{
  "best_global_step": 3600,
  "best_metric": 0.6341557778664263,
  "best_model_checkpoint": "./experiments/modernbert-large-router-lr5e-6-ep5-20250808-213234/checkpoint-3600",
  "epoch": 2.719033232628399,
  "eval_steps": 100,
  "global_step": 3600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0007552870090634441,
      "grad_norm": 26.948183059692383,
      "learning_rate": 0.0,
      "loss": 1.3654,
      "step": 1
    },
    {
      "epoch": 0.0188821752265861,
      "grad_norm": 13.511099815368652,
      "learning_rate": 2.4000000000000003e-07,
      "loss": 1.686,
      "step": 25
    },
    {
      "epoch": 0.0377643504531722,
      "grad_norm": 11.786803245544434,
      "learning_rate": 4.900000000000001e-07,
      "loss": 1.508,
      "step": 50
    },
    {
      "epoch": 0.05664652567975831,
      "grad_norm": 12.238818168640137,
      "learning_rate": 7.4e-07,
      "loss": 1.4226,
      "step": 75
    },
    {
      "epoch": 0.0755287009063444,
      "grad_norm": 9.133111953735352,
      "learning_rate": 9.9e-07,
      "loss": 1.4262,
      "step": 100
    },
    {
      "epoch": 0.0755287009063444,
      "eval_accuracy": 0.5558073654390935,
      "eval_f1": 0.5538780809735232,
      "eval_loss": 0.6949775218963623,
      "eval_precision": 0.5525289134150683,
      "eval_recall": 0.5558073654390934,
      "eval_runtime": 26.307,
      "eval_samples_per_second": 201.277,
      "eval_steps_per_second": 12.582,
      "step": 100
    },
    {
      "epoch": 0.09441087613293052,
      "grad_norm": 7.815569877624512,
      "learning_rate": 1.2400000000000002e-06,
      "loss": 1.3694,
      "step": 125
    },
    {
      "epoch": 0.11329305135951662,
      "grad_norm": 18.00971794128418,
      "learning_rate": 1.4900000000000001e-06,
      "loss": 1.3346,
      "step": 150
    },
    {
      "epoch": 0.13217522658610273,
      "grad_norm": 8.447665214538574,
      "learning_rate": 1.74e-06,
      "loss": 1.3443,
      "step": 175
    },
    {
      "epoch": 0.1510574018126888,
      "grad_norm": 9.353869438171387,
      "learning_rate": 1.9900000000000004e-06,
      "loss": 1.3471,
      "step": 200
    },
    {
      "epoch": 0.1510574018126888,
      "eval_accuracy": 0.5933899905571294,
      "eval_f1": 0.5870828666987644,
      "eval_loss": 0.6730061173439026,
      "eval_precision": 0.5858765908974611,
      "eval_recall": 0.5933899905571294,
      "eval_runtime": 25.7751,
      "eval_samples_per_second": 205.431,
      "eval_steps_per_second": 12.842,
      "step": 200
    },
    {
      "epoch": 0.16993957703927492,
      "grad_norm": 6.845251560211182,
      "learning_rate": 2.24e-06,
      "loss": 1.3692,
      "step": 225
    },
    {
      "epoch": 0.18882175226586104,
      "grad_norm": 14.863338470458984,
      "learning_rate": 2.4900000000000003e-06,
      "loss": 1.3566,
      "step": 250
    },
    {
      "epoch": 0.20770392749244712,
      "grad_norm": 10.733919143676758,
      "learning_rate": 2.7400000000000004e-06,
      "loss": 1.3426,
      "step": 275
    },
    {
      "epoch": 0.22658610271903323,
      "grad_norm": 8.589818000793457,
      "learning_rate": 2.99e-06,
      "loss": 1.3253,
      "step": 300
    },
    {
      "epoch": 0.22658610271903323,
      "eval_accuracy": 0.6073654390934844,
      "eval_f1": 0.5758152044311772,
      "eval_loss": 0.6623404622077942,
      "eval_precision": 0.5958916118010293,
      "eval_recall": 0.6073654390934844,
      "eval_runtime": 25.6943,
      "eval_samples_per_second": 206.077,
      "eval_steps_per_second": 12.882,
      "step": 300
    },
    {
      "epoch": 0.24546827794561935,
      "grad_norm": 7.530769348144531,
      "learning_rate": 3.2400000000000003e-06,
      "loss": 1.3273,
      "step": 325
    },
    {
      "epoch": 0.26435045317220546,
      "grad_norm": 16.52275276184082,
      "learning_rate": 3.49e-06,
      "loss": 1.33,
      "step": 350
    },
    {
      "epoch": 0.28323262839879154,
      "grad_norm": 9.330696105957031,
      "learning_rate": 3.74e-06,
      "loss": 1.3396,
      "step": 375
    },
    {
      "epoch": 0.3021148036253776,
      "grad_norm": 4.76742696762085,
      "learning_rate": 3.990000000000001e-06,
      "loss": 1.359,
      "step": 400
    },
    {
      "epoch": 0.3021148036253776,
      "eval_accuracy": 0.6239848914069878,
      "eval_f1": 0.5901741272637462,
      "eval_loss": 0.6522805094718933,
      "eval_precision": 0.6190840821922633,
      "eval_recall": 0.6239848914069878,
      "eval_runtime": 27.5423,
      "eval_samples_per_second": 192.25,
      "eval_steps_per_second": 12.018,
      "step": 400
    },
    {
      "epoch": 0.32099697885196377,
      "grad_norm": 7.44769811630249,
      "learning_rate": 4.24e-06,
      "loss": 1.3093,
      "step": 425
    },
    {
      "epoch": 0.33987915407854985,
      "grad_norm": 8.336981773376465,
      "learning_rate": 4.49e-06,
      "loss": 1.2871,
      "step": 450
    },
    {
      "epoch": 0.35876132930513593,
      "grad_norm": 4.4486918449401855,
      "learning_rate": 4.74e-06,
      "loss": 1.2957,
      "step": 475
    },
    {
      "epoch": 0.3776435045317221,
      "grad_norm": 5.152225017547607,
      "learning_rate": 4.9900000000000005e-06,
      "loss": 1.2706,
      "step": 500
    },
    {
      "epoch": 0.3776435045317221,
      "eval_accuracy": 0.6251180358829084,
      "eval_f1": 0.5859680282877181,
      "eval_loss": 0.654725193977356,
      "eval_precision": 0.6233075274864223,
      "eval_recall": 0.6251180358829084,
      "eval_runtime": 26.3536,
      "eval_samples_per_second": 200.921,
      "eval_steps_per_second": 12.56,
      "step": 500
    },
    {
      "epoch": 0.39652567975830816,
      "grad_norm": 5.1736931800842285,
      "learning_rate": 4.980392156862746e-06,
      "loss": 1.3144,
      "step": 525
    },
    {
      "epoch": 0.41540785498489424,
      "grad_norm": 8.337306022644043,
      "learning_rate": 4.959967320261438e-06,
      "loss": 1.3004,
      "step": 550
    },
    {
      "epoch": 0.4342900302114804,
      "grad_norm": 21.556997299194336,
      "learning_rate": 4.939542483660132e-06,
      "loss": 1.281,
      "step": 575
    },
    {
      "epoch": 0.45317220543806647,
      "grad_norm": 10.470200538635254,
      "learning_rate": 4.919117647058823e-06,
      "loss": 1.3114,
      "step": 600
    },
    {
      "epoch": 0.45317220543806647,
      "eval_accuracy": 0.6168083097261567,
      "eval_f1": 0.5444872481570541,
      "eval_loss": 0.6602128148078918,
      "eval_precision": 0.6358703772903485,
      "eval_recall": 0.6168083097261567,
      "eval_runtime": 26.4011,
      "eval_samples_per_second": 200.56,
      "eval_steps_per_second": 12.537,
      "step": 600
    },
    {
      "epoch": 0.47205438066465255,
      "grad_norm": 7.532138824462891,
      "learning_rate": 4.8986928104575166e-06,
      "loss": 1.2788,
      "step": 625
    },
    {
      "epoch": 0.4909365558912387,
      "grad_norm": 8.277374267578125,
      "learning_rate": 4.87826797385621e-06,
      "loss": 1.283,
      "step": 650
    },
    {
      "epoch": 0.5098187311178247,
      "grad_norm": 4.102471828460693,
      "learning_rate": 4.857843137254902e-06,
      "loss": 1.2815,
      "step": 675
    },
    {
      "epoch": 0.5287009063444109,
      "grad_norm": 9.688183784484863,
      "learning_rate": 4.8374183006535956e-06,
      "loss": 1.3075,
      "step": 700
    },
    {
      "epoch": 0.5287009063444109,
      "eval_accuracy": 0.6356940509915014,
      "eval_f1": 0.6084915332524169,
      "eval_loss": 0.6435697078704834,
      "eval_precision": 0.6320921608809188,
      "eval_recall": 0.6356940509915014,
      "eval_runtime": 28.2231,
      "eval_samples_per_second": 187.612,
      "eval_steps_per_second": 11.728,
      "step": 700
    },
    {
      "epoch": 0.547583081570997,
      "grad_norm": 11.222074508666992,
      "learning_rate": 4.816993464052288e-06,
      "loss": 1.2944,
      "step": 725
    },
    {
      "epoch": 0.5664652567975831,
      "grad_norm": 5.912996768951416,
      "learning_rate": 4.796568627450981e-06,
      "loss": 1.3052,
      "step": 750
    },
    {
      "epoch": 0.5853474320241692,
      "grad_norm": 6.690607070922852,
      "learning_rate": 4.776143790849674e-06,
      "loss": 1.2849,
      "step": 775
    },
    {
      "epoch": 0.6042296072507553,
      "grad_norm": 3.7729241847991943,
      "learning_rate": 4.755718954248366e-06,
      "loss": 1.2988,
      "step": 800
    },
    {
      "epoch": 0.6042296072507553,
      "eval_accuracy": 0.6326723323890463,
      "eval_f1": 0.5875834641292582,
      "eval_loss": 0.6444017887115479,
      "eval_precision": 0.6395765540326241,
      "eval_recall": 0.6326723323890463,
      "eval_runtime": 26.5576,
      "eval_samples_per_second": 199.378,
      "eval_steps_per_second": 12.463,
      "step": 800
    },
    {
      "epoch": 0.6231117824773413,
      "grad_norm": 4.947546005249023,
      "learning_rate": 4.7352941176470594e-06,
      "loss": 1.2908,
      "step": 825
    },
    {
      "epoch": 0.6419939577039275,
      "grad_norm": 6.54949951171875,
      "learning_rate": 4.714869281045752e-06,
      "loss": 1.3008,
      "step": 850
    },
    {
      "epoch": 0.6608761329305136,
      "grad_norm": 5.186257362365723,
      "learning_rate": 4.694444444444445e-06,
      "loss": 1.3012,
      "step": 875
    },
    {
      "epoch": 0.6797583081570997,
      "grad_norm": 4.848501682281494,
      "learning_rate": 4.674019607843138e-06,
      "loss": 1.3748,
      "step": 900
    },
    {
      "epoch": 0.6797583081570997,
      "eval_accuracy": 0.6196411709159585,
      "eval_f1": 0.5413471089456644,
      "eval_loss": 0.6539095044136047,
      "eval_precision": 0.6522323982900737,
      "eval_recall": 0.6196411709159585,
      "eval_runtime": 26.3133,
      "eval_samples_per_second": 201.229,
      "eval_steps_per_second": 12.579,
      "step": 900
    },
    {
      "epoch": 0.6986404833836858,
      "grad_norm": 8.589735984802246,
      "learning_rate": 4.653594771241831e-06,
      "loss": 1.3267,
      "step": 925
    },
    {
      "epoch": 0.7175226586102719,
      "grad_norm": 7.5751848220825195,
      "learning_rate": 4.633169934640523e-06,
      "loss": 1.2539,
      "step": 950
    },
    {
      "epoch": 0.736404833836858,
      "grad_norm": 4.0868120193481445,
      "learning_rate": 4.612745098039217e-06,
      "loss": 1.312,
      "step": 975
    },
    {
      "epoch": 0.7552870090634441,
      "grad_norm": 5.95460844039917,
      "learning_rate": 4.592320261437909e-06,
      "loss": 1.2903,
      "step": 1000
    },
    {
      "epoch": 0.7552870090634441,
      "eval_accuracy": 0.6385269121813031,
      "eval_f1": 0.6173864182737019,
      "eval_loss": 0.6376426815986633,
      "eval_precision": 0.6332296117395951,
      "eval_recall": 0.6385269121813031,
      "eval_runtime": 28.415,
      "eval_samples_per_second": 186.345,
      "eval_steps_per_second": 11.649,
      "step": 1000
    },
    {
      "epoch": 0.7741691842900302,
      "grad_norm": 6.253991603851318,
      "learning_rate": 4.5718954248366015e-06,
      "loss": 1.3063,
      "step": 1025
    },
    {
      "epoch": 0.7930513595166163,
      "grad_norm": 3.0682971477508545,
      "learning_rate": 4.551470588235295e-06,
      "loss": 1.2785,
      "step": 1050
    },
    {
      "epoch": 0.8119335347432024,
      "grad_norm": 4.491682052612305,
      "learning_rate": 4.531045751633987e-06,
      "loss": 1.276,
      "step": 1075
    },
    {
      "epoch": 0.8308157099697885,
      "grad_norm": 7.531124114990234,
      "learning_rate": 4.5106209150326805e-06,
      "loss": 1.2757,
      "step": 1100
    },
    {
      "epoch": 0.8308157099697885,
      "eval_accuracy": 0.6417374881964117,
      "eval_f1": 0.6312473050221801,
      "eval_loss": 0.6380065679550171,
      "eval_precision": 0.6351146396632452,
      "eval_recall": 0.6417374881964117,
      "eval_runtime": 25.9777,
      "eval_samples_per_second": 203.828,
      "eval_steps_per_second": 12.742,
      "step": 1100
    },
    {
      "epoch": 0.8496978851963746,
      "grad_norm": 7.8424201011657715,
      "learning_rate": 4.490196078431373e-06,
      "loss": 1.2807,
      "step": 1125
    },
    {
      "epoch": 0.8685800604229608,
      "grad_norm": 8.517139434814453,
      "learning_rate": 4.469771241830066e-06,
      "loss": 1.2705,
      "step": 1150
    },
    {
      "epoch": 0.8874622356495468,
      "grad_norm": 5.245467185974121,
      "learning_rate": 4.449346405228759e-06,
      "loss": 1.2839,
      "step": 1175
    },
    {
      "epoch": 0.9063444108761329,
      "grad_norm": 6.94294548034668,
      "learning_rate": 4.428921568627451e-06,
      "loss": 1.2711,
      "step": 1200
    },
    {
      "epoch": 0.9063444108761329,
      "eval_accuracy": 0.6321057601510859,
      "eval_f1": 0.5846932538581225,
      "eval_loss": 0.6422750949859619,
      "eval_precision": 0.6407048384547473,
      "eval_recall": 0.6321057601510859,
      "eval_runtime": 25.7628,
      "eval_samples_per_second": 205.529,
      "eval_steps_per_second": 12.848,
      "step": 1200
    },
    {
      "epoch": 0.925226586102719,
      "grad_norm": 3.8326165676116943,
      "learning_rate": 4.408496732026144e-06,
      "loss": 1.2671,
      "step": 1225
    },
    {
      "epoch": 0.9441087613293051,
      "grad_norm": 4.557371139526367,
      "learning_rate": 4.388071895424837e-06,
      "loss": 1.3158,
      "step": 1250
    },
    {
      "epoch": 0.9629909365558912,
      "grad_norm": 3.9651763439178467,
      "learning_rate": 4.36764705882353e-06,
      "loss": 1.2823,
      "step": 1275
    },
    {
      "epoch": 0.9818731117824774,
      "grad_norm": 3.4279274940490723,
      "learning_rate": 4.3472222222222225e-06,
      "loss": 1.3348,
      "step": 1300
    },
    {
      "epoch": 0.9818731117824774,
      "eval_accuracy": 0.6245514636449481,
      "eval_f1": 0.6260536519033915,
      "eval_loss": 0.6523462533950806,
      "eval_precision": 0.6287153689730779,
      "eval_recall": 0.6245514636449481,
      "eval_runtime": 27.6559,
      "eval_samples_per_second": 191.46,
      "eval_steps_per_second": 11.969,
      "step": 1300
    },
    {
      "epoch": 1.0007552870090635,
      "grad_norm": 6.088149547576904,
      "learning_rate": 4.326797385620916e-06,
      "loss": 1.3225,
      "step": 1325
    },
    {
      "epoch": 1.0196374622356494,
      "grad_norm": 10.229233741760254,
      "learning_rate": 4.306372549019608e-06,
      "loss": 1.243,
      "step": 1350
    },
    {
      "epoch": 1.0385196374622356,
      "grad_norm": 3.2610249519348145,
      "learning_rate": 4.285947712418301e-06,
      "loss": 1.2981,
      "step": 1375
    },
    {
      "epoch": 1.0574018126888218,
      "grad_norm": 11.988770484924316,
      "learning_rate": 4.265522875816994e-06,
      "loss": 1.2404,
      "step": 1400
    },
    {
      "epoch": 1.0574018126888218,
      "eval_accuracy": 0.624929178470255,
      "eval_f1": 0.6264983238137061,
      "eval_loss": 0.6569391489028931,
      "eval_precision": 0.6293753800764041,
      "eval_recall": 0.624929178470255,
      "eval_runtime": 26.1047,
      "eval_samples_per_second": 202.837,
      "eval_steps_per_second": 12.68,
      "step": 1400
    },
    {
      "epoch": 1.0762839879154078,
      "grad_norm": 3.8238422870635986,
      "learning_rate": 4.245098039215686e-06,
      "loss": 1.2335,
      "step": 1425
    },
    {
      "epoch": 1.095166163141994,
      "grad_norm": 8.899502754211426,
      "learning_rate": 4.22467320261438e-06,
      "loss": 1.2353,
      "step": 1450
    },
    {
      "epoch": 1.11404833836858,
      "grad_norm": 7.564877510070801,
      "learning_rate": 4.204248366013072e-06,
      "loss": 1.2154,
      "step": 1475
    },
    {
      "epoch": 1.1329305135951662,
      "grad_norm": 4.982959270477295,
      "learning_rate": 4.183823529411765e-06,
      "loss": 1.2499,
      "step": 1500
    },
    {
      "epoch": 1.1329305135951662,
      "eval_accuracy": 0.6396600566572238,
      "eval_f1": 0.6198342963492217,
      "eval_loss": 0.6338247656822205,
      "eval_precision": 0.6341757111657282,
      "eval_recall": 0.6396600566572238,
      "eval_runtime": 26.367,
      "eval_samples_per_second": 200.819,
      "eval_steps_per_second": 12.554,
      "step": 1500
    },
    {
      "epoch": 1.1518126888217524,
      "grad_norm": 6.588780879974365,
      "learning_rate": 4.163398692810458e-06,
      "loss": 1.2348,
      "step": 1525
    },
    {
      "epoch": 1.1706948640483383,
      "grad_norm": 46.35456848144531,
      "learning_rate": 4.142973856209151e-06,
      "loss": 1.2517,
      "step": 1550
    },
    {
      "epoch": 1.1895770392749245,
      "grad_norm": 4.0942277908325195,
      "learning_rate": 4.1225490196078435e-06,
      "loss": 1.226,
      "step": 1575
    },
    {
      "epoch": 1.2084592145015105,
      "grad_norm": 30.18272590637207,
      "learning_rate": 4.102124183006536e-06,
      "loss": 1.2077,
      "step": 1600
    },
    {
      "epoch": 1.2084592145015105,
      "eval_accuracy": 0.6381491973559962,
      "eval_f1": 0.6242245896900378,
      "eval_loss": 0.6355974674224854,
      "eval_precision": 0.6312162931573174,
      "eval_recall": 0.6381491973559962,
      "eval_runtime": 26.2286,
      "eval_samples_per_second": 201.879,
      "eval_steps_per_second": 12.62,
      "step": 1600
    },
    {
      "epoch": 1.2273413897280967,
      "grad_norm": 5.6306843757629395,
      "learning_rate": 4.081699346405229e-06,
      "loss": 1.2495,
      "step": 1625
    },
    {
      "epoch": 1.2462235649546827,
      "grad_norm": 7.549539566040039,
      "learning_rate": 4.061274509803922e-06,
      "loss": 1.2438,
      "step": 1650
    },
    {
      "epoch": 1.2651057401812689,
      "grad_norm": 7.462466716766357,
      "learning_rate": 4.040849673202615e-06,
      "loss": 1.264,
      "step": 1675
    },
    {
      "epoch": 1.283987915407855,
      "grad_norm": 4.440084934234619,
      "learning_rate": 4.020424836601307e-06,
      "loss": 1.2455,
      "step": 1700
    },
    {
      "epoch": 1.283987915407855,
      "eval_accuracy": 0.6379603399433428,
      "eval_f1": 0.6099089285109587,
      "eval_loss": 0.6329221725463867,
      "eval_precision": 0.6355277706702459,
      "eval_recall": 0.6379603399433428,
      "eval_runtime": 27.9156,
      "eval_samples_per_second": 189.679,
      "eval_steps_per_second": 11.857,
      "step": 1700
    },
    {
      "epoch": 1.302870090634441,
      "grad_norm": 12.334700584411621,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.2148,
      "step": 1725
    },
    {
      "epoch": 1.3217522658610272,
      "grad_norm": 12.755899429321289,
      "learning_rate": 3.979575163398693e-06,
      "loss": 1.185,
      "step": 1750
    },
    {
      "epoch": 1.3406344410876132,
      "grad_norm": 4.876974582672119,
      "learning_rate": 3.9591503267973855e-06,
      "loss": 1.1808,
      "step": 1775
    },
    {
      "epoch": 1.3595166163141994,
      "grad_norm": 7.091792106628418,
      "learning_rate": 3.938725490196079e-06,
      "loss": 1.243,
      "step": 1800
    },
    {
      "epoch": 1.3595166163141994,
      "eval_accuracy": 0.6383380547686497,
      "eval_f1": 0.6127733372599106,
      "eval_loss": 0.6369228959083557,
      "eval_precision": 0.6347855200764598,
      "eval_recall": 0.6383380547686497,
      "eval_runtime": 25.6226,
      "eval_samples_per_second": 206.654,
      "eval_steps_per_second": 12.918,
      "step": 1800
    },
    {
      "epoch": 1.3783987915407856,
      "grad_norm": 20.21983528137207,
      "learning_rate": 3.918300653594771e-06,
      "loss": 1.2483,
      "step": 1825
    },
    {
      "epoch": 1.3972809667673716,
      "grad_norm": 10.870339393615723,
      "learning_rate": 3.8978758169934645e-06,
      "loss": 1.2401,
      "step": 1850
    },
    {
      "epoch": 1.4161631419939578,
      "grad_norm": 5.997837543487549,
      "learning_rate": 3.877450980392157e-06,
      "loss": 1.2504,
      "step": 1875
    },
    {
      "epoch": 1.4350453172205437,
      "grad_norm": 17.45680046081543,
      "learning_rate": 3.85702614379085e-06,
      "loss": 1.2222,
      "step": 1900
    },
    {
      "epoch": 1.4350453172205437,
      "eval_accuracy": 0.6368271954674221,
      "eval_f1": 0.5909856645024059,
      "eval_loss": 0.6384395956993103,
      "eval_precision": 0.647404076427884,
      "eval_recall": 0.6368271954674221,
      "eval_runtime": 25.6817,
      "eval_samples_per_second": 206.178,
      "eval_steps_per_second": 12.889,
      "step": 1900
    },
    {
      "epoch": 1.45392749244713,
      "grad_norm": 10.445354461669922,
      "learning_rate": 3.836601307189543e-06,
      "loss": 1.2701,
      "step": 1925
    },
    {
      "epoch": 1.4728096676737161,
      "grad_norm": 4.692181587219238,
      "learning_rate": 3.816176470588235e-06,
      "loss": 1.2394,
      "step": 1950
    },
    {
      "epoch": 1.491691842900302,
      "grad_norm": 10.470508575439453,
      "learning_rate": 3.7957516339869284e-06,
      "loss": 1.1962,
      "step": 1975
    },
    {
      "epoch": 1.510574018126888,
      "grad_norm": 7.276846885681152,
      "learning_rate": 3.7753267973856213e-06,
      "loss": 1.2455,
      "step": 2000
    },
    {
      "epoch": 1.510574018126888,
      "eval_accuracy": 0.6445703493862134,
      "eval_f1": 0.6293429417019504,
      "eval_loss": 0.636458694934845,
      "eval_precision": 0.6387100201588051,
      "eval_recall": 0.6445703493862134,
      "eval_runtime": 27.9815,
      "eval_samples_per_second": 189.232,
      "eval_steps_per_second": 11.829,
      "step": 2000
    },
    {
      "epoch": 1.5294561933534743,
      "grad_norm": 10.635672569274902,
      "learning_rate": 3.754901960784314e-06,
      "loss": 1.2269,
      "step": 2025
    },
    {
      "epoch": 1.5483383685800605,
      "grad_norm": 11.313166618347168,
      "learning_rate": 3.734477124183007e-06,
      "loss": 1.251,
      "step": 2050
    },
    {
      "epoch": 1.5672205438066467,
      "grad_norm": 6.484985828399658,
      "learning_rate": 3.7140522875817e-06,
      "loss": 1.2621,
      "step": 2075
    },
    {
      "epoch": 1.5861027190332326,
      "grad_norm": 4.878018379211426,
      "learning_rate": 3.6936274509803927e-06,
      "loss": 1.2292,
      "step": 2100
    },
    {
      "epoch": 1.5861027190332326,
      "eval_accuracy": 0.6449480642115203,
      "eval_f1": 0.6179400117774106,
      "eval_loss": 0.6318739652633667,
      "eval_precision": 0.6442848843529498,
      "eval_recall": 0.6449480642115203,
      "eval_runtime": 26.4812,
      "eval_samples_per_second": 199.953,
      "eval_steps_per_second": 12.499,
      "step": 2100
    },
    {
      "epoch": 1.6049848942598186,
      "grad_norm": 14.128438949584961,
      "learning_rate": 3.6732026143790856e-06,
      "loss": 1.2229,
      "step": 2125
    },
    {
      "epoch": 1.6238670694864048,
      "grad_norm": 8.369355201721191,
      "learning_rate": 3.652777777777778e-06,
      "loss": 1.2459,
      "step": 2150
    },
    {
      "epoch": 1.642749244712991,
      "grad_norm": 6.0107808113098145,
      "learning_rate": 3.632352941176471e-06,
      "loss": 1.2377,
      "step": 2175
    },
    {
      "epoch": 1.6616314199395772,
      "grad_norm": 24.867504119873047,
      "learning_rate": 3.6119281045751637e-06,
      "loss": 1.2272,
      "step": 2200
    },
    {
      "epoch": 1.6616314199395772,
      "eval_accuracy": 0.6237960339943343,
      "eval_f1": 0.5542731710957122,
      "eval_loss": 0.6526551842689514,
      "eval_precision": 0.6494505292663029,
      "eval_recall": 0.6237960339943343,
      "eval_runtime": 26.5153,
      "eval_samples_per_second": 199.696,
      "eval_steps_per_second": 12.483,
      "step": 2200
    },
    {
      "epoch": 1.6805135951661632,
      "grad_norm": 9.593733787536621,
      "learning_rate": 3.5915032679738566e-06,
      "loss": 1.2012,
      "step": 2225
    },
    {
      "epoch": 1.6993957703927491,
      "grad_norm": 6.158414840698242,
      "learning_rate": 3.5710784313725494e-06,
      "loss": 1.2514,
      "step": 2250
    },
    {
      "epoch": 1.7182779456193353,
      "grad_norm": 7.738510608673096,
      "learning_rate": 3.5506535947712423e-06,
      "loss": 1.2585,
      "step": 2275
    },
    {
      "epoch": 1.7371601208459215,
      "grad_norm": 6.2952423095703125,
      "learning_rate": 3.530228758169935e-06,
      "loss": 1.2085,
      "step": 2300
    },
    {
      "epoch": 1.7371601208459215,
      "eval_accuracy": 0.6474032105760151,
      "eval_f1": 0.6281159209295336,
      "eval_loss": 0.630427360534668,
      "eval_precision": 0.6433494506393627,
      "eval_recall": 0.6474032105760151,
      "eval_runtime": 28.1713,
      "eval_samples_per_second": 187.957,
      "eval_steps_per_second": 11.75,
      "step": 2300
    },
    {
      "epoch": 1.7560422960725075,
      "grad_norm": 4.894021987915039,
      "learning_rate": 3.5098039215686276e-06,
      "loss": 1.2198,
      "step": 2325
    },
    {
      "epoch": 1.7749244712990937,
      "grad_norm": 10.691727638244629,
      "learning_rate": 3.4893790849673205e-06,
      "loss": 1.2269,
      "step": 2350
    },
    {
      "epoch": 1.7938066465256797,
      "grad_norm": 6.024422645568848,
      "learning_rate": 3.4689542483660133e-06,
      "loss": 1.2162,
      "step": 2375
    },
    {
      "epoch": 1.8126888217522659,
      "grad_norm": 7.545328140258789,
      "learning_rate": 3.448529411764706e-06,
      "loss": 1.2114,
      "step": 2400
    },
    {
      "epoch": 1.8126888217522659,
      "eval_accuracy": 0.6394711992445703,
      "eval_f1": 0.6140748119530789,
      "eval_loss": 0.636786162853241,
      "eval_precision": 0.6361754382587972,
      "eval_recall": 0.6394711992445703,
      "eval_runtime": 27.0479,
      "eval_samples_per_second": 195.764,
      "eval_steps_per_second": 12.238,
      "step": 2400
    },
    {
      "epoch": 1.831570996978852,
      "grad_norm": 11.729330062866211,
      "learning_rate": 3.428104575163399e-06,
      "loss": 1.2473,
      "step": 2425
    },
    {
      "epoch": 1.850453172205438,
      "grad_norm": 8.259448051452637,
      "learning_rate": 3.407679738562092e-06,
      "loss": 1.2086,
      "step": 2450
    },
    {
      "epoch": 1.869335347432024,
      "grad_norm": 7.117493629455566,
      "learning_rate": 3.3872549019607848e-06,
      "loss": 1.2435,
      "step": 2475
    },
    {
      "epoch": 1.8882175226586102,
      "grad_norm": 11.000151634216309,
      "learning_rate": 3.3668300653594776e-06,
      "loss": 1.2081,
      "step": 2500
    },
    {
      "epoch": 1.8882175226586102,
      "eval_accuracy": 0.6457034938621341,
      "eval_f1": 0.6192035548672061,
      "eval_loss": 0.632973849773407,
      "eval_precision": 0.6449753360936373,
      "eval_recall": 0.6457034938621341,
      "eval_runtime": 26.2323,
      "eval_samples_per_second": 201.851,
      "eval_steps_per_second": 12.618,
      "step": 2500
    },
    {
      "epoch": 1.9070996978851964,
      "grad_norm": 6.117466926574707,
      "learning_rate": 3.34640522875817e-06,
      "loss": 1.2099,
      "step": 2525
    },
    {
      "epoch": 1.9259818731117826,
      "grad_norm": 7.307370662689209,
      "learning_rate": 3.325980392156863e-06,
      "loss": 1.1788,
      "step": 2550
    },
    {
      "epoch": 1.9448640483383686,
      "grad_norm": 9.010822296142578,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 1.2212,
      "step": 2575
    },
    {
      "epoch": 1.9637462235649545,
      "grad_norm": 7.040300369262695,
      "learning_rate": 3.2851307189542486e-06,
      "loss": 1.2359,
      "step": 2600
    },
    {
      "epoch": 1.9637462235649545,
      "eval_accuracy": 0.6436260623229462,
      "eval_f1": 0.618869473778857,
      "eval_loss": 0.6305481195449829,
      "eval_precision": 0.6412356250382201,
      "eval_recall": 0.6436260623229462,
      "eval_runtime": 27.5332,
      "eval_samples_per_second": 192.313,
      "eval_steps_per_second": 12.022,
      "step": 2600
    },
    {
      "epoch": 1.9826283987915407,
      "grad_norm": 14.115243911743164,
      "learning_rate": 3.2647058823529415e-06,
      "loss": 1.2123,
      "step": 2625
    },
    {
      "epoch": 2.001510574018127,
      "grad_norm": 13.984496116638184,
      "learning_rate": 3.2442810457516343e-06,
      "loss": 1.2124,
      "step": 2650
    },
    {
      "epoch": 2.020392749244713,
      "grad_norm": 9.57017993927002,
      "learning_rate": 3.223856209150327e-06,
      "loss": 1.0599,
      "step": 2675
    },
    {
      "epoch": 2.039274924471299,
      "grad_norm": 10.857102394104004,
      "learning_rate": 3.20343137254902e-06,
      "loss": 1.113,
      "step": 2700
    },
    {
      "epoch": 2.039274924471299,
      "eval_accuracy": 0.6494806421152031,
      "eval_f1": 0.6320126937395264,
      "eval_loss": 0.6417879462242126,
      "eval_precision": 0.6451152563709769,
      "eval_recall": 0.6494806421152031,
      "eval_runtime": 25.6285,
      "eval_samples_per_second": 206.606,
      "eval_steps_per_second": 12.915,
      "step": 2700
    },
    {
      "epoch": 2.058157099697885,
      "grad_norm": 7.998616695404053,
      "learning_rate": 3.1830065359477125e-06,
      "loss": 1.0953,
      "step": 2725
    },
    {
      "epoch": 2.0770392749244713,
      "grad_norm": 11.49893856048584,
      "learning_rate": 3.1625816993464054e-06,
      "loss": 1.0805,
      "step": 2750
    },
    {
      "epoch": 2.0959214501510575,
      "grad_norm": 33.727813720703125,
      "learning_rate": 3.1421568627450982e-06,
      "loss": 1.113,
      "step": 2775
    },
    {
      "epoch": 2.1148036253776437,
      "grad_norm": 13.618017196655273,
      "learning_rate": 3.121732026143791e-06,
      "loss": 1.0903,
      "step": 2800
    },
    {
      "epoch": 2.1148036253776437,
      "eval_accuracy": 0.6343720491029273,
      "eval_f1": 0.5939358416916795,
      "eval_loss": 0.6593873500823975,
      "eval_precision": 0.6384799226464456,
      "eval_recall": 0.6343720491029273,
      "eval_runtime": 25.6715,
      "eval_samples_per_second": 206.26,
      "eval_steps_per_second": 12.894,
      "step": 2800
    },
    {
      "epoch": 2.1336858006042294,
      "grad_norm": 17.023202896118164,
      "learning_rate": 3.101307189542484e-06,
      "loss": 1.092,
      "step": 2825
    },
    {
      "epoch": 2.1525679758308156,
      "grad_norm": 93.96994018554688,
      "learning_rate": 3.080882352941177e-06,
      "loss": 1.0704,
      "step": 2850
    },
    {
      "epoch": 2.171450151057402,
      "grad_norm": 18.93490982055664,
      "learning_rate": 3.0604575163398697e-06,
      "loss": 1.0752,
      "step": 2875
    },
    {
      "epoch": 2.190332326283988,
      "grad_norm": 16.731698989868164,
      "learning_rate": 3.040032679738562e-06,
      "loss": 1.0768,
      "step": 2900
    },
    {
      "epoch": 2.190332326283988,
      "eval_accuracy": 0.6311614730878187,
      "eval_f1": 0.6227425810648518,
      "eval_loss": 0.6634938716888428,
      "eval_precision": 0.624132285789951,
      "eval_recall": 0.6311614730878187,
      "eval_runtime": 26.3588,
      "eval_samples_per_second": 200.881,
      "eval_steps_per_second": 12.557,
      "step": 2900
    },
    {
      "epoch": 2.209214501510574,
      "grad_norm": 11.118818283081055,
      "learning_rate": 3.019607843137255e-06,
      "loss": 1.0888,
      "step": 2925
    },
    {
      "epoch": 2.22809667673716,
      "grad_norm": 10.985156059265137,
      "learning_rate": 2.999183006535948e-06,
      "loss": 1.1436,
      "step": 2950
    },
    {
      "epoch": 2.246978851963746,
      "grad_norm": 19.878847122192383,
      "learning_rate": 2.9787581699346407e-06,
      "loss": 1.1073,
      "step": 2975
    },
    {
      "epoch": 2.2658610271903323,
      "grad_norm": 38.21413803100586,
      "learning_rate": 2.9583333333333335e-06,
      "loss": 1.0574,
      "step": 3000
    },
    {
      "epoch": 2.2658610271903323,
      "eval_accuracy": 0.6349386213408876,
      "eval_f1": 0.627970344866798,
      "eval_loss": 0.6681957840919495,
      "eval_precision": 0.6285863561364663,
      "eval_recall": 0.6349386213408876,
      "eval_runtime": 27.7894,
      "eval_samples_per_second": 190.54,
      "eval_steps_per_second": 11.911,
      "step": 3000
    },
    {
      "epoch": 2.2847432024169185,
      "grad_norm": 12.937191009521484,
      "learning_rate": 2.9379084967320264e-06,
      "loss": 1.0699,
      "step": 3025
    },
    {
      "epoch": 2.3036253776435047,
      "grad_norm": 20.07396125793457,
      "learning_rate": 2.9174836601307192e-06,
      "loss": 1.051,
      "step": 3050
    },
    {
      "epoch": 2.3225075528700905,
      "grad_norm": 15.067835807800293,
      "learning_rate": 2.897058823529412e-06,
      "loss": 1.0947,
      "step": 3075
    },
    {
      "epoch": 2.3413897280966767,
      "grad_norm": 42.31669998168945,
      "learning_rate": 2.8766339869281045e-06,
      "loss": 1.0713,
      "step": 3100
    },
    {
      "epoch": 2.3413897280966767,
      "eval_accuracy": 0.6277620396600566,
      "eval_f1": 0.6194707679474215,
      "eval_loss": 0.6624400615692139,
      "eval_precision": 0.6205786035913383,
      "eval_recall": 0.6277620396600566,
      "eval_runtime": 25.7958,
      "eval_samples_per_second": 205.266,
      "eval_steps_per_second": 12.832,
      "step": 3100
    },
    {
      "epoch": 2.360271903323263,
      "grad_norm": 18.090736389160156,
      "learning_rate": 2.8562091503267974e-06,
      "loss": 1.0245,
      "step": 3125
    },
    {
      "epoch": 2.379154078549849,
      "grad_norm": 52.456302642822266,
      "learning_rate": 2.8357843137254903e-06,
      "loss": 1.0917,
      "step": 3150
    },
    {
      "epoch": 2.398036253776435,
      "grad_norm": 13.822253227233887,
      "learning_rate": 2.815359477124183e-06,
      "loss": 1.0362,
      "step": 3175
    },
    {
      "epoch": 2.416918429003021,
      "grad_norm": 32.68608474731445,
      "learning_rate": 2.794934640522876e-06,
      "loss": 1.0118,
      "step": 3200
    },
    {
      "epoch": 2.416918429003021,
      "eval_accuracy": 0.6315391879131256,
      "eval_f1": 0.6185399761211638,
      "eval_loss": 0.6814322471618652,
      "eval_precision": 0.6238170726799035,
      "eval_recall": 0.6315391879131256,
      "eval_runtime": 25.2547,
      "eval_samples_per_second": 209.664,
      "eval_steps_per_second": 13.106,
      "step": 3200
    },
    {
      "epoch": 2.435800604229607,
      "grad_norm": 28.337984085083008,
      "learning_rate": 2.774509803921569e-06,
      "loss": 1.0688,
      "step": 3225
    },
    {
      "epoch": 2.4546827794561934,
      "grad_norm": 17.661848068237305,
      "learning_rate": 2.7540849673202617e-06,
      "loss": 1.0645,
      "step": 3250
    },
    {
      "epoch": 2.4735649546827796,
      "grad_norm": 24.61591339111328,
      "learning_rate": 2.733660130718954e-06,
      "loss": 1.084,
      "step": 3275
    },
    {
      "epoch": 2.4924471299093653,
      "grad_norm": 19.35943603515625,
      "learning_rate": 2.713235294117647e-06,
      "loss": 1.0458,
      "step": 3300
    },
    {
      "epoch": 2.4924471299093653,
      "eval_accuracy": 0.6343720491029273,
      "eval_f1": 0.6261022786359158,
      "eval_loss": 0.6693563461303711,
      "eval_precision": 0.6275730183984842,
      "eval_recall": 0.6343720491029273,
      "eval_runtime": 27.7312,
      "eval_samples_per_second": 190.94,
      "eval_steps_per_second": 11.936,
      "step": 3300
    },
    {
      "epoch": 2.5113293051359515,
      "grad_norm": 27.008237838745117,
      "learning_rate": 2.69281045751634e-06,
      "loss": 1.0649,
      "step": 3325
    },
    {
      "epoch": 2.5302114803625377,
      "grad_norm": 22.106470108032227,
      "learning_rate": 2.6723856209150327e-06,
      "loss": 1.0598,
      "step": 3350
    },
    {
      "epoch": 2.549093655589124,
      "grad_norm": 35.57805633544922,
      "learning_rate": 2.6519607843137256e-06,
      "loss": 1.0983,
      "step": 3375
    },
    {
      "epoch": 2.56797583081571,
      "grad_norm": 19.551307678222656,
      "learning_rate": 2.6315359477124184e-06,
      "loss": 1.0895,
      "step": 3400
    },
    {
      "epoch": 2.56797583081571,
      "eval_accuracy": 0.6330500472143532,
      "eval_f1": 0.6033695321475454,
      "eval_loss": 0.6630004644393921,
      "eval_precision": 0.6297347803000727,
      "eval_recall": 0.6330500472143532,
      "eval_runtime": 25.2264,
      "eval_samples_per_second": 209.899,
      "eval_steps_per_second": 13.121,
      "step": 3400
    },
    {
      "epoch": 2.586858006042296,
      "grad_norm": 17.806697845458984,
      "learning_rate": 2.6111111111111113e-06,
      "loss": 1.078,
      "step": 3425
    },
    {
      "epoch": 2.605740181268882,
      "grad_norm": 12.314886093139648,
      "learning_rate": 2.5906862745098046e-06,
      "loss": 1.0708,
      "step": 3450
    },
    {
      "epoch": 2.6246223564954683,
      "grad_norm": 24.937700271606445,
      "learning_rate": 2.5702614379084966e-06,
      "loss": 1.0807,
      "step": 3475
    },
    {
      "epoch": 2.6435045317220545,
      "grad_norm": 19.148723602294922,
      "learning_rate": 2.5498366013071894e-06,
      "loss": 1.0981,
      "step": 3500
    },
    {
      "epoch": 2.6435045317220545,
      "eval_accuracy": 0.6290840415486307,
      "eval_f1": 0.61111999303902,
      "eval_loss": 0.6673372387886047,
      "eval_precision": 0.6212227015642893,
      "eval_recall": 0.6290840415486307,
      "eval_runtime": 25.7251,
      "eval_samples_per_second": 205.83,
      "eval_steps_per_second": 12.867,
      "step": 3500
    },
    {
      "epoch": 2.6623867069486407,
      "grad_norm": 18.942354202270508,
      "learning_rate": 2.5294117647058823e-06,
      "loss": 1.0382,
      "step": 3525
    },
    {
      "epoch": 2.6812688821752264,
      "grad_norm": 13.00023365020752,
      "learning_rate": 2.508986928104575e-06,
      "loss": 1.0019,
      "step": 3550
    },
    {
      "epoch": 2.7001510574018126,
      "grad_norm": 18.894309997558594,
      "learning_rate": 2.488562091503268e-06,
      "loss": 1.0346,
      "step": 3575
    },
    {
      "epoch": 2.719033232628399,
      "grad_norm": 27.026264190673828,
      "learning_rate": 2.4681372549019613e-06,
      "loss": 1.0459,
      "step": 3600
    },
    {
      "epoch": 2.719033232628399,
      "eval_accuracy": 0.6356940509915014,
      "eval_f1": 0.6341557778664263,
      "eval_loss": 0.6878824234008789,
      "eval_precision": 0.6332597508413698,
      "eval_recall": 0.6356940509915014,
      "eval_runtime": 27.4703,
      "eval_samples_per_second": 192.754,
      "eval_steps_per_second": 12.049,
      "step": 3600
    }
  ],
  "logging_steps": 25,
  "max_steps": 6620,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.4268243718705856e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}