907,
      "eval_f1_macro": 0.3664294602272974,
      "eval_f1_micro": 0.7665082507046622,
      "eval_loss": 0.04939533770084381,
      "eval_runtime": 493.8302,
      "eval_samples_per_second": 116.236,
      "eval_steps_per_second": 1.816,
      "learning_rate": 1e-05,
      "step": 352521
    },
    {
      "epoch": 131.17800074321812,
      "grad_norm": 0.16268064081668854,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 353000
    },
    {
      "epoch": 131.36380527684875,
      "grad_norm": 0.22975020110607147,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 353500
    },
    {
      "epoch": 131.54960981047938,
      "grad_norm": 0.16796068847179413,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 354000
    },
    {
      "epoch": 131.73541434410998,
      "grad_norm": 0.2104586660861969,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 354500
    },
    {
      "epoch": 131.9212188777406,
      "grad_norm": 0.17215226590633392,
      "learning_rate": 1e-05,
      "loss": 0.0528,
      "step": 355000
    },
    {
      "epoch": 132.0,
      "eval_accuracy": 0.717583317363809,
      "eval_f1_macro": 0.3651176446191739,
      "eval_f1_micro": 0.7664564319910658,
      "eval_loss": 0.04943186417222023,
      "eval_runtime": 487.4839,
      "eval_samples_per_second": 117.75,
      "eval_steps_per_second": 1.84,
      "learning_rate": 1e-05,
      "step": 355212
    },
    {
      "epoch": 132.10702341137124,
      "grad_norm": 0.1763976663351059,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 355500
    },
    {
      "epoch": 132.29282794500185,
      "grad_norm": 0.19165627658367157,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 356000
    },
    {
      "epoch": 132.47863247863248,
      "grad_norm": 0.18794529139995575,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 356500
    },
    {
      "epoch": 132.6644370122631,
      "grad_norm": 0.19783887267112732,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 357000
    },
    {
      "epoch": 132.8502415458937,
      "grad_norm": 0.14160636067390442,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 357500
    },
    {
      "epoch": 133.0,
      "eval_accuracy": 0.714621696486124,
      "eval_f1_macro": 0.36115748131858555,
      "eval_f1_micro": 0.7658437005098911,
      "eval_loss": 0.049288176000118256,
      "eval_runtime": 479.1036,
      "eval_samples_per_second": 119.809,
      "eval_steps_per_second": 1.872,
      "learning_rate": 1e-05,
      "step": 357903
    },
    {
      "epoch": 133.03604607952434,
      "grad_norm": 0.18090824782848358,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 358000
    },
    {
      "epoch": 133.22185061315497,
      "grad_norm": 0.19353412091732025,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 358500
    },
    {
      "epoch": 133.40765514678557,
      "grad_norm": 0.22670786082744598,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 359000
    },
    {
      "epoch": 133.5934596804162,
      "grad_norm": 0.18295426666736603,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 359500
    },
    {
      "epoch": 133.77926421404683,
      "grad_norm": 0.20159971714019775,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 360000
    },
    {
      "epoch": 133.96506874767744,
      "grad_norm": 0.21708709001541138,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 360500
    },
    {
      "epoch": 134.0,
      "eval_accuracy": 0.7154753401508684,
      "eval_f1_macro": 0.3677607284274943,
      "eval_f1_micro": 0.7659699195779215,
      "eval_loss": 0.04927274212241173,
      "eval_runtime": 492.3636,
      "eval_samples_per_second": 116.583,
      "eval_steps_per_second": 1.822,
      "learning_rate": 1e-05,
      "step": 360594
    },
    {
      "epoch": 134.15087328130807,
      "grad_norm": 0.1716473251581192,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 361000
    },
    {
      "epoch": 134.3366778149387,
      "grad_norm": 0.1569896936416626,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 361500
    },
    {
      "epoch": 134.5224823485693,
      "grad_norm": 0.18859770894050598,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 362000
    },
    {
      "epoch": 134.70828688219993,
      "grad_norm": 0.18753333389759064,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 362500
    },
    {
      "epoch": 134.89409141583056,
      "grad_norm": 0.19746656715869904,
      "learning_rate": 1e-05,
      "loss": 0.0528,
      "step": 363000
    },
    {
      "epoch": 135.0,
      "eval_accuracy": 0.7189944426055295,
      "eval_f1_macro": 0.37226237632641596,
      "eval_f1_micro": 0.7674080308866179,
      "eval_loss": 0.04929700121283531,
      "eval_runtime": 476.5646,
      "eval_samples_per_second": 120.447,
      "eval_steps_per_second": 1.882,
      "learning_rate": 1e-05,
      "step": 363285
    },
    {
      "epoch": 135.07989594946116,
      "grad_norm": 0.1843235194683075,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 363500
    },
    {
      "epoch": 135.2657004830918,
      "grad_norm": 0.17511311173439026,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 364000
    },
    {
      "epoch": 135.45150501672242,
      "grad_norm": 0.20711584389209747,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 364500
    },
    {
      "epoch": 135.63730955035302,
      "grad_norm": 0.19971944391727448,
      "learning_rate": 1e-05,
      "loss": 0.0529,
      "step": 365000
    },
    {
      "epoch": 135.82311408398365,
      "grad_norm": 0.2045559585094452,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 365500
    },
    {
      "epoch": 136.0,
      "eval_accuracy": 0.7151094928659779,
      "eval_f1_macro": 0.35969130867283144,
      "eval_f1_micro": 0.766711291239524,
      "eval_loss": 0.049187980592250824,
      "eval_runtime": 502.7189,
      "eval_samples_per_second": 114.181,
      "eval_steps_per_second": 1.784,
      "learning_rate": 1e-05,
      "step": 365976
    },
    {
      "epoch": 136.00891861761426,
      "grad_norm": 0.17318807542324066,
      "learning_rate": 1e-05,
      "loss": 0.0529,
      "step": 366000
    },
    {
      "epoch": 136.1947231512449,
      "grad_norm": 0.20163071155548096,
      "learning_rate": 1e-05,
      "loss": 0.0528,
      "step": 366500
    },
    {
      "epoch": 136.38052768487552,
      "grad_norm": 0.17554792761802673,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 367000
    },
    {
      "epoch": 136.56633221850612,
      "grad_norm": 0.20222993195056915,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 367500
    },
    {
      "epoch": 136.75213675213675,
      "grad_norm": 0.2069316953420639,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 368000
    },
    {
      "epoch": 136.93794128576738,
      "grad_norm": 0.17365820705890656,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 368500
    },
    {
      "epoch": 137.0,
      "eval_accuracy": 0.7157889235379175,
      "eval_f1_macro": 0.3631043131303743,
      "eval_f1_micro": 0.7664713487937058,
      "eval_loss": 0.0491538941860199,
      "eval_runtime": 518.3587,
      "eval_samples_per_second": 110.736,
      "eval_steps_per_second": 1.73,
      "learning_rate": 1e-05,
      "step": 368667
    },
    {
      "epoch": 137.12374581939798,
      "grad_norm": 0.19572311639785767,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 369000
    },
    {
      "epoch": 137.3095503530286,
      "grad_norm": 0.1570323258638382,
      "learning_rate": 1e-05,
      "loss": 0.0528,
      "step": 369500
    },
    {
      "epoch": 137.49535488665924,
      "grad_norm": 0.16947729885578156,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 370000
    },
    {
      "epoch": 137.68115942028984,
      "grad_norm": 0.16606110334396362,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 370500
    },
    {
      "epoch": 137.86696395392048,
      "grad_norm": 0.2260875552892685,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 371000
    },
    {
      "epoch": 138.0,
      "eval_accuracy": 0.7178446368530165,
      "eval_f1_macro": 0.3687935119842292,
      "eval_f1_micro": 0.7665450277813434,
      "eval_loss": 0.04930136725306511,
      "eval_runtime": 510.8639,
      "eval_samples_per_second": 112.361,
      "eval_steps_per_second": 1.756,
      "learning_rate": 1e-05,
      "step": 371358
    },
    {
      "epoch": 138.0527684875511,
      "grad_norm": 0.163988396525383,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 371500
    },
    {
      "epoch": 138.2385730211817,
      "grad_norm": 0.20549984276294708,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 372000
    },
    {
      "epoch": 138.42437755481234,
      "grad_norm": 0.18827009201049805,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 372500
    },
    {
      "epoch": 138.61018208844297,
      "grad_norm": 0.2782110571861267,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 373000
    },
    {
      "epoch": 138.79598662207357,
      "grad_norm": 0.1959678679704666,
      "learning_rate": 1e-05,
      "loss": 0.0538,
      "step": 373500
    },
    {
      "epoch": 138.9817911557042,
      "grad_norm": 0.18641294538974762,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 374000
    },
    {
      "epoch": 139.0,
      "eval_accuracy": 0.7182279054371875,
      "eval_f1_macro": 0.35626916649899915,
      "eval_f1_micro": 0.766155421092079,
      "eval_loss": 0.04927237331867218,
      "eval_runtime": 508.9115,
      "eval_samples_per_second": 112.792,
      "eval_steps_per_second": 1.763,
      "learning_rate": 1e-05,
      "step": 374049
    },
    {
      "epoch": 139.16759568933483,
      "grad_norm": 0.16375960409641266,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 374500
    },
    {
      "epoch": 139.35340022296543,
      "grad_norm": 0.19818973541259766,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 375000
    },
    {
      "epoch": 139.53920475659606,
      "grad_norm": 0.19776101410388947,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 375500
    },
    {
      "epoch": 139.7250092902267,
      "grad_norm": 0.24181506037712097,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 376000
    },
    {
      "epoch": 139.9108138238573,
      "grad_norm": 0.21174757182598114,
      "learning_rate": 1e-05,
      "loss": 0.0528,
      "step": 376500
    },
    {
      "epoch": 140.0,
      "eval_accuracy": 0.7197958223724326,
      "eval_f1_macro": 0.3699733355385332,
      "eval_f1_micro": 0.767376184687937,
      "eval_loss": 0.04918988421559334,
      "eval_runtime": 505.3732,
      "eval_samples_per_second": 113.581,
      "eval_steps_per_second": 1.775,
      "learning_rate": 1e-05,
      "step": 376740
    },
    {
      "epoch": 140.09661835748793,
      "grad_norm": 0.18043100833892822,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 377000
    },
    {
      "epoch": 140.28242289111856,
      "grad_norm": 0.18150250613689423,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 377500
    },
    {
      "epoch": 140.46822742474916,
      "grad_norm": 0.17685070633888245,
      "learning_rate": 1e-05,
      "loss": 0.0523,
      "step": 378000
    },
    {
      "epoch": 140.6540319583798,
      "grad_norm": 0.2478715479373932,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 378500
    },
    {
      "epoch": 140.8398364920104,
      "grad_norm": 0.20271550118923187,
      "learning_rate": 1e-05,
      "loss": 0.0529,
      "step": 379000
    },
    {
      "epoch": 141.0,
      "eval_accuracy": 0.716468354209857,
      "eval_f1_macro": 0.35072596287625124,
      "eval_f1_micro": 0.7666041104041745,
      "eval_loss": 0.04920462518930435,
      "eval_runtime": 510.1227,
      "eval_samples_per_second": 112.524,
      "eval_steps_per_second": 1.758,
      "learning_rate": 1e-05,
      "step": 379431
    },
    {
      "epoch": 141.02564102564102,
      "grad_norm": 0.25310391187667847,
      "learning_rate": 1e-05,
      "loss": 0.0539,
      "step": 379500
    },
    {
      "epoch": 141.21144555927165,
      "grad_norm": 0.2286742925643921,
      "learning_rate": 1e-05,
      "loss": 0.0532,
      "step": 380000
    },
    {
      "epoch": 141.39725009290225,
      "grad_norm": 0.2335425764322281,
      "learning_rate": 1e-05,
      "loss": 0.0536,
      "step": 380500
    },
    {
      "epoch": 141.58305462653288,
      "grad_norm": 0.21884822845458984,
      "learning_rate": 1e-05,
      "loss": 0.0531,
      "step": 381000
    },
    {
      "epoch": 141.76885916016352,
      "grad_norm": 0.23641781508922577,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 381500
    },
    {
      "epoch": 141.95466369379412,
      "grad_norm": 0.19402460753917694,
      "learning_rate": 1e-05,
      "loss": 0.0529,
      "step": 382000
    },
    {
      "epoch": 142.0,
      "eval_accuracy": 0.7194473963868225,
      "eval_f1_macro": 0.36600085102264546,
      "eval_f1_micro": 0.7669340748803981,
      "eval_loss": 0.04919710010290146,
      "eval_runtime": 506.4483,
      "eval_samples_per_second": 113.34,
      "eval_steps_per_second": 1.771,
      "learning_rate": 1e-05,
      "step": 382122
    },
    {
      "epoch": 142.14046822742475,
      "grad_norm": 0.21041558682918549,
      "learning_rate": 1e-05,
      "loss": 0.0529,
      "step": 382500
    },
    {
      "epoch": 142.32627276105538,
      "grad_norm": 0.21750488877296448,
      "learning_rate": 1e-05,
      "loss": 0.0533,
      "step": 383000
    },
    {
      "epoch": 142.51207729468598,
      "grad_norm": 0.2097017467021942,
      "learning_rate": 1e-05,
      "loss": 0.053,
      "step": 383500
    },
    {
      "epoch": 142.6978818283166,
      "grad_norm": 0.2606968879699707,
      "learning_rate": 1e-05,
      "loss": 0.0535,
      "step": 384000
    },
    {
      "epoch": 142.88368636194724,
      "grad_norm": 0.18851463496685028,
      "learning_rate": 1e-05,
      "loss": 0.0534,
      "step": 384500
    },
    {
      "epoch": 143.0,
      "eval_accuracy": 0.7168342014947475,
      "eval_f1_macro": 0.3673237139632794,
      "eval_f1_micro": 0.765517685242224,
      "eval_loss": 0.04930509999394417,
      "eval_runtime": 513.7066,
      "eval_samples_per_second": 111.739,
      "eval_steps_per_second": 1.746,
      "learning_rate": 1e-05,
      "step": 384813
    },
    {
      "epoch": 143.06949089557784,
      "grad_norm": 0.243364155292511,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0527,
      "step": 385000
    },
    {
      "epoch": 143.25529542920847,
      "grad_norm": 0.2838144600391388,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.053,
      "step": 385500
    },
    {
      "epoch": 143.4410999628391,
      "grad_norm": 0.24525409936904907,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0528,
      "step": 386000
    },
    {
      "epoch": 143.6269044964697,
      "grad_norm": 0.2145887315273285,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0527,
      "step": 386500
    },
    {
      "epoch": 143.81270903010034,
      "grad_norm": 0.16669905185699463,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0531,
      "step": 387000
    },
    {
      "epoch": 143.99851356373097,
      "grad_norm": 0.23091119527816772,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0528,
      "step": 387500
    },
    {
      "epoch": 144.0,
      "eval_accuracy": 0.7171477848817965,
      "eval_f1_macro": 0.3554021435508309,
      "eval_f1_micro": 0.7667940015206897,
      "eval_loss": 0.0490318201482296,
      "eval_runtime": 525.9642,
      "eval_samples_per_second": 109.135,
      "eval_steps_per_second": 1.705,
      "learning_rate": 1.0000000000000002e-06,
      "step": 387504
    },
    {
      "epoch": 144.18431809736157,
      "grad_norm": 0.2074396163225174,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0531,
      "step": 388000
    },
    {
      "epoch": 144.3701226309922,
      "grad_norm": 0.2579312026500702,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0529,
      "step": 388500
    },
    {
      "epoch": 144.55592716462283,
      "grad_norm": 0.1861879676580429,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0528,
      "step": 389000
    },
    {
      "epoch": 144.74173169825343,
      "grad_norm": 0.21441423892974854,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0526,
      "step": 389500
    },
    {
      "epoch": 144.92753623188406,
      "grad_norm": 0.23621511459350586,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0534,
      "step": 390000
    },
    {
      "epoch": 145.0,
      "eval_accuracy": 0.7201616696573231,
      "eval_f1_macro": 0.3711029550898432,
      "eval_f1_micro": 0.7677822164123848,
      "eval_loss": 0.04918621480464935,
      "eval_runtime": 507.392,
      "eval_samples_per_second": 113.129,
      "eval_steps_per_second": 1.768,
      "learning_rate": 1.0000000000000002e-06,
      "step": 390195
    },
    {
      "epoch": 145.1133407655147,
      "grad_norm": 0.2494996339082718,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0533,
      "step": 390500
    },
    {
      "epoch": 145.2991452991453,
      "grad_norm": 0.19221335649490356,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0526,
      "step": 391000
    },
    {
      "epoch": 145.48494983277592,
      "grad_norm": 0.19597986340522766,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0533,
      "step": 391500
    },
    {
      "epoch": 145.67075436640653,
      "grad_norm": 0.19304433465003967,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0526,
      "step": 392000
    },
    {
      "epoch": 145.85655890003716,
      "grad_norm": 0.21061711013317108,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0529,
      "step": 392500
    },
    {
      "epoch": 146.0,
      "eval_accuracy": 0.717130363582516,
      "eval_f1_macro": 0.368326447075977,
      "eval_f1_micro": 0.7665065530257804,
      "eval_loss": 0.04903709515929222,
      "eval_runtime": 590.7348,
      "eval_samples_per_second": 97.169,
      "eval_steps_per_second": 1.518,
      "learning_rate": 1.0000000000000002e-06,
      "step": 392886
    },
    {
      "epoch": 146.0423634336678,
      "grad_norm": 0.21325454115867615,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0529,
      "step": 393000
    },
    {
      "epoch": 146.2281679672984,
      "grad_norm": 0.15669451653957367,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0525,
      "step": 393500
    },
    {
      "epoch": 146.41397250092902,
      "grad_norm": 0.22324424982070923,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.053,
      "step": 394000
    },
    {
      "epoch": 146.59977703455965,
      "grad_norm": 0.192140132188797,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0531,
      "step": 394500
    },
    {
      "epoch": 146.78558156819025,
      "grad_norm": 0.20880526304244995,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0532,
      "step": 395000
    },
    {
      "epoch": 146.97138610182088,
      "grad_norm": 0.21843858063220978,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0532,
      "step": 395500
    },
    {
      "epoch": 147.0,
      "eval_accuracy": 0.720823679029982,
      "eval_f1_macro": 0.37476196073915324,
      "eval_f1_micro": 0.768544776459646,
      "eval_loss": 0.049094948917627335,
      "eval_runtime": 525.4912,
      "eval_samples_per_second": 109.233,
      "eval_steps_per_second": 1.707,
      "learning_rate": 1.0000000000000002e-06,
      "step": 395577
    },
    {
      "epoch": 147.1571906354515,
      "grad_norm": 0.17876408994197845,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0521,
      "step": 396000
    },
    {
      "epoch": 147.34299516908212,
      "grad_norm": 0.25762057304382324,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0528,
      "step": 396500
    },
    {
      "epoch": 147.52879970271275,
      "grad_norm": 0.20070230960845947,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.053,
      "step": 397000
    },
    {
      "epoch": 147.71460423634338,
      "grad_norm": 0.20209959149360657,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0536,
      "step": 397500
    },
    {
      "epoch": 147.90040876997398,
      "grad_norm": 0.22516289353370667,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.053,
      "step": 398000
    },
    {
      "epoch": 148.0,
      "eval_accuracy": 0.7167296736990645,
      "eval_f1_macro": 0.3649988602534311,
      "eval_f1_micro": 0.7667018106807243,
      "eval_loss": 0.04907181113958359,
      "eval_runtime": 525.537,
      "eval_samples_per_second": 109.224,
      "eval_steps_per_second": 1.707,
      "learning_rate": 1.0000000000000002e-06,
      "step": 398268
    },
    {
      "epoch": 148.0862133036046,
      "grad_norm": 0.2316817343235016,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0532,
      "step": 398500
    },
    {
      "epoch": 148.27201783723524,
      "grad_norm": 0.2039523720741272,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0525,
      "step": 399000
    },
    {
      "epoch": 148.45782237086584,
      "grad_norm": 0.22887806594371796,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0533,
      "step": 399500
    },
    {
      "epoch": 148.64362690449647,
      "grad_norm": 0.20459462702274323,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0527,
      "step": 400000
    },
    {
      "epoch": 148.8294314381271,
      "grad_norm": 0.20509076118469238,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0526,
      "step": 400500
    },
    {
      "epoch": 149.0,
      "eval_accuracy": 0.718210484137907,
      "eval_f1_macro": 0.3787860055208151,
      "eval_f1_micro": 0.7671139893046166,
      "eval_loss": 0.04904184117913246,
      "eval_runtime": 553.2009,
      "eval_samples_per_second": 103.762,
      "eval_steps_per_second": 1.621,
      "learning_rate": 1.0000000000000002e-06,
      "step": 400959
    },
    {
      "epoch": 149.0152359717577,
      "grad_norm": 0.2034793645143509,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0536,
      "step": 401000
    },
    {
      "epoch": 149.20104050538833,
      "grad_norm": 0.21667377650737762,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0531,
      "step": 401500
    },
    {
      "epoch": 149.38684503901897,
      "grad_norm": 0.23273786902427673,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0526,
      "step": 402000
    },
    {
      "epoch": 149.57264957264957,
      "grad_norm": 0.2024523764848709,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0528,
      "step": 402500
    },
    {
      "epoch": 149.7584541062802,
      "grad_norm": 0.19701418280601501,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0519,
      "step": 403000
    },
    {
      "epoch": 149.94425863991083,
      "grad_norm": 0.20358909666538239,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0535,
      "step": 403500
    },
    {
      "epoch": 150.0,
      "eval_accuracy": 0.7154404975523074,
      "eval_f1_macro": 0.3726446424747262,
      "eval_f1_micro": 0.7667920374277589,
      "eval_loss": 0.04912904277443886,
      "eval_runtime": 540.2131,
      "eval_samples_per_second": 106.256,
      "eval_steps_per_second": 1.66,
      "learning_rate": 1.0000000000000002e-06,
      "step": 403650
    },
    {
      "epoch": 150.0,
      "learning_rate": 1.0000000000000002e-07,
      "step": 403650,
      "total_flos": 2.3283011301956885e+21,
      "train_loss": 0.019644906779846472,
      "train_runtime": 108552.2715,
      "train_samples_per_second": 237.919,
      "train_steps_per_second": 3.718
    }
  ],
  "logging_steps": 500,
  "max_steps": 403650,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 150,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 10,
        "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.3283011301956885e+21,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}