{
  "best_metric": 0.9710622429847717,
  "best_model_checkpoint": "/scratch/czm5kz/NEW_finetuned_llama27b32_1_0.0003_alternate/checkpoint-1400",
  "epoch": 0.9975062344139651,
  "eval_steps": 20,
  "global_step": 1400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 1.9070518016815186,
      "learning_rate": 0.0002989308624376336,
      "loss": 3.4521,
      "step": 5
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.083425998687744,
      "learning_rate": 0.00029786172487526725,
      "loss": 2.6104,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.840538740158081,
      "learning_rate": 0.0002967925873129009,
      "loss": 2.0815,
      "step": 15
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.4302124977111816,
      "learning_rate": 0.00029572344975053457,
      "loss": 1.7791,
      "step": 20
    },
    {
      "epoch": 0.01,
      "eval_loss": 1.6359353065490723,
      "eval_runtime": 233.285,
      "eval_samples_per_second": 48.134,
      "eval_steps_per_second": 6.018,
      "step": 20
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.1669527292251587,
      "learning_rate": 0.00029465431218816815,
      "loss": 1.6289,
      "step": 25
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.1892757415771484,
      "learning_rate": 0.00029358517462580184,
      "loss": 1.4635,
      "step": 30
    },
    {
      "epoch": 0.02,
      "grad_norm": 0.9651042819023132,
      "learning_rate": 0.0002925160370634355,
      "loss": 1.3605,
      "step": 35
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3424170017242432,
      "learning_rate": 0.0002914468995010691,
      "loss": 1.3334,
      "step": 40
    },
    {
      "epoch": 0.03,
      "eval_loss": 1.2492247819900513,
      "eval_runtime": 233.1335,
      "eval_samples_per_second": 48.166,
      "eval_steps_per_second": 6.022,
      "step": 40
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.857738196849823,
      "learning_rate": 0.00029037776193870275,
      "loss": 1.1967,
      "step": 45
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.9371346235275269,
      "learning_rate": 0.0002893086243763364,
      "loss": 1.192,
      "step": 50
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.9375379681587219,
      "learning_rate": 0.00028823948681397,
      "loss": 1.1894,
      "step": 55
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.9225698113441467,
      "learning_rate": 0.0002871703492516037,
      "loss": 1.2074,
      "step": 60
    },
    {
      "epoch": 0.04,
      "eval_loss": 1.1329890489578247,
      "eval_runtime": 233.4936,
      "eval_samples_per_second": 48.091,
      "eval_steps_per_second": 6.013,
      "step": 60
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.7983834147453308,
      "learning_rate": 0.0002861012116892373,
      "loss": 1.1401,
      "step": 65
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.8089339733123779,
      "learning_rate": 0.000285032074126871,
      "loss": 1.0932,
      "step": 70
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.9041706323623657,
      "learning_rate": 0.0002839629365645046,
      "loss": 1.081,
      "step": 75
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.7338858842849731,
      "learning_rate": 0.00028289379900213826,
      "loss": 1.1233,
      "step": 80
    },
    {
      "epoch": 0.06,
      "eval_loss": 1.091849446296692,
      "eval_runtime": 233.0902,
      "eval_samples_per_second": 48.174,
      "eval_steps_per_second": 6.023,
      "step": 80
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.816347062587738,
      "learning_rate": 0.0002818246614397719,
      "loss": 1.1108,
      "step": 85
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.6140106916427612,
      "learning_rate": 0.00028075552387740553,
      "loss": 1.1325,
      "step": 90
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.6668590903282166,
      "learning_rate": 0.00027968638631503917,
      "loss": 1.0967,
      "step": 95
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.6617318391799927,
      "learning_rate": 0.0002786172487526728,
      "loss": 1.0699,
      "step": 100
    },
    {
      "epoch": 0.07,
      "eval_loss": 1.0730466842651367,
      "eval_runtime": 233.2743,
      "eval_samples_per_second": 48.136,
      "eval_steps_per_second": 6.019,
      "step": 100
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.7199958562850952,
      "learning_rate": 0.00027754811119030644,
      "loss": 1.1094,
      "step": 105
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.5779143571853638,
      "learning_rate": 0.0002764789736279401,
      "loss": 1.0849,
      "step": 110
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.5297559499740601,
      "learning_rate": 0.00027540983606557377,
      "loss": 1.059,
      "step": 115
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.794755220413208,
      "learning_rate": 0.0002743406985032074,
      "loss": 1.0519,
      "step": 120
    },
    {
      "epoch": 0.09,
      "eval_loss": 1.0585691928863525,
      "eval_runtime": 233.4391,
      "eval_samples_per_second": 48.102,
      "eval_steps_per_second": 6.014,
      "step": 120
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.5636884570121765,
      "learning_rate": 0.00027327156094084104,
      "loss": 1.0797,
      "step": 125
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.760064423084259,
      "learning_rate": 0.0002722024233784747,
      "loss": 1.0599,
      "step": 130
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.4716905951499939,
      "learning_rate": 0.0002711332858161083,
      "loss": 1.0662,
      "step": 135
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0044249296188354,
      "learning_rate": 0.00027006414825374195,
      "loss": 1.0562,
      "step": 140
    },
    {
      "epoch": 0.1,
      "eval_loss": 1.0480350255966187,
      "eval_runtime": 232.9379,
      "eval_samples_per_second": 48.206,
      "eval_steps_per_second": 6.027,
      "step": 140
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.5207091569900513,
      "learning_rate": 0.00026899501069137564,
      "loss": 1.0867,
      "step": 145
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.5663835406303406,
      "learning_rate": 0.0002679258731290092,
      "loss": 1.0417,
      "step": 150
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.6968746185302734,
      "learning_rate": 0.0002668567355666429,
      "loss": 1.0675,
      "step": 155
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.5393630862236023,
      "learning_rate": 0.00026578759800427654,
      "loss": 1.0537,
      "step": 160
    },
    {
      "epoch": 0.11,
      "eval_loss": 1.0404675006866455,
      "eval_runtime": 233.2342,
      "eval_samples_per_second": 48.145,
      "eval_steps_per_second": 6.02,
      "step": 160
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.6900107860565186,
      "learning_rate": 0.0002647184604419102,
      "loss": 1.028,
      "step": 165
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.6130568981170654,
      "learning_rate": 0.0002636493228795438,
      "loss": 1.0394,
      "step": 170
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.5123544931411743,
      "learning_rate": 0.00026258018531717745,
      "loss": 1.0185,
      "step": 175
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.5719022750854492,
      "learning_rate": 0.0002615110477548111,
      "loss": 1.0661,
      "step": 180
    },
    {
      "epoch": 0.13,
      "eval_loss": 1.0330013036727905,
      "eval_runtime": 232.9041,
      "eval_samples_per_second": 48.213,
      "eval_steps_per_second": 6.028,
      "step": 180
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.5089533925056458,
      "learning_rate": 0.0002604419101924447,
      "loss": 1.0173,
      "step": 185
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.4675886332988739,
      "learning_rate": 0.00025937277263007836,
      "loss": 0.9911,
      "step": 190
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.4334254264831543,
      "learning_rate": 0.000258303635067712,
      "loss": 1.047,
      "step": 195
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.4760352671146393,
      "learning_rate": 0.0002572344975053457,
      "loss": 1.0267,
      "step": 200
    },
    {
      "epoch": 0.14,
      "eval_loss": 1.027119755744934,
      "eval_runtime": 232.824,
      "eval_samples_per_second": 48.23,
      "eval_steps_per_second": 6.03,
      "step": 200
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.6061151027679443,
      "learning_rate": 0.0002561653599429793,
      "loss": 1.0051,
      "step": 205
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.5186893939971924,
      "learning_rate": 0.00025509622238061296,
      "loss": 1.012,
      "step": 210
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.5924214124679565,
      "learning_rate": 0.0002540270848182466,
      "loss": 1.0089,
      "step": 215
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.5817935466766357,
      "learning_rate": 0.00025295794725588023,
      "loss": 1.0317,
      "step": 220
    },
    {
      "epoch": 0.16,
      "eval_loss": 1.0244611501693726,
      "eval_runtime": 233.3845,
      "eval_samples_per_second": 48.114,
      "eval_steps_per_second": 6.016,
      "step": 220
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.5199680924415588,
      "learning_rate": 0.00025188880969351387,
      "loss": 0.9895,
      "step": 225
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.5646855235099792,
      "learning_rate": 0.00025081967213114756,
      "loss": 1.0858,
      "step": 230
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.47028979659080505,
      "learning_rate": 0.00024975053456878114,
      "loss": 1.0456,
      "step": 235
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.528374433517456,
      "learning_rate": 0.00024868139700641483,
      "loss": 1.0268,
      "step": 240
    },
    {
      "epoch": 0.17,
      "eval_loss": 1.026130199432373,
      "eval_runtime": 233.4901,
      "eval_samples_per_second": 48.092,
      "eval_steps_per_second": 6.013,
      "step": 240
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.557716429233551,
      "learning_rate": 0.00024761225944404847,
      "loss": 1.0249,
      "step": 245
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.6109133362770081,
      "learning_rate": 0.0002465431218816821,
      "loss": 1.0497,
      "step": 250
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.5170979499816895,
      "learning_rate": 0.00024547398431931574,
      "loss": 1.0874,
      "step": 255
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.48054125905036926,
      "learning_rate": 0.0002444048467569494,
      "loss": 1.0261,
      "step": 260
    },
    {
      "epoch": 0.19,
      "eval_loss": 1.0192761421203613,
      "eval_runtime": 233.4201,
      "eval_samples_per_second": 48.106,
      "eval_steps_per_second": 6.015,
      "step": 260
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.5639621019363403,
      "learning_rate": 0.000243335709194583,
      "loss": 1.001,
      "step": 265
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.5815649032592773,
      "learning_rate": 0.00024226657163221665,
      "loss": 0.9984,
      "step": 270
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.5705720782279968,
      "learning_rate": 0.0002411974340698503,
      "loss": 1.0214,
      "step": 275
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.5422970652580261,
      "learning_rate": 0.00024012829650748392,
      "loss": 1.0365,
      "step": 280
    },
    {
      "epoch": 0.2,
      "eval_loss": 1.0164554119110107,
      "eval_runtime": 233.0466,
      "eval_samples_per_second": 48.183,
      "eval_steps_per_second": 6.025,
      "step": 280
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.5650832653045654,
      "learning_rate": 0.00023905915894511758,
      "loss": 1.054,
      "step": 285
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.5521314740180969,
      "learning_rate": 0.00023799002138275122,
      "loss": 1.0778,
      "step": 290
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.5239119529724121,
      "learning_rate": 0.00023692088382038488,
      "loss": 0.9975,
      "step": 295
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.43719667196273804,
      "learning_rate": 0.0002358517462580185,
      "loss": 1.035,
      "step": 300
    },
    {
      "epoch": 0.21,
      "eval_loss": 1.0123059749603271,
      "eval_runtime": 233.8927,
      "eval_samples_per_second": 48.009,
      "eval_steps_per_second": 6.003,
      "step": 300
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.5832995772361755,
      "learning_rate": 0.00023478260869565215,
      "loss": 0.9933,
      "step": 305
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.47054553031921387,
      "learning_rate": 0.0002337134711332858,
      "loss": 1.0032,
      "step": 310
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.6376765370368958,
      "learning_rate": 0.00023264433357091945,
      "loss": 1.0231,
      "step": 315
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.47957855463027954,
      "learning_rate": 0.0002315751960085531,
      "loss": 1.0272,
      "step": 320
    },
    {
      "epoch": 0.23,
      "eval_loss": 1.0109120607376099,
      "eval_runtime": 233.0467,
      "eval_samples_per_second": 48.183,
      "eval_steps_per_second": 6.025,
      "step": 320
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.44474130868911743,
      "learning_rate": 0.00023050605844618672,
      "loss": 1.0308,
      "step": 325
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.5061362981796265,
      "learning_rate": 0.00022943692088382036,
      "loss": 1.0351,
      "step": 330
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.5571821331977844,
      "learning_rate": 0.00022836778332145402,
      "loss": 1.0549,
      "step": 335
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.4734516739845276,
      "learning_rate": 0.00022729864575908766,
      "loss": 0.9756,
      "step": 340
    },
    {
      "epoch": 0.24,
      "eval_loss": 1.0074673891067505,
      "eval_runtime": 233.4351,
      "eval_samples_per_second": 48.103,
      "eval_steps_per_second": 6.015,
      "step": 340
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.38644713163375854,
      "learning_rate": 0.00022622950819672127,
      "loss": 1.0128,
      "step": 345
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.6090478301048279,
      "learning_rate": 0.00022516037063435493,
      "loss": 1.0577,
      "step": 350
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.6028711795806885,
      "learning_rate": 0.00022409123307198857,
      "loss": 1.0256,
      "step": 355
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.5490003228187561,
      "learning_rate": 0.00022302209550962223,
      "loss": 1.0364,
      "step": 360
    },
    {
      "epoch": 0.26,
      "eval_loss": 1.005479097366333,
      "eval_runtime": 233.0631,
      "eval_samples_per_second": 48.18,
      "eval_steps_per_second": 6.024,
      "step": 360
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.4859737753868103,
      "learning_rate": 0.00022195295794725584,
      "loss": 1.0368,
      "step": 365
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.5286576151847839,
      "learning_rate": 0.0002208838203848895,
      "loss": 0.9919,
      "step": 370
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.5190165638923645,
      "learning_rate": 0.00021981468282252314,
      "loss": 1.0574,
      "step": 375
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.5330237746238708,
      "learning_rate": 0.0002187455452601568,
      "loss": 0.9796,
      "step": 380
    },
    {
      "epoch": 0.27,
      "eval_loss": 1.0031410455703735,
      "eval_runtime": 233.5388,
      "eval_samples_per_second": 48.082,
      "eval_steps_per_second": 6.012,
      "step": 380
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.598227858543396,
      "learning_rate": 0.0002176764076977904,
      "loss": 1.0521,
      "step": 385
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.4702519178390503,
      "learning_rate": 0.00021660727013542407,
      "loss": 1.0167,
      "step": 390
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.550994336605072,
      "learning_rate": 0.0002155381325730577,
      "loss": 1.0502,
      "step": 395
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.5564207434654236,
      "learning_rate": 0.00021446899501069137,
      "loss": 1.0414,
      "step": 400
    },
    {
      "epoch": 0.29,
      "eval_loss": 1.0028647184371948,
      "eval_runtime": 233.0518,
      "eval_samples_per_second": 48.182,
      "eval_steps_per_second": 6.024,
      "step": 400
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.4342661499977112,
      "learning_rate": 0.00021339985744832498,
      "loss": 0.9645,
      "step": 405
    },
    {
      "epoch": 0.29,
      "grad_norm": 2.041062593460083,
      "learning_rate": 0.00021233071988595865,
      "loss": 1.0252,
      "step": 410
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.516720712184906,
      "learning_rate": 0.00021126158232359228,
      "loss": 0.976,
      "step": 415
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.5575417876243591,
      "learning_rate": 0.00021019244476122595,
      "loss": 1.0176,
      "step": 420
    },
    {
      "epoch": 0.3,
      "eval_loss": 1.001124620437622,
      "eval_runtime": 233.1135,
      "eval_samples_per_second": 48.17,
      "eval_steps_per_second": 6.023,
      "step": 420
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.4411332309246063,
      "learning_rate": 0.00020912330719885958,
      "loss": 1.0407,
      "step": 425
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.3926582336425781,
      "learning_rate": 0.0002080541696364932,
      "loss": 1.0158,
      "step": 430
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.4546189308166504,
      "learning_rate": 0.00020698503207412685,
      "loss": 1.0102,
      "step": 435
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.5215502977371216,
      "learning_rate": 0.0002059158945117605,
      "loss": 1.0285,
      "step": 440
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.9982317686080933,
      "eval_runtime": 233.2445,
      "eval_samples_per_second": 48.143,
      "eval_steps_per_second": 6.019,
      "step": 440
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.6208043694496155,
      "learning_rate": 0.00020484675694939415,
      "loss": 1.0089,
      "step": 445
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.5218257904052734,
      "learning_rate": 0.00020377761938702776,
      "loss": 0.9841,
      "step": 450
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.5310038924217224,
      "learning_rate": 0.00020270848182466143,
      "loss": 0.9717,
      "step": 455
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.5639017820358276,
      "learning_rate": 0.00020163934426229506,
      "loss": 1.0441,
      "step": 460
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.9958406686782837,
      "eval_runtime": 233.4156,
      "eval_samples_per_second": 48.107,
      "eval_steps_per_second": 6.015,
      "step": 460
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.5642719864845276,
      "learning_rate": 0.00020057020669992872,
      "loss": 1.055,
      "step": 465
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.4470948576927185,
      "learning_rate": 0.00019950106913756233,
      "loss": 1.0306,
      "step": 470
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.5050873160362244,
      "learning_rate": 0.000198431931575196,
      "loss": 1.0047,
      "step": 475
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.3829382658004761,
      "learning_rate": 0.00019736279401282963,
      "loss": 1.0449,
      "step": 480
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.9962899684906006,
      "eval_runtime": 233.2659,
      "eval_samples_per_second": 48.138,
      "eval_steps_per_second": 6.019,
      "step": 480
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.36466968059539795,
      "learning_rate": 0.0001962936564504633,
      "loss": 1.067,
      "step": 485
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.4244692623615265,
      "learning_rate": 0.0001952245188880969,
      "loss": 0.985,
      "step": 490
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.38433513045310974,
      "learning_rate": 0.00019415538132573057,
      "loss": 0.988,
      "step": 495
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.41374388337135315,
      "learning_rate": 0.0001930862437633642,
      "loss": 0.9988,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.9949624538421631,
      "eval_runtime": 233.13,
      "eval_samples_per_second": 48.166,
      "eval_steps_per_second": 6.022,
      "step": 500
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.4772859811782837,
      "learning_rate": 0.00019201710620099787,
      "loss": 1.0462,
      "step": 505
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.48039695620536804,
      "learning_rate": 0.00019094796863863148,
      "loss": 1.0354,
      "step": 510
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.5138806104660034,
      "learning_rate": 0.0001898788310762651,
      "loss": 0.9772,
      "step": 515
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.4885447025299072,
      "learning_rate": 0.00018880969351389878,
      "loss": 1.0409,
      "step": 520
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.9945592880249023,
      "eval_runtime": 233.5393,
      "eval_samples_per_second": 48.082,
      "eval_steps_per_second": 6.012,
      "step": 520
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.44714003801345825,
      "learning_rate": 0.0001877405559515324,
      "loss": 0.9899,
      "step": 525
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.4275243282318115,
      "learning_rate": 0.00018667141838916605,
      "loss": 0.9873,
      "step": 530
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.4732975959777832,
      "learning_rate": 0.00018560228082679968,
      "loss": 1.006,
      "step": 535
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.4305308163166046,
      "learning_rate": 0.00018453314326443335,
      "loss": 0.9871,
      "step": 540
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.9933722019195557,
      "eval_runtime": 233.0714,
      "eval_samples_per_second": 48.178,
      "eval_steps_per_second": 6.024,
      "step": 540
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.4475422203540802,
      "learning_rate": 0.00018346400570206698,
      "loss": 0.9713,
      "step": 545
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.4801812767982483,
      "learning_rate": 0.00018239486813970065,
      "loss": 1.0059,
      "step": 550
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.48028790950775146,
      "learning_rate": 0.00018132573057733425,
      "loss": 0.9594,
      "step": 555
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.48806053400039673,
      "learning_rate": 0.00018025659301496792,
      "loss": 1.0144,
      "step": 560
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.9934199452400208,
      "eval_runtime": 232.9538,
      "eval_samples_per_second": 48.203,
      "eval_steps_per_second": 6.027,
      "step": 560
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.5405058860778809,
      "learning_rate": 0.00017918745545260155,
      "loss": 1.0135,
      "step": 565
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.3488529622554779,
      "learning_rate": 0.00017811831789023522,
      "loss": 0.9895,
      "step": 570
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.4228030741214752,
      "learning_rate": 0.00017704918032786883,
      "loss": 1.0373,
      "step": 575
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.3694140613079071,
      "learning_rate": 0.0001759800427655025,
      "loss": 0.9522,
      "step": 580
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.990608274936676,
      "eval_runtime": 233.3892,
      "eval_samples_per_second": 48.113,
      "eval_steps_per_second": 6.016,
      "step": 580
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.4826660752296448,
      "learning_rate": 0.00017491090520313613,
      "loss": 1.0129,
      "step": 585
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.40139585733413696,
      "learning_rate": 0.0001738417676407698,
      "loss": 0.9783,
      "step": 590
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.5151722431182861,
      "learning_rate": 0.0001727726300784034,
      "loss": 1.0676,
      "step": 595
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.37561264634132385,
      "learning_rate": 0.00017170349251603703,
      "loss": 1.0592,
      "step": 600
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.9911465644836426,
      "eval_runtime": 233.5858,
      "eval_samples_per_second": 48.072,
      "eval_steps_per_second": 6.011,
      "step": 600
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.4334346354007721,
      "learning_rate": 0.0001706343549536707,
      "loss": 0.9892,
      "step": 605
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.3643469512462616,
      "learning_rate": 0.00016956521739130433,
      "loss": 1.0195,
      "step": 610
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.4546561539173126,
      "learning_rate": 0.00016849607982893797,
      "loss": 0.9758,
      "step": 615
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.48632940649986267,
      "learning_rate": 0.0001674269422665716,
      "loss": 1.0421,
      "step": 620
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.9907721877098083,
      "eval_runtime": 233.4632,
      "eval_samples_per_second": 48.098,
      "eval_steps_per_second": 6.014,
      "step": 620
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.3957265317440033,
      "learning_rate": 0.00016635780470420527,
      "loss": 0.9755,
      "step": 625
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.5378085374832153,
      "learning_rate": 0.0001652886671418389,
      "loss": 1.0281,
      "step": 630
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.556155800819397,
      "learning_rate": 0.00016421952957947254,
      "loss": 0.9628,
      "step": 635
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.43633490800857544,
      "learning_rate": 0.00016315039201710618,
      "loss": 1.0179,
      "step": 640
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.9871463775634766,
      "eval_runtime": 233.9947,
      "eval_samples_per_second": 47.988,
      "eval_steps_per_second": 6.0,
      "step": 640
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.3990333378314972,
      "learning_rate": 0.00016208125445473984,
      "loss": 0.9964,
      "step": 645
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.36454010009765625,
      "learning_rate": 0.00016101211689237348,
      "loss": 1.0073,
      "step": 650
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.3725971579551697,
      "learning_rate": 0.00015994297933000714,
      "loss": 0.9967,
      "step": 655
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.5494146943092346,
      "learning_rate": 0.00015887384176764075,
      "loss": 0.9956,
      "step": 660
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.9869644641876221,
      "eval_runtime": 233.2617,
      "eval_samples_per_second": 48.139,
      "eval_steps_per_second": 6.019,
      "step": 660
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.4139350652694702,
      "learning_rate": 0.0001578047042052744,
      "loss": 1.022,
      "step": 665
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.44353967905044556,
      "learning_rate": 0.00015673556664290805,
      "loss": 1.0101,
      "step": 670
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.4779268205165863,
      "learning_rate": 0.0001556664290805417,
      "loss": 1.0261,
      "step": 675
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.37415191531181335,
      "learning_rate": 0.00015459729151817532,
      "loss": 0.9729,
      "step": 680
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.9881226420402527,
      "eval_runtime": 233.1792,
      "eval_samples_per_second": 48.156,
      "eval_steps_per_second": 6.021,
      "step": 680
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.4490317106246948,
      "learning_rate": 0.00015352815395580896,
      "loss": 0.9895,
      "step": 685
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.4588363766670227,
      "learning_rate": 0.00015245901639344262,
      "loss": 1.0078,
      "step": 690
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.4649104177951813,
      "learning_rate": 0.00015138987883107623,
      "loss": 0.9891,
      "step": 695
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.47752678394317627,
      "learning_rate": 0.0001503207412687099,
      "loss": 0.9936,
      "step": 700
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.9856240153312683,
      "eval_runtime": 233.0604,
      "eval_samples_per_second": 48.181,
      "eval_steps_per_second": 6.024,
      "step": 700
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.5587136745452881,
      "learning_rate": 0.00014925160370634355,
      "loss": 1.0269,
      "step": 705
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.4135577082633972,
      "learning_rate": 0.0001481824661439772,
      "loss": 0.9948,
      "step": 710
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.4373073875904083,
      "learning_rate": 0.00014711332858161083,
      "loss": 0.9559,
      "step": 715
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.4726034998893738,
      "learning_rate": 0.00014604419101924446,
      "loss": 0.9894,
      "step": 720
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.984818160533905,
      "eval_runtime": 233.4048,
      "eval_samples_per_second": 48.11,
      "eval_steps_per_second": 6.015,
      "step": 720
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.4481181800365448,
      "learning_rate": 0.0001449750534568781,
      "loss": 0.9723,
      "step": 725
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.3406706154346466,
      "learning_rate": 0.00014390591589451173,
      "loss": 0.9888,
      "step": 730
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.5130278468132019,
      "learning_rate": 0.0001428367783321454,
      "loss": 0.9928,
      "step": 735
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.36185356974601746,
      "learning_rate": 0.00014176764076977903,
      "loss": 0.9778,
      "step": 740
    },
    {
      "epoch": 0.53,
      "eval_loss": 0.9835929274559021,
      "eval_runtime": 233.3843,
      "eval_samples_per_second": 48.114,
      "eval_steps_per_second": 6.016,
      "step": 740
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.5359773635864258,
      "learning_rate": 0.00014069850320741267,
      "loss": 0.9865,
      "step": 745
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.5931240320205688,
      "learning_rate": 0.0001396293656450463,
      "loss": 1.011,
      "step": 750
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.44104117155075073,
      "learning_rate": 0.00013856022808267997,
      "loss": 1.0204,
      "step": 755
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.4355945587158203,
      "learning_rate": 0.0001374910905203136,
      "loss": 0.9671,
      "step": 760
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.9838274717330933,
      "eval_runtime": 233.2563,
      "eval_samples_per_second": 48.14,
      "eval_steps_per_second": 6.019,
      "step": 760
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.39338964223861694,
      "learning_rate": 0.00013642195295794724,
      "loss": 1.0193,
      "step": 765
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.43264439702033997,
      "learning_rate": 0.00013535281539558088,
      "loss": 0.9929,
      "step": 770
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.4693204462528229,
      "learning_rate": 0.00013428367783321454,
      "loss": 1.0113,
      "step": 775
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.3989047110080719,
      "learning_rate": 0.00013321454027084818,
      "loss": 1.0049,
      "step": 780
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.9828758239746094,
      "eval_runtime": 233.0119,
      "eval_samples_per_second": 48.191,
      "eval_steps_per_second": 6.025,
      "step": 780
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.3822576701641083,
      "learning_rate": 0.0001321454027084818,
      "loss": 0.998,
      "step": 785
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.3883315920829773,
      "learning_rate": 0.00013107626514611545,
      "loss": 1.0123,
      "step": 790
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.48141029477119446,
      "learning_rate": 0.0001300071275837491,
      "loss": 0.9956,
      "step": 795
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.43159806728363037,
      "learning_rate": 0.00012893799002138275,
      "loss": 0.9517,
      "step": 800
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.9815587401390076,
      "eval_runtime": 233.5763,
      "eval_samples_per_second": 48.074,
      "eval_steps_per_second": 6.011,
      "step": 800
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.4243312180042267,
      "learning_rate": 0.00012786885245901638,
      "loss": 0.9875,
      "step": 805
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.32536423206329346,
      "learning_rate": 0.00012679971489665002,
      "loss": 1.028,
      "step": 810
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.3888213038444519,
      "learning_rate": 0.00012573057733428366,
      "loss": 0.9964,
      "step": 815
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.35847336053848267,
      "learning_rate": 0.0001246614397719173,
      "loss": 0.9995,
      "step": 820
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.9827173948287964,
      "eval_runtime": 233.1259,
      "eval_samples_per_second": 48.167,
      "eval_steps_per_second": 6.022,
      "step": 820
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.43905025720596313,
      "learning_rate": 0.00012359230220955095,
      "loss": 1.0293,
      "step": 825
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.44588425755500793,
      "learning_rate": 0.0001225231646471846,
      "loss": 0.962,
      "step": 830
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.336024671792984,
      "learning_rate": 0.00012145402708481824,
      "loss": 1.0679,
      "step": 835
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.30530259013175964,
      "learning_rate": 0.00012038488952245188,
      "loss": 1.0149,
      "step": 840
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.9809694290161133,
      "eval_runtime": 232.8618,
      "eval_samples_per_second": 48.222,
      "eval_steps_per_second": 6.029,
      "step": 840
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.4134106934070587,
      "learning_rate": 0.00011931575196008553,
      "loss": 0.9827,
      "step": 845
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.4589226543903351,
      "learning_rate": 0.00011824661439771916,
      "loss": 0.9659,
      "step": 850
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.4869944155216217,
      "learning_rate": 0.00011717747683535281,
      "loss": 0.9849,
      "step": 855
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.4100353717803955,
      "learning_rate": 0.00011610833927298645,
      "loss": 0.9797,
      "step": 860
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.9812436103820801,
      "eval_runtime": 233.0856,
      "eval_samples_per_second": 48.175,
      "eval_steps_per_second": 6.024,
      "step": 860
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.4007159173488617,
      "learning_rate": 0.0001150392017106201,
      "loss": 1.0056,
      "step": 865
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.4553622603416443,
      "learning_rate": 0.00011397006414825373,
      "loss": 0.9957,
      "step": 870
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.3927219808101654,
      "learning_rate": 0.00011290092658588738,
      "loss": 0.9725,
      "step": 875
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.363148033618927,
      "learning_rate": 0.00011183178902352102,
      "loss": 0.9903,
      "step": 880
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.9790927171707153,
      "eval_runtime": 233.4773,
      "eval_samples_per_second": 48.095,
      "eval_steps_per_second": 6.013,
      "step": 880
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.4857732653617859,
      "learning_rate": 0.00011076265146115467,
      "loss": 1.0079,
      "step": 885
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.4485923647880554,
      "learning_rate": 0.00010969351389878829,
      "loss": 0.9998,
      "step": 890
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.40661126375198364,
      "learning_rate": 0.00010862437633642194,
      "loss": 1.0143,
      "step": 895
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.47750282287597656,
      "learning_rate": 0.00010755523877405558,
      "loss": 0.9909,
      "step": 900
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.9794048070907593,
      "eval_runtime": 232.8011,
      "eval_samples_per_second": 48.234,
      "eval_steps_per_second": 6.031,
      "step": 900
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.49147945642471313,
      "learning_rate": 0.00010648610121168923,
      "loss": 1.0004,
      "step": 905
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.4777437448501587,
      "learning_rate": 0.00010541696364932286,
      "loss": 1.0,
      "step": 910
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.43491894006729126,
      "learning_rate": 0.00010434782608695651,
      "loss": 1.0349,
      "step": 915
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.4582345187664032,
      "learning_rate": 0.00010327868852459015,
      "loss": 1.0298,
      "step": 920
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.9796016812324524,
      "eval_runtime": 233.4945,
      "eval_samples_per_second": 48.091,
      "eval_steps_per_second": 6.013,
      "step": 920
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.4664052128791809,
      "learning_rate": 0.0001022095509622238,
      "loss": 1.0192,
      "step": 925
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.4029270112514496,
      "learning_rate": 0.00010114041339985743,
      "loss": 0.9866,
      "step": 930
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.5014646649360657,
      "learning_rate": 0.00010007127583749108,
      "loss": 0.9919,
      "step": 935
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.4741760790348053,
      "learning_rate": 9.900213827512472e-05,
      "loss": 0.9845,
      "step": 940
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.9778198599815369,
      "eval_runtime": 233.2505,
      "eval_samples_per_second": 48.141,
      "eval_steps_per_second": 6.019,
      "step": 940
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.3994337320327759,
      "learning_rate": 9.793300071275837e-05,
      "loss": 0.992,
      "step": 945
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.3798719048500061,
      "learning_rate": 9.686386315039202e-05,
      "loss": 1.0055,
      "step": 950
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.5074355006217957,
      "learning_rate": 9.579472558802566e-05,
      "loss": 1.0449,
      "step": 955
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.39967235922813416,
      "learning_rate": 9.47255880256593e-05,
      "loss": 0.9388,
      "step": 960
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.978825569152832,
      "eval_runtime": 233.051,
      "eval_samples_per_second": 48.183,
      "eval_steps_per_second": 6.024,
      "step": 960
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.47815588116645813,
      "learning_rate": 9.365645046329294e-05,
      "loss": 0.9669,
      "step": 965
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.41418758034706116,
      "learning_rate": 9.258731290092659e-05,
      "loss": 1.0471,
      "step": 970
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.47331690788269043,
      "learning_rate": 9.151817533856021e-05,
      "loss": 0.9924,
      "step": 975
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.6328206062316895,
      "learning_rate": 9.044903777619385e-05,
      "loss": 0.9483,
      "step": 980
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.9771455526351929,
      "eval_runtime": 233.1232,
      "eval_samples_per_second": 48.168,
      "eval_steps_per_second": 6.023,
      "step": 980
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.33134448528289795,
      "learning_rate": 8.93799002138275e-05,
      "loss": 0.9692,
      "step": 985
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.41730013489723206,
      "learning_rate": 8.831076265146115e-05,
      "loss": 0.97,
      "step": 990
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.3953218460083008,
      "learning_rate": 8.724162508909478e-05,
      "loss": 0.9786,
      "step": 995
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.5037208199501038,
      "learning_rate": 8.617248752672843e-05,
      "loss": 0.9988,
      "step": 1000
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.9770392775535583,
      "eval_runtime": 233.7995,
      "eval_samples_per_second": 48.028,
      "eval_steps_per_second": 6.005,
      "step": 1000
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.6080517172813416,
      "learning_rate": 8.510334996436207e-05,
      "loss": 0.9739,
      "step": 1005
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.41217270493507385,
      "learning_rate": 8.403421240199572e-05,
      "loss": 0.9599,
      "step": 1010
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.4043317139148712,
      "learning_rate": 8.296507483962936e-05,
      "loss": 0.958,
      "step": 1015
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.42344191670417786,
      "learning_rate": 8.1895937277263e-05,
      "loss": 0.9474,
      "step": 1020
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.9775366187095642,
      "eval_runtime": 233.032,
      "eval_samples_per_second": 48.187,
      "eval_steps_per_second": 6.025,
      "step": 1020
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.4266872704029083,
      "learning_rate": 8.082679971489664e-05,
      "loss": 0.9744,
      "step": 1025
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.53822261095047,
      "learning_rate": 7.975766215253029e-05,
      "loss": 1.0066,
      "step": 1030
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.47148677706718445,
      "learning_rate": 7.868852459016393e-05,
      "loss": 1.0076,
      "step": 1035
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.4493381977081299,
      "learning_rate": 7.761938702779758e-05,
      "loss": 1.0063,
      "step": 1040
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.9757411479949951,
      "eval_runtime": 233.1816,
      "eval_samples_per_second": 48.156,
      "eval_steps_per_second": 6.021,
      "step": 1040
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.3722420334815979,
      "learning_rate": 7.655024946543121e-05,
      "loss": 0.9636,
      "step": 1045
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.3606567680835724,
      "learning_rate": 7.548111190306486e-05,
      "loss": 1.0114,
      "step": 1050
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.520788311958313,
      "learning_rate": 7.44119743406985e-05,
      "loss": 1.0032,
      "step": 1055
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.7319127321243286,
      "learning_rate": 7.334283677833213e-05,
      "loss": 1.0007,
      "step": 1060
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.9754329323768616,
      "eval_runtime": 233.5812,
      "eval_samples_per_second": 48.073,
      "eval_steps_per_second": 6.011,
      "step": 1060
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.4526929557323456,
      "learning_rate": 7.227369921596578e-05,
      "loss": 0.989,
      "step": 1065
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.4485223889350891,
      "learning_rate": 7.120456165359942e-05,
      "loss": 0.9985,
      "step": 1070
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.46079307794570923,
      "learning_rate": 7.013542409123307e-05,
      "loss": 1.0313,
      "step": 1075
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.43368515372276306,
      "learning_rate": 6.90662865288667e-05,
      "loss": 0.9722,
      "step": 1080
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.9747573137283325,
      "eval_runtime": 232.8905,
      "eval_samples_per_second": 48.216,
      "eval_steps_per_second": 6.029,
      "step": 1080
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.30304279923439026,
      "learning_rate": 6.799714896650034e-05,
      "loss": 0.9795,
      "step": 1085
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.46647265553474426,
      "learning_rate": 6.692801140413399e-05,
      "loss": 0.9787,
      "step": 1090
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.502044677734375,
      "learning_rate": 6.585887384176763e-05,
      "loss": 0.9896,
      "step": 1095
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.36911284923553467,
      "learning_rate": 6.478973627940128e-05,
      "loss": 0.9907,
      "step": 1100
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.9756913185119629,
      "eval_runtime": 233.558,
      "eval_samples_per_second": 48.078,
      "eval_steps_per_second": 6.011,
      "step": 1100
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.41324666142463684,
      "learning_rate": 6.372059871703493e-05,
      "loss": 0.9344,
      "step": 1105
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.543196976184845,
      "learning_rate": 6.265146115466856e-05,
      "loss": 0.9829,
      "step": 1110
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.45038723945617676,
      "learning_rate": 6.158232359230221e-05,
      "loss": 0.9591,
      "step": 1115
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.3415057957172394,
      "learning_rate": 6.051318602993584e-05,
      "loss": 0.9425,
      "step": 1120
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.9745773673057556,
      "eval_runtime": 233.0909,
      "eval_samples_per_second": 48.174,
      "eval_steps_per_second": 6.023,
      "step": 1120
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.5441497564315796,
      "learning_rate": 5.9444048467569485e-05,
      "loss": 0.9796,
      "step": 1125
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.42310476303100586,
      "learning_rate": 5.837491090520313e-05,
      "loss": 0.9625,
      "step": 1130
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.42857611179351807,
      "learning_rate": 5.730577334283677e-05,
      "loss": 0.9652,
      "step": 1135
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.3687695562839508,
      "learning_rate": 5.6236635780470413e-05,
      "loss": 0.9465,
      "step": 1140
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.974454939365387,
      "eval_runtime": 233.5503,
      "eval_samples_per_second": 48.08,
      "eval_steps_per_second": 6.012,
      "step": 1140
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.4901082217693329,
      "learning_rate": 5.5167498218104056e-05,
      "loss": 0.992,
      "step": 1145
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.5039891600608826,
      "learning_rate": 5.40983606557377e-05,
      "loss": 0.9943,
      "step": 1150
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.4841081202030182,
      "learning_rate": 5.302922309337134e-05,
      "loss": 0.9436,
      "step": 1155
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.37224721908569336,
      "learning_rate": 5.196008553100499e-05,
      "loss": 0.997,
      "step": 1160
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.9734376668930054,
      "eval_runtime": 233.22,
      "eval_samples_per_second": 48.148,
      "eval_steps_per_second": 6.02,
      "step": 1160
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.41128888726234436,
      "learning_rate": 5.089094796863862e-05,
      "loss": 0.9735,
      "step": 1165
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.406338095664978,
      "learning_rate": 4.9821810406272264e-05,
      "loss": 0.9955,
      "step": 1170
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.551630973815918,
      "learning_rate": 4.875267284390591e-05,
      "loss": 1.0032,
      "step": 1175
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.36810871958732605,
      "learning_rate": 4.7683535281539556e-05,
      "loss": 0.9371,
      "step": 1180
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.9730872511863708,
      "eval_runtime": 233.3639,
      "eval_samples_per_second": 48.118,
      "eval_steps_per_second": 6.016,
      "step": 1180
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.43923842906951904,
      "learning_rate": 4.66143977191732e-05,
      "loss": 0.9796,
      "step": 1185
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.4551779329776764,
      "learning_rate": 4.554526015680684e-05,
      "loss": 0.9973,
      "step": 1190
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.40367963910102844,
      "learning_rate": 4.4476122594440485e-05,
      "loss": 0.9833,
      "step": 1195
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.41799184679985046,
      "learning_rate": 4.340698503207413e-05,
      "loss": 0.9681,
      "step": 1200
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.9727602601051331,
      "eval_runtime": 233.417,
      "eval_samples_per_second": 48.107,
      "eval_steps_per_second": 6.015,
      "step": 1200
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.6011839509010315,
      "learning_rate": 4.2337847469707764e-05,
      "loss": 1.0271,
      "step": 1205
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.386572927236557,
      "learning_rate": 4.1268709907341407e-05,
      "loss": 0.9882,
      "step": 1210
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.5409009456634521,
      "learning_rate": 4.019957234497505e-05,
      "loss": 0.9772,
      "step": 1215
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.4123440980911255,
      "learning_rate": 3.913043478260869e-05,
      "loss": 0.988,
      "step": 1220
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.9725460410118103,
      "eval_runtime": 233.2561,
      "eval_samples_per_second": 48.14,
      "eval_steps_per_second": 6.019,
      "step": 1220
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.448880136013031,
      "learning_rate": 3.8061297220242335e-05,
      "loss": 0.9838,
      "step": 1225
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.33402350544929504,
      "learning_rate": 3.699215965787598e-05,
      "loss": 0.9564,
      "step": 1230
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.5954431891441345,
      "learning_rate": 3.592302209550962e-05,
      "loss": 0.9882,
      "step": 1235
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.410119891166687,
      "learning_rate": 3.485388453314326e-05,
      "loss": 0.9635,
      "step": 1240
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.9721638560295105,
      "eval_runtime": 233.3651,
      "eval_samples_per_second": 48.118,
      "eval_steps_per_second": 6.016,
      "step": 1240
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.3925500214099884,
      "learning_rate": 3.3784746970776906e-05,
      "loss": 0.9998,
      "step": 1245
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.4033336937427521,
      "learning_rate": 3.271560940841055e-05,
      "loss": 0.9943,
      "step": 1250
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.45633894205093384,
      "learning_rate": 3.164647184604419e-05,
      "loss": 1.0161,
      "step": 1255
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.44340837001800537,
      "learning_rate": 3.057733428367783e-05,
      "loss": 1.0065,
      "step": 1260
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.9722057580947876,
      "eval_runtime": 233.2613,
      "eval_samples_per_second": 48.139,
      "eval_steps_per_second": 6.019,
      "step": 1260
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.371636301279068,
      "learning_rate": 2.950819672131147e-05,
      "loss": 0.9746,
      "step": 1265
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.35404834151268005,
      "learning_rate": 2.8439059158945114e-05,
      "loss": 1.0016,
      "step": 1270
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.3891022503376007,
      "learning_rate": 2.736992159657876e-05,
      "loss": 0.9371,
      "step": 1275
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.4783947765827179,
      "learning_rate": 2.63007840342124e-05,
      "loss": 0.9776,
      "step": 1280
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.9719991087913513,
      "eval_runtime": 233.0799,
      "eval_samples_per_second": 48.177,
      "eval_steps_per_second": 6.024,
      "step": 1280
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.5107392072677612,
      "learning_rate": 2.5231646471846042e-05,
      "loss": 0.9662,
      "step": 1285
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.46247023344039917,
      "learning_rate": 2.4162508909479685e-05,
      "loss": 0.9845,
      "step": 1290
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.42233213782310486,
      "learning_rate": 2.3093371347113328e-05,
      "loss": 0.9892,
      "step": 1295
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.439802348613739,
      "learning_rate": 2.2024233784746968e-05,
      "loss": 1.0365,
      "step": 1300
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.9716612696647644,
      "eval_runtime": 232.9368,
      "eval_samples_per_second": 48.206,
      "eval_steps_per_second": 6.027,
      "step": 1300
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.3681885004043579,
      "learning_rate": 2.095509622238061e-05,
      "loss": 0.9908,
      "step": 1305
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.49940550327301025,
      "learning_rate": 1.9885958660014253e-05,
      "loss": 1.0436,
      "step": 1310
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.44985833764076233,
      "learning_rate": 1.8816821097647896e-05,
      "loss": 0.9962,
      "step": 1315
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.44092249870300293,
      "learning_rate": 1.774768353528154e-05,
      "loss": 0.9739,
      "step": 1320
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.9714374542236328,
      "eval_runtime": 232.9847,
      "eval_samples_per_second": 48.196,
      "eval_steps_per_second": 6.026,
      "step": 1320
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.49097079038619995,
      "learning_rate": 1.6678545972915182e-05,
      "loss": 1.0104,
      "step": 1325
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.40521782636642456,
      "learning_rate": 1.560940841054882e-05,
      "loss": 0.9506,
      "step": 1330
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.40690892934799194,
      "learning_rate": 1.4540270848182466e-05,
      "loss": 0.9733,
      "step": 1335
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.3609257638454437,
      "learning_rate": 1.3471133285816107e-05,
      "loss": 0.9915,
      "step": 1340
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.9715495109558105,
      "eval_runtime": 233.3093,
      "eval_samples_per_second": 48.129,
      "eval_steps_per_second": 6.018,
      "step": 1340
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.3654801547527313,
      "learning_rate": 1.240199572344975e-05,
      "loss": 0.9821,
      "step": 1345
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.46048375964164734,
      "learning_rate": 1.1332858161083391e-05,
      "loss": 0.9414,
      "step": 1350
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.7182672023773193,
      "learning_rate": 1.0263720598717034e-05,
      "loss": 0.9658,
      "step": 1355
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.4436035752296448,
      "learning_rate": 9.194583036350677e-06,
      "loss": 0.9906,
      "step": 1360
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.9712263941764832,
      "eval_runtime": 232.8136,
      "eval_samples_per_second": 48.232,
      "eval_steps_per_second": 6.031,
      "step": 1360
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.4624147415161133,
      "learning_rate": 8.12544547398432e-06,
      "loss": 0.9995,
      "step": 1365
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.36057719588279724,
      "learning_rate": 7.0563079116179615e-06,
      "loss": 0.9749,
      "step": 1370
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.4826163351535797,
      "learning_rate": 5.9871703492516035e-06,
      "loss": 0.9691,
      "step": 1375
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.46810442209243774,
      "learning_rate": 4.9180327868852455e-06,
      "loss": 0.9769,
      "step": 1380
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.9711294174194336,
      "eval_runtime": 233.1267,
      "eval_samples_per_second": 48.167,
      "eval_steps_per_second": 6.022,
      "step": 1380
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.4403376877307892,
      "learning_rate": 3.848895224518888e-06,
      "loss": 1.0044,
      "step": 1385
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.3866402804851532,
      "learning_rate": 2.7797576621525303e-06,
      "loss": 0.9635,
      "step": 1390
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.4226088523864746,
      "learning_rate": 1.7106200997861725e-06,
      "loss": 1.0251,
      "step": 1395
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.48196887969970703,
      "learning_rate": 6.414825374198146e-07,
      "loss": 0.9691,
      "step": 1400
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.9710622429847717,
      "eval_runtime": 233.1637,
      "eval_samples_per_second": 48.159,
      "eval_steps_per_second": 6.022,
      "step": 1400
    }
  ],
  "logging_steps": 5,
  "max_steps": 1403,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 20,
  "total_flos": 2.7134894097432576e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}