{
  "best_metric": 0.034271180629730225,
  "best_model_checkpoint": "saves/psy-course/Llama3-OpenBioLLM-8B/train/fold7/checkpoint-1950",
  "epoch": 4.9961802902979375,
  "eval_steps": 50,
  "global_step": 3270,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.015278838808250574,
      "grad_norm": 5.722738265991211,
      "learning_rate": 3.0581039755351682e-06,
      "loss": 1.3515,
      "step": 10
    },
    {
      "epoch": 0.030557677616501147,
      "grad_norm": 5.476120471954346,
      "learning_rate": 6.1162079510703365e-06,
      "loss": 1.2833,
      "step": 20
    },
    {
      "epoch": 0.04583651642475172,
      "grad_norm": 3.9035332202911377,
      "learning_rate": 9.174311926605506e-06,
      "loss": 1.0237,
      "step": 30
    },
    {
      "epoch": 0.061115355233002294,
      "grad_norm": 2.4191477298736572,
      "learning_rate": 1.2232415902140673e-05,
      "loss": 0.8029,
      "step": 40
    },
    {
      "epoch": 0.07639419404125286,
      "grad_norm": 2.650960683822632,
      "learning_rate": 1.5290519877675842e-05,
      "loss": 0.5143,
      "step": 50
    },
    {
      "epoch": 0.07639419404125286,
      "eval_loss": 0.30689939856529236,
      "eval_runtime": 192.1143,
      "eval_samples_per_second": 6.059,
      "eval_steps_per_second": 6.059,
      "step": 50
    },
    {
      "epoch": 0.09167303284950344,
      "grad_norm": 1.2615604400634766,
      "learning_rate": 1.834862385321101e-05,
      "loss": 0.2999,
      "step": 60
    },
    {
      "epoch": 0.10695187165775401,
      "grad_norm": 1.4181287288665771,
      "learning_rate": 2.140672782874618e-05,
      "loss": 0.2094,
      "step": 70
    },
    {
      "epoch": 0.12223071046600459,
      "grad_norm": 1.572482705116272,
      "learning_rate": 2.4464831804281346e-05,
      "loss": 0.1623,
      "step": 80
    },
    {
      "epoch": 0.13750954927425516,
      "grad_norm": 1.0281938314437866,
      "learning_rate": 2.7522935779816515e-05,
      "loss": 0.1138,
      "step": 90
    },
    {
      "epoch": 0.15278838808250572,
      "grad_norm": 1.8321815729141235,
      "learning_rate": 3.0581039755351684e-05,
      "loss": 0.1115,
      "step": 100
    },
    {
      "epoch": 0.15278838808250572,
      "eval_loss": 0.08002135157585144,
      "eval_runtime": 191.5562,
      "eval_samples_per_second": 6.077,
      "eval_steps_per_second": 6.077,
      "step": 100
    },
    {
      "epoch": 0.16806722689075632,
      "grad_norm": 1.0883564949035645,
      "learning_rate": 3.363914373088685e-05,
      "loss": 0.0747,
      "step": 110
    },
    {
      "epoch": 0.18334606569900688,
      "grad_norm": 1.412797212600708,
      "learning_rate": 3.669724770642202e-05,
      "loss": 0.0824,
      "step": 120
    },
    {
      "epoch": 0.19862490450725745,
      "grad_norm": 1.3624087572097778,
      "learning_rate": 3.9755351681957185e-05,
      "loss": 0.0737,
      "step": 130
    },
    {
      "epoch": 0.21390374331550802,
      "grad_norm": 1.5038727521896362,
      "learning_rate": 4.281345565749236e-05,
      "loss": 0.0822,
      "step": 140
    },
    {
      "epoch": 0.22918258212375858,
      "grad_norm": 1.0052672624588013,
      "learning_rate": 4.587155963302753e-05,
      "loss": 0.0781,
      "step": 150
    },
    {
      "epoch": 0.22918258212375858,
      "eval_loss": 0.06360089778900146,
      "eval_runtime": 192.0197,
      "eval_samples_per_second": 6.062,
      "eval_steps_per_second": 6.062,
      "step": 150
    },
    {
      "epoch": 0.24446142093200918,
      "grad_norm": 1.2365893125534058,
      "learning_rate": 4.892966360856269e-05,
      "loss": 0.0754,
      "step": 160
    },
    {
      "epoch": 0.2597402597402597,
      "grad_norm": 0.8989318609237671,
      "learning_rate": 5.1987767584097854e-05,
      "loss": 0.0666,
      "step": 170
    },
    {
      "epoch": 0.2750190985485103,
      "grad_norm": 0.5810503959655762,
      "learning_rate": 5.504587155963303e-05,
      "loss": 0.0705,
      "step": 180
    },
    {
      "epoch": 0.2902979373567609,
      "grad_norm": 1.2738935947418213,
      "learning_rate": 5.81039755351682e-05,
      "loss": 0.0741,
      "step": 190
    },
    {
      "epoch": 0.30557677616501144,
      "grad_norm": 0.8120205402374268,
      "learning_rate": 6.116207951070337e-05,
      "loss": 0.0677,
      "step": 200
    },
    {
      "epoch": 0.30557677616501144,
      "eval_loss": 0.059481341391801834,
      "eval_runtime": 190.8939,
      "eval_samples_per_second": 6.098,
      "eval_steps_per_second": 6.098,
      "step": 200
    },
    {
      "epoch": 0.32085561497326204,
      "grad_norm": 0.9474322199821472,
      "learning_rate": 6.422018348623854e-05,
      "loss": 0.0619,
      "step": 210
    },
    {
      "epoch": 0.33613445378151263,
      "grad_norm": 0.926907479763031,
      "learning_rate": 6.72782874617737e-05,
      "loss": 0.0806,
      "step": 220
    },
    {
      "epoch": 0.35141329258976317,
      "grad_norm": 0.8283106684684753,
      "learning_rate": 7.033639143730886e-05,
      "loss": 0.0677,
      "step": 230
    },
    {
      "epoch": 0.36669213139801377,
      "grad_norm": 0.6710531115531921,
      "learning_rate": 7.339449541284404e-05,
      "loss": 0.0606,
      "step": 240
    },
    {
      "epoch": 0.3819709702062643,
      "grad_norm": 0.729621946811676,
      "learning_rate": 7.645259938837921e-05,
      "loss": 0.0609,
      "step": 250
    },
    {
      "epoch": 0.3819709702062643,
      "eval_loss": 0.04995494335889816,
      "eval_runtime": 190.9738,
      "eval_samples_per_second": 6.095,
      "eval_steps_per_second": 6.095,
      "step": 250
    },
    {
      "epoch": 0.3972498090145149,
      "grad_norm": 0.7670791149139404,
      "learning_rate": 7.951070336391437e-05,
      "loss": 0.0462,
      "step": 260
    },
    {
      "epoch": 0.4125286478227655,
      "grad_norm": 1.007689118385315,
      "learning_rate": 8.256880733944955e-05,
      "loss": 0.0799,
      "step": 270
    },
    {
      "epoch": 0.42780748663101603,
      "grad_norm": 0.5677183270454407,
      "learning_rate": 8.562691131498472e-05,
      "loss": 0.0758,
      "step": 280
    },
    {
      "epoch": 0.4430863254392666,
      "grad_norm": 0.5293294787406921,
      "learning_rate": 8.868501529051988e-05,
      "loss": 0.0554,
      "step": 290
    },
    {
      "epoch": 0.45836516424751717,
      "grad_norm": 0.7656412720680237,
      "learning_rate": 9.174311926605506e-05,
      "loss": 0.0442,
      "step": 300
    },
    {
      "epoch": 0.45836516424751717,
      "eval_loss": 0.051944799721241,
      "eval_runtime": 191.224,
      "eval_samples_per_second": 6.087,
      "eval_steps_per_second": 6.087,
      "step": 300
    },
    {
      "epoch": 0.47364400305576776,
      "grad_norm": 0.5527473092079163,
      "learning_rate": 9.480122324159021e-05,
      "loss": 0.0433,
      "step": 310
    },
    {
      "epoch": 0.48892284186401835,
      "grad_norm": 0.3190058469772339,
      "learning_rate": 9.785932721712538e-05,
      "loss": 0.0546,
      "step": 320
    },
    {
      "epoch": 0.5042016806722689,
      "grad_norm": 0.486265629529953,
      "learning_rate": 9.999974360983129e-05,
      "loss": 0.0597,
      "step": 330
    },
    {
      "epoch": 0.5194805194805194,
      "grad_norm": 0.5260328650474548,
      "learning_rate": 9.999518563553522e-05,
      "loss": 0.0626,
      "step": 340
    },
    {
      "epoch": 0.5347593582887701,
      "grad_norm": 0.6759735941886902,
      "learning_rate": 9.998493069976636e-05,
      "loss": 0.0605,
      "step": 350
    },
    {
      "epoch": 0.5347593582887701,
      "eval_loss": 0.052965257316827774,
      "eval_runtime": 192.2041,
      "eval_samples_per_second": 6.056,
      "eval_steps_per_second": 6.056,
      "step": 350
    },
    {
      "epoch": 0.5500381970970206,
      "grad_norm": 0.8925366401672363,
      "learning_rate": 9.99689799710767e-05,
      "loss": 0.0647,
      "step": 360
    },
    {
      "epoch": 0.5653170359052712,
      "grad_norm": 0.5414925217628479,
      "learning_rate": 9.994733526705501e-05,
      "loss": 0.0662,
      "step": 370
    },
    {
      "epoch": 0.5805958747135218,
      "grad_norm": 0.5703091621398926,
      "learning_rate": 9.991999905411966e-05,
      "loss": 0.044,
      "step": 380
    },
    {
      "epoch": 0.5958747135217723,
      "grad_norm": 0.2930344045162201,
      "learning_rate": 9.988697444723762e-05,
      "loss": 0.0395,
      "step": 390
    },
    {
      "epoch": 0.6111535523300229,
      "grad_norm": 0.26469382643699646,
      "learning_rate": 9.984826520956949e-05,
      "loss": 0.0415,
      "step": 400
    },
    {
      "epoch": 0.6111535523300229,
      "eval_loss": 0.042674873024225235,
      "eval_runtime": 191.5669,
      "eval_samples_per_second": 6.076,
      "eval_steps_per_second": 6.076,
      "step": 400
    },
    {
      "epoch": 0.6264323911382735,
      "grad_norm": 0.45453405380249023,
      "learning_rate": 9.980387575204072e-05,
      "loss": 0.0442,
      "step": 410
    },
    {
      "epoch": 0.6417112299465241,
      "grad_norm": 0.5698592662811279,
      "learning_rate": 9.975381113283891e-05,
      "loss": 0.035,
      "step": 420
    },
    {
      "epoch": 0.6569900687547746,
      "grad_norm": 0.5528980493545532,
      "learning_rate": 9.969807705683751e-05,
      "loss": 0.0449,
      "step": 430
    },
    {
      "epoch": 0.6722689075630253,
      "grad_norm": 0.8066198229789734,
      "learning_rate": 9.96366798749457e-05,
      "loss": 0.03,
      "step": 440
    },
    {
      "epoch": 0.6875477463712758,
      "grad_norm": 0.3302541971206665,
      "learning_rate": 9.956962658338473e-05,
      "loss": 0.0616,
      "step": 450
    },
    {
      "epoch": 0.6875477463712758,
      "eval_loss": 0.04298438876867294,
      "eval_runtime": 190.871,
      "eval_samples_per_second": 6.098,
      "eval_steps_per_second": 6.098,
      "step": 450
    },
    {
      "epoch": 0.7028265851795263,
      "grad_norm": 0.3077934980392456,
      "learning_rate": 9.94969248228907e-05,
      "loss": 0.0393,
      "step": 460
    },
    {
      "epoch": 0.7181054239877769,
      "grad_norm": 0.46446335315704346,
      "learning_rate": 9.941858287784383e-05,
      "loss": 0.0514,
      "step": 470
    },
    {
      "epoch": 0.7333842627960275,
      "grad_norm": 0.5021407604217529,
      "learning_rate": 9.933460967532453e-05,
      "loss": 0.0513,
      "step": 480
    },
    {
      "epoch": 0.7486631016042781,
      "grad_norm": 0.5457306504249573,
      "learning_rate": 9.924501478409618e-05,
      "loss": 0.0527,
      "step": 490
    },
    {
      "epoch": 0.7639419404125286,
      "grad_norm": 0.38309529423713684,
      "learning_rate": 9.914980841351465e-05,
      "loss": 0.0412,
      "step": 500
    },
    {
      "epoch": 0.7639419404125286,
      "eval_loss": 0.03890955075621605,
      "eval_runtime": 191.337,
      "eval_samples_per_second": 6.084,
      "eval_steps_per_second": 6.084,
      "step": 500
    },
    {
      "epoch": 0.7792207792207793,
      "grad_norm": 0.4321908950805664,
      "learning_rate": 9.904900141236506e-05,
      "loss": 0.0329,
      "step": 510
    },
    {
      "epoch": 0.7944996180290298,
      "grad_norm": 0.5725101232528687,
      "learning_rate": 9.894260526762548e-05,
      "loss": 0.0507,
      "step": 520
    },
    {
      "epoch": 0.8097784568372803,
      "grad_norm": 0.7723375558853149,
      "learning_rate": 9.883063210315804e-05,
      "loss": 0.0627,
      "step": 530
    },
    {
      "epoch": 0.825057295645531,
      "grad_norm": 0.49958401918411255,
      "learning_rate": 9.871309467832738e-05,
      "loss": 0.0471,
      "step": 540
    },
    {
      "epoch": 0.8403361344537815,
      "grad_norm": 0.32487210631370544,
      "learning_rate": 9.859000638654674e-05,
      "loss": 0.0306,
      "step": 550
    },
    {
      "epoch": 0.8403361344537815,
      "eval_loss": 0.03846091032028198,
      "eval_runtime": 191.5057,
      "eval_samples_per_second": 6.078,
      "eval_steps_per_second": 6.078,
      "step": 550
    },
    {
      "epoch": 0.8556149732620321,
      "grad_norm": 0.5225566029548645,
      "learning_rate": 9.846138125375175e-05,
      "loss": 0.0523,
      "step": 560
    },
    {
      "epoch": 0.8708938120702827,
      "grad_norm": 0.3816261291503906,
      "learning_rate": 9.83272339368022e-05,
      "loss": 0.0352,
      "step": 570
    },
    {
      "epoch": 0.8861726508785333,
      "grad_norm": 0.3570096790790558,
      "learning_rate": 9.818757972181191e-05,
      "loss": 0.0547,
      "step": 580
    },
    {
      "epoch": 0.9014514896867838,
      "grad_norm": 0.6031997799873352,
      "learning_rate": 9.804243452240675e-05,
      "loss": 0.0377,
      "step": 590
    },
    {
      "epoch": 0.9167303284950343,
      "grad_norm": 0.955195963382721,
      "learning_rate": 9.789181487791146e-05,
      "loss": 0.0579,
      "step": 600
    },
    {
      "epoch": 0.9167303284950343,
      "eval_loss": 0.03784943372011185,
      "eval_runtime": 191.7708,
      "eval_samples_per_second": 6.07,
      "eval_steps_per_second": 6.07,
      "step": 600
    },
    {
      "epoch": 0.932009167303285,
      "grad_norm": 0.5558242201805115,
      "learning_rate": 9.773573795146485e-05,
      "loss": 0.0581,
      "step": 610
    },
    {
      "epoch": 0.9472880061115355,
      "grad_norm": 0.3050296902656555,
      "learning_rate": 9.757422152806415e-05,
      "loss": 0.0342,
      "step": 620
    },
    {
      "epoch": 0.9625668449197861,
      "grad_norm": 0.41481053829193115,
      "learning_rate": 9.74072840125383e-05,
      "loss": 0.0506,
      "step": 630
    },
    {
      "epoch": 0.9778456837280367,
      "grad_norm": 0.6876723170280457,
      "learning_rate": 9.723494442745085e-05,
      "loss": 0.0474,
      "step": 640
    },
    {
      "epoch": 0.9931245225362872,
      "grad_norm": 0.638076663017273,
      "learning_rate": 9.705722241093223e-05,
      "loss": 0.0625,
      "step": 650
    },
    {
      "epoch": 0.9931245225362872,
      "eval_loss": 0.042443711310625076,
      "eval_runtime": 191.6589,
      "eval_samples_per_second": 6.073,
      "eval_steps_per_second": 6.073,
      "step": 650
    },
    {
      "epoch": 1.0084033613445378,
      "grad_norm": 0.3265399634838104,
      "learning_rate": 9.687413821444199e-05,
      "loss": 0.0475,
      "step": 660
    },
    {
      "epoch": 1.0236822001527883,
      "grad_norm": 0.2271881401538849,
      "learning_rate": 9.668571270046122e-05,
      "loss": 0.04,
      "step": 670
    },
    {
      "epoch": 1.0389610389610389,
      "grad_norm": 0.20923583209514618,
      "learning_rate": 9.649196734011519e-05,
      "loss": 0.0353,
      "step": 680
    },
    {
      "epoch": 1.0542398777692896,
      "grad_norm": 0.9111831784248352,
      "learning_rate": 9.629292421072671e-05,
      "loss": 0.029,
      "step": 690
    },
    {
      "epoch": 1.0695187165775402,
      "grad_norm": 0.21794196963310242,
      "learning_rate": 9.608860599330048e-05,
      "loss": 0.0363,
      "step": 700
    },
    {
      "epoch": 1.0695187165775402,
      "eval_loss": 0.0368080735206604,
      "eval_runtime": 190.2458,
      "eval_samples_per_second": 6.118,
      "eval_steps_per_second": 6.118,
      "step": 700
    },
    {
      "epoch": 1.0847975553857907,
      "grad_norm": 0.1954701691865921,
      "learning_rate": 9.587903596993854e-05,
      "loss": 0.0337,
      "step": 710
    },
    {
      "epoch": 1.1000763941940412,
      "grad_norm": 0.3832463324069977,
      "learning_rate": 9.566423802118724e-05,
      "loss": 0.036,
      "step": 720
    },
    {
      "epoch": 1.1153552330022918,
      "grad_norm": 0.120274618268013,
      "learning_rate": 9.544423662331612e-05,
      "loss": 0.048,
      "step": 730
    },
    {
      "epoch": 1.1306340718105423,
      "grad_norm": 0.8174816966056824,
      "learning_rate": 9.521905684552877e-05,
      "loss": 0.0354,
      "step": 740
    },
    {
      "epoch": 1.1459129106187929,
      "grad_norm": 0.31460389494895935,
      "learning_rate": 9.498872434710623e-05,
      "loss": 0.024,
      "step": 750
    },
    {
      "epoch": 1.1459129106187929,
      "eval_loss": 0.037662845104932785,
      "eval_runtime": 188.5908,
      "eval_samples_per_second": 6.172,
      "eval_steps_per_second": 6.172,
      "step": 750
    },
    {
      "epoch": 1.1611917494270436,
      "grad_norm": 0.26857221126556396,
      "learning_rate": 9.475326537448307e-05,
      "loss": 0.0303,
      "step": 760
    },
    {
      "epoch": 1.1764705882352942,
      "grad_norm": 0.2545701861381531,
      "learning_rate": 9.451270675825665e-05,
      "loss": 0.0275,
      "step": 770
    },
    {
      "epoch": 1.1917494270435447,
      "grad_norm": 0.41096410155296326,
      "learning_rate": 9.426707591012976e-05,
      "loss": 0.0333,
      "step": 780
    },
    {
      "epoch": 1.2070282658517952,
      "grad_norm": 0.383953332901001,
      "learning_rate": 9.4016400819787e-05,
      "loss": 0.0413,
      "step": 790
    },
    {
      "epoch": 1.2223071046600458,
      "grad_norm": 0.6644073724746704,
      "learning_rate": 9.376071005170539e-05,
      "loss": 0.0331,
      "step": 800
    },
    {
      "epoch": 1.2223071046600458,
      "eval_loss": 0.037403445690870285,
      "eval_runtime": 187.4305,
      "eval_samples_per_second": 6.21,
      "eval_steps_per_second": 6.21,
      "step": 800
    },
    {
      "epoch": 1.2375859434682965,
      "grad_norm": 0.2898555099964142,
      "learning_rate": 9.350003274189949e-05,
      "loss": 0.0371,
      "step": 810
    },
    {
      "epoch": 1.2528647822765469,
      "grad_norm": 0.6009165644645691,
      "learning_rate": 9.323439859460122e-05,
      "loss": 0.0376,
      "step": 820
    },
    {
      "epoch": 1.2681436210847976,
      "grad_norm": 0.3177230954170227,
      "learning_rate": 9.296383787887519e-05,
      "loss": 0.0342,
      "step": 830
    },
    {
      "epoch": 1.2834224598930482,
      "grad_norm": 0.3252789080142975,
      "learning_rate": 9.268838142516943e-05,
      "loss": 0.0446,
      "step": 840
    },
    {
      "epoch": 1.2987012987012987,
      "grad_norm": 0.22259239852428436,
      "learning_rate": 9.240806062180234e-05,
      "loss": 0.0368,
      "step": 850
    },
    {
      "epoch": 1.2987012987012987,
      "eval_loss": 0.03829605504870415,
      "eval_runtime": 187.1071,
      "eval_samples_per_second": 6.221,
      "eval_steps_per_second": 6.221,
      "step": 850
    },
    {
      "epoch": 1.3139801375095492,
      "grad_norm": 0.17059855163097382,
      "learning_rate": 9.212290741138592e-05,
      "loss": 0.0184,
      "step": 860
    },
    {
      "epoch": 1.3292589763177998,
      "grad_norm": 0.18967171013355255,
      "learning_rate": 9.183295428718592e-05,
      "loss": 0.0318,
      "step": 870
    },
    {
      "epoch": 1.3445378151260505,
      "grad_norm": 0.3216451406478882,
      "learning_rate": 9.153823428941924e-05,
      "loss": 0.0403,
      "step": 880
    },
    {
      "epoch": 1.359816653934301,
      "grad_norm": 0.32410064339637756,
      "learning_rate": 9.1238781001489e-05,
      "loss": 0.0311,
      "step": 890
    },
    {
      "epoch": 1.3750954927425516,
      "grad_norm": 0.2728131115436554,
      "learning_rate": 9.093462854615766e-05,
      "loss": 0.039,
      "step": 900
    },
    {
      "epoch": 1.3750954927425516,
      "eval_loss": 0.03588823974132538,
      "eval_runtime": 186.8024,
      "eval_samples_per_second": 6.231,
      "eval_steps_per_second": 6.231,
      "step": 900
    },
    {
      "epoch": 1.3903743315508021,
      "grad_norm": 0.28432366251945496,
      "learning_rate": 9.062581158165876e-05,
      "loss": 0.0353,
      "step": 910
    },
    {
      "epoch": 1.4056531703590527,
      "grad_norm": 0.3692541718482971,
      "learning_rate": 9.031236529774764e-05,
      "loss": 0.0277,
      "step": 920
    },
    {
      "epoch": 1.4209320091673032,
      "grad_norm": 0.20645208656787872,
      "learning_rate": 8.999432541169145e-05,
      "loss": 0.035,
      "step": 930
    },
    {
      "epoch": 1.4362108479755538,
      "grad_norm": 0.28323686122894287,
      "learning_rate": 8.967172816419927e-05,
      "loss": 0.0319,
      "step": 940
    },
    {
      "epoch": 1.4514896867838045,
      "grad_norm": 0.3196757435798645,
      "learning_rate": 8.934461031529242e-05,
      "loss": 0.0428,
      "step": 950
    },
    {
      "epoch": 1.4514896867838045,
      "eval_loss": 0.03870726749300957,
      "eval_runtime": 188.2883,
      "eval_samples_per_second": 6.182,
      "eval_steps_per_second": 6.182,
      "step": 950
    },
    {
      "epoch": 1.466768525592055,
      "grad_norm": 0.42527279257774353,
      "learning_rate": 8.901300914011569e-05,
      "loss": 0.0339,
      "step": 960
    },
    {
      "epoch": 1.4820473644003056,
      "grad_norm": 0.12177124619483948,
      "learning_rate": 8.867696242468976e-05,
      "loss": 0.0247,
      "step": 970
    },
    {
      "epoch": 1.4973262032085561,
      "grad_norm": 0.33116525411605835,
      "learning_rate": 8.833650846160555e-05,
      "loss": 0.0413,
      "step": 980
    },
    {
      "epoch": 1.5126050420168067,
      "grad_norm": 0.33314356207847595,
      "learning_rate": 8.79916860456607e-05,
      "loss": 0.0292,
      "step": 990
    },
    {
      "epoch": 1.5278838808250574,
      "grad_norm": 0.23561441898345947,
      "learning_rate": 8.7642534469439e-05,
      "loss": 0.0285,
      "step": 1000
    },
    {
      "epoch": 1.5278838808250574,
      "eval_loss": 0.03498704358935356,
      "eval_runtime": 189.0743,
      "eval_samples_per_second": 6.156,
      "eval_steps_per_second": 6.156,
      "step": 1000
    },
    {
      "epoch": 1.5431627196333078,
      "grad_norm": 0.6858423352241516,
      "learning_rate": 8.728909351883283e-05,
      "loss": 0.0279,
      "step": 1010
    },
    {
      "epoch": 1.5584415584415585,
      "grad_norm": 0.43968966603279114,
      "learning_rate": 8.693140346850975e-05,
      "loss": 0.0401,
      "step": 1020
    },
    {
      "epoch": 1.573720397249809,
      "grad_norm": 0.24436284601688385,
      "learning_rate": 8.656950507732303e-05,
      "loss": 0.0275,
      "step": 1030
    },
    {
      "epoch": 1.5889992360580596,
      "grad_norm": 0.15837892889976501,
      "learning_rate": 8.620343958366718e-05,
      "loss": 0.0358,
      "step": 1040
    },
    {
      "epoch": 1.6042780748663101,
      "grad_norm": 0.40071073174476624,
      "learning_rate": 8.5833248700779e-05,
      "loss": 0.0292,
      "step": 1050
    },
    {
      "epoch": 1.6042780748663101,
      "eval_loss": 0.0367753803730011,
      "eval_runtime": 189.1042,
      "eval_samples_per_second": 6.155,
      "eval_steps_per_second": 6.155,
      "step": 1050
    },
    {
      "epoch": 1.6195569136745607,
      "grad_norm": 0.12202508002519608,
      "learning_rate": 8.545897461198413e-05,
      "loss": 0.0326,
      "step": 1060
    },
    {
      "epoch": 1.6348357524828114,
      "grad_norm": 0.18900205194950104,
      "learning_rate": 8.508065996589036e-05,
      "loss": 0.0285,
      "step": 1070
    },
    {
      "epoch": 1.6501145912910617,
      "grad_norm": 0.15240703523159027,
      "learning_rate": 8.469834787152783e-05,
      "loss": 0.0352,
      "step": 1080
    },
    {
      "epoch": 1.6653934300993125,
      "grad_norm": 0.28794217109680176,
      "learning_rate": 8.43120818934367e-05,
      "loss": 0.0306,
      "step": 1090
    },
    {
      "epoch": 1.680672268907563,
      "grad_norm": 0.4002177119255066,
      "learning_rate": 8.392190604670293e-05,
      "loss": 0.0317,
      "step": 1100
    },
    {
      "epoch": 1.680672268907563,
      "eval_loss": 0.036450888961553574,
      "eval_runtime": 189.1055,
      "eval_samples_per_second": 6.155,
      "eval_steps_per_second": 6.155,
      "step": 1100
    },
    {
      "epoch": 1.6959511077158136,
      "grad_norm": 0.2935492992401123,
      "learning_rate": 8.352786479194288e-05,
      "loss": 0.0383,
      "step": 1110
    },
    {
      "epoch": 1.7112299465240641,
      "grad_norm": 0.3728632926940918,
      "learning_rate": 8.313000303023688e-05,
      "loss": 0.0287,
      "step": 1120
    },
    {
      "epoch": 1.7265087853323147,
      "grad_norm": 0.2868274450302124,
      "learning_rate": 8.27283660980128e-05,
      "loss": 0.0369,
      "step": 1130
    },
    {
      "epoch": 1.7417876241405654,
      "grad_norm": 0.210665762424469,
      "learning_rate": 8.232299976187999e-05,
      "loss": 0.0361,
      "step": 1140
    },
    {
      "epoch": 1.7570664629488157,
      "grad_norm": 0.2601427733898163,
      "learning_rate": 8.191395021341408e-05,
      "loss": 0.0506,
      "step": 1150
    },
    {
      "epoch": 1.7570664629488157,
      "eval_loss": 0.03487667441368103,
      "eval_runtime": 188.1603,
      "eval_samples_per_second": 6.186,
      "eval_steps_per_second": 6.186,
      "step": 1150
    },
    {
      "epoch": 1.7723453017570665,
      "grad_norm": 0.34721285104751587,
      "learning_rate": 8.150126406389352e-05,
      "loss": 0.0303,
      "step": 1160
    },
    {
      "epoch": 1.787624140565317,
      "grad_norm": 0.3729540705680847,
      "learning_rate": 8.108498833898815e-05,
      "loss": 0.038,
      "step": 1170
    },
    {
      "epoch": 1.8029029793735676,
      "grad_norm": 0.3638748526573181,
      "learning_rate": 8.066517047340066e-05,
      "loss": 0.032,
      "step": 1180
    },
    {
      "epoch": 1.8181818181818183,
      "grad_norm": 0.14144481718540192,
      "learning_rate": 8.02418583054614e-05,
      "loss": 0.0198,
      "step": 1190
    },
    {
      "epoch": 1.8334606569900687,
      "grad_norm": 0.40133315324783325,
      "learning_rate": 7.981510007167719e-05,
      "loss": 0.0329,
      "step": 1200
    },
    {
      "epoch": 1.8334606569900687,
      "eval_loss": 0.035291481763124466,
      "eval_runtime": 187.8454,
      "eval_samples_per_second": 6.197,
      "eval_steps_per_second": 6.197,
      "step": 1200
    },
    {
      "epoch": 1.8487394957983194,
      "grad_norm": 0.3257274329662323,
      "learning_rate": 7.938494440123468e-05,
      "loss": 0.031,
      "step": 1210
    },
    {
      "epoch": 1.8640183346065697,
      "grad_norm": 0.16669751703739166,
      "learning_rate": 7.895144031045918e-05,
      "loss": 0.0349,
      "step": 1220
    },
    {
      "epoch": 1.8792971734148205,
      "grad_norm": 0.2631896138191223,
      "learning_rate": 7.851463719722913e-05,
      "loss": 0.0234,
      "step": 1230
    },
    {
      "epoch": 1.894576012223071,
      "grad_norm": 0.33011606335639954,
      "learning_rate": 7.80745848353473e-05,
      "loss": 0.0334,
      "step": 1240
    },
    {
      "epoch": 1.9098548510313216,
      "grad_norm": 0.27665475010871887,
      "learning_rate": 7.763133336886892e-05,
      "loss": 0.0352,
      "step": 1250
    },
    {
      "epoch": 1.9098548510313216,
      "eval_loss": 0.03766217082738876,
      "eval_runtime": 188.3048,
      "eval_samples_per_second": 6.181,
      "eval_steps_per_second": 6.181,
      "step": 1250
    },
    {
      "epoch": 1.9251336898395723,
      "grad_norm": 0.15627293288707733,
      "learning_rate": 7.718493330638789e-05,
      "loss": 0.0245,
      "step": 1260
    },
    {
      "epoch": 1.9404125286478227,
      "grad_norm": 0.11793619394302368,
      "learning_rate": 7.673543551528122e-05,
      "loss": 0.0391,
      "step": 1270
    },
    {
      "epoch": 1.9556913674560734,
      "grad_norm": 0.46485260128974915,
      "learning_rate": 7.628289121591277e-05,
      "loss": 0.0365,
      "step": 1280
    },
    {
      "epoch": 1.970970206264324,
      "grad_norm": 0.28952929377555847,
      "learning_rate": 7.582735197579656e-05,
      "loss": 0.0318,
      "step": 1290
    },
    {
      "epoch": 1.9862490450725745,
      "grad_norm": 0.24003995954990387,
      "learning_rate": 7.536886970372078e-05,
      "loss": 0.0289,
      "step": 1300
    },
    {
      "epoch": 1.9862490450725745,
      "eval_loss": 0.036530643701553345,
      "eval_runtime": 188.1847,
      "eval_samples_per_second": 6.185,
      "eval_steps_per_second": 6.185,
      "step": 1300
    },
    {
      "epoch": 2.0015278838808253,
      "grad_norm": 0.23869691789150238,
      "learning_rate": 7.490749664383271e-05,
      "loss": 0.0428,
      "step": 1310
    },
    {
      "epoch": 2.0168067226890756,
      "grad_norm": 0.18940101563930511,
      "learning_rate": 7.444328536968538e-05,
      "loss": 0.0265,
      "step": 1320
    },
    {
      "epoch": 2.0320855614973263,
      "grad_norm": 0.4082794487476349,
      "learning_rate": 7.397628877824701e-05,
      "loss": 0.0227,
      "step": 1330
    },
    {
      "epoch": 2.0473644003055766,
      "grad_norm": 0.13684681057929993,
      "learning_rate": 7.350656008387327e-05,
      "loss": 0.0201,
      "step": 1340
    },
    {
      "epoch": 2.0626432391138274,
      "grad_norm": 0.1932537704706192,
      "learning_rate": 7.303415281224346e-05,
      "loss": 0.0202,
      "step": 1350
    },
    {
      "epoch": 2.0626432391138274,
      "eval_loss": 0.03559866175055504,
      "eval_runtime": 188.1857,
      "eval_samples_per_second": 6.185,
      "eval_steps_per_second": 6.185,
      "step": 1350
    },
    {
      "epoch": 2.0779220779220777,
      "grad_norm": 0.1311611384153366,
      "learning_rate": 7.255912079426136e-05,
      "loss": 0.0164,
      "step": 1360
    },
    {
      "epoch": 2.0932009167303285,
      "grad_norm": 0.40545451641082764,
      "learning_rate": 7.208151815992107e-05,
      "loss": 0.0198,
      "step": 1370
    },
    {
      "epoch": 2.1084797555385792,
      "grad_norm": 0.45390045642852783,
      "learning_rate": 7.160139933213898e-05,
      "loss": 0.0272,
      "step": 1380
    },
    {
      "epoch": 2.1237585943468296,
      "grad_norm": 0.31190603971481323,
      "learning_rate": 7.111881902055223e-05,
      "loss": 0.0196,
      "step": 1390
    },
    {
      "epoch": 2.1390374331550803,
      "grad_norm": 0.49418479204177856,
      "learning_rate": 7.06338322152845e-05,
      "loss": 0.0174,
      "step": 1400
    },
    {
      "epoch": 2.1390374331550803,
      "eval_loss": 0.035685863345861435,
      "eval_runtime": 187.4939,
      "eval_samples_per_second": 6.208,
      "eval_steps_per_second": 6.208,
      "step": 1400
    },
    {
      "epoch": 2.1543162719633306,
      "grad_norm": 0.2847372889518738,
      "learning_rate": 7.014649418067994e-05,
      "loss": 0.0185,
      "step": 1410
    },
    {
      "epoch": 2.1695951107715814,
      "grad_norm": 0.16619278490543365,
      "learning_rate": 6.965686044900577e-05,
      "loss": 0.0149,
      "step": 1420
    },
    {
      "epoch": 2.184873949579832,
      "grad_norm": 0.44972145557403564,
      "learning_rate": 6.91649868141243e-05,
      "loss": 0.0149,
      "step": 1430
    },
    {
      "epoch": 2.2001527883880825,
      "grad_norm": 0.15861999988555908,
      "learning_rate": 6.86709293251353e-05,
      "loss": 0.0106,
      "step": 1440
    },
    {
      "epoch": 2.2154316271963332,
      "grad_norm": 0.5113412141799927,
      "learning_rate": 6.817474427998916e-05,
      "loss": 0.0134,
      "step": 1450
    },
    {
      "epoch": 2.2154316271963332,
      "eval_loss": 0.03945557773113251,
      "eval_runtime": 187.1894,
      "eval_samples_per_second": 6.218,
      "eval_steps_per_second": 6.218,
      "step": 1450
    },
    {
      "epoch": 2.2307104660045836,
      "grad_norm": 0.7107367515563965,
      "learning_rate": 6.767648821907172e-05,
      "loss": 0.0208,
      "step": 1460
    },
    {
      "epoch": 2.2459893048128343,
      "grad_norm": 0.2532961666584015,
      "learning_rate": 6.717621791876147e-05,
      "loss": 0.0317,
      "step": 1470
    },
    {
      "epoch": 2.2612681436210846,
      "grad_norm": 0.25595831871032715,
      "learning_rate": 6.667399038495986e-05,
      "loss": 0.0228,
      "step": 1480
    },
    {
      "epoch": 2.2765469824293354,
      "grad_norm": 0.2862846851348877,
      "learning_rate": 6.616986284659556e-05,
      "loss": 0.0238,
      "step": 1490
    },
    {
      "epoch": 2.2918258212375857,
      "grad_norm": 0.1290646344423294,
      "learning_rate": 6.566389274910309e-05,
      "loss": 0.02,
      "step": 1500
    },
    {
      "epoch": 2.2918258212375857,
      "eval_loss": 0.03613383322954178,
      "eval_runtime": 187.5145,
      "eval_samples_per_second": 6.208,
      "eval_steps_per_second": 6.208,
      "step": 1500
    },
    {
      "epoch": 2.3071046600458365,
      "grad_norm": 0.1208987608551979,
      "learning_rate": 6.515613774787697e-05,
      "loss": 0.0161,
      "step": 1510
    },
    {
      "epoch": 2.3223834988540872,
      "grad_norm": 0.14849679172039032,
      "learning_rate": 6.464665570170186e-05,
      "loss": 0.0226,
      "step": 1520
    },
    {
      "epoch": 2.3376623376623376,
      "grad_norm": 0.1842430830001831,
      "learning_rate": 6.413550466615952e-05,
      "loss": 0.0188,
      "step": 1530
    },
    {
      "epoch": 2.3529411764705883,
      "grad_norm": 0.22144414484500885,
      "learning_rate": 6.362274288701342e-05,
      "loss": 0.016,
      "step": 1540
    },
    {
      "epoch": 2.3682200152788386,
      "grad_norm": 0.2850957214832306,
      "learning_rate": 6.310842879357158e-05,
      "loss": 0.0189,
      "step": 1550
    },
    {
      "epoch": 2.3682200152788386,
      "eval_loss": 0.03735971450805664,
      "eval_runtime": 187.5259,
      "eval_samples_per_second": 6.207,
      "eval_steps_per_second": 6.207,
      "step": 1550
    },
    {
      "epoch": 2.3834988540870894,
      "grad_norm": 0.28386402130126953,
      "learning_rate": 6.25926209920285e-05,
      "loss": 0.0129,
      "step": 1560
    },
    {
      "epoch": 2.39877769289534,
      "grad_norm": 0.3930361866950989,
      "learning_rate": 6.207537825878708e-05,
      "loss": 0.0177,
      "step": 1570
    },
    {
      "epoch": 2.4140565317035905,
      "grad_norm": 0.1254359781742096,
      "learning_rate": 6.155675953376095e-05,
      "loss": 0.0278,
      "step": 1580
    },
    {
      "epoch": 2.4293353705118412,
      "grad_norm": 0.2683659493923187,
      "learning_rate": 6.103682391365828e-05,
      "loss": 0.0247,
      "step": 1590
    },
    {
      "epoch": 2.4446142093200915,
      "grad_norm": 0.20878136157989502,
      "learning_rate": 6.05156306452477e-05,
      "loss": 0.0162,
      "step": 1600
    },
    {
      "epoch": 2.4446142093200915,
      "eval_loss": 0.03479843959212303,
      "eval_runtime": 187.7453,
      "eval_samples_per_second": 6.2,
      "eval_steps_per_second": 6.2,
      "step": 1600
    },
    {
      "epoch": 2.4598930481283423,
      "grad_norm": 0.2738895118236542,
      "learning_rate": 5.9993239118607124e-05,
      "loss": 0.0255,
      "step": 1610
    },
    {
      "epoch": 2.475171886936593,
      "grad_norm": 0.23021462559700012,
      "learning_rate": 5.9469708860356246e-05,
      "loss": 0.0173,
      "step": 1620
    },
    {
      "epoch": 2.4904507257448434,
      "grad_norm": 0.23016996681690216,
      "learning_rate": 5.89450995268734e-05,
      "loss": 0.0198,
      "step": 1630
    },
    {
      "epoch": 2.5057295645530937,
      "grad_norm": 0.37153613567352295,
      "learning_rate": 5.841947089749783e-05,
      "loss": 0.0226,
      "step": 1640
    },
    {
      "epoch": 2.5210084033613445,
      "grad_norm": 0.3245033919811249,
      "learning_rate": 5.78928828677177e-05,
      "loss": 0.0252,
      "step": 1650
    },
    {
      "epoch": 2.5210084033613445,
      "eval_loss": 0.037063825875520706,
      "eval_runtime": 187.1008,
      "eval_samples_per_second": 6.221,
      "eval_steps_per_second": 6.221,
      "step": 1650
    },
    {
      "epoch": 2.5362872421695952,
      "grad_norm": 0.7465736865997314,
      "learning_rate": 5.7365395442345085e-05,
      "loss": 0.0186,
      "step": 1660
    },
    {
      "epoch": 2.5515660809778455,
      "grad_norm": 0.25831788778305054,
      "learning_rate": 5.683706872867833e-05,
      "loss": 0.0179,
      "step": 1670
    },
    {
      "epoch": 2.5668449197860963,
      "grad_norm": 0.32766786217689514,
      "learning_rate": 5.630796292965288e-05,
      "loss": 0.0271,
      "step": 1680
    },
    {
      "epoch": 2.5821237585943466,
      "grad_norm": 0.38363370299339294,
      "learning_rate": 5.57781383369811e-05,
      "loss": 0.0204,
      "step": 1690
    },
    {
      "epoch": 2.5974025974025974,
      "grad_norm": 0.14547067880630493,
      "learning_rate": 5.524765532428203e-05,
      "loss": 0.0175,
      "step": 1700
    },
    {
      "epoch": 2.5974025974025974,
      "eval_loss": 0.03661783039569855,
      "eval_runtime": 186.6526,
      "eval_samples_per_second": 6.236,
      "eval_steps_per_second": 6.236,
      "step": 1700
    },
    {
      "epoch": 2.612681436210848,
      "grad_norm": 0.29803889989852905,
      "learning_rate": 5.471657434020182e-05,
      "loss": 0.0212,
      "step": 1710
    },
    {
      "epoch": 2.6279602750190985,
      "grad_norm": 0.31547999382019043,
      "learning_rate": 5.418495590152557e-05,
      "loss": 0.0233,
      "step": 1720
    },
    {
      "epoch": 2.643239113827349,
      "grad_norm": 0.19628596305847168,
      "learning_rate": 5.365286058628145e-05,
      "loss": 0.0236,
      "step": 1730
    },
    {
      "epoch": 2.6585179526355995,
      "grad_norm": 0.3903220593929291,
      "learning_rate": 5.312034902683779e-05,
      "loss": 0.0264,
      "step": 1740
    },
    {
      "epoch": 2.6737967914438503,
      "grad_norm": 0.42966052889823914,
      "learning_rate": 5.258748190299404e-05,
      "loss": 0.0222,
      "step": 1750
    },
    {
      "epoch": 2.6737967914438503,
      "eval_loss": 0.03461781516671181,
      "eval_runtime": 184.1896,
      "eval_samples_per_second": 6.32,
      "eval_steps_per_second": 6.32,
      "step": 1750
    },
    {
      "epoch": 2.689075630252101,
      "grad_norm": 0.2390514761209488,
      "learning_rate": 5.20543199350663e-05,
      "loss": 0.0203,
      "step": 1760
    },
    {
      "epoch": 2.7043544690603514,
      "grad_norm": 0.17239676415920258,
      "learning_rate": 5.152092387696821e-05,
      "loss": 0.0235,
      "step": 1770
    },
    {
      "epoch": 2.719633307868602,
      "grad_norm": 0.29174742102622986,
      "learning_rate": 5.0987354509287985e-05,
      "loss": 0.0189,
      "step": 1780
    },
    {
      "epoch": 2.7349121466768525,
      "grad_norm": 0.35961881279945374,
      "learning_rate": 5.045367263236257e-05,
      "loss": 0.016,
      "step": 1790
    },
    {
      "epoch": 2.750190985485103,
      "grad_norm": 0.15872785449028015,
      "learning_rate": 4.991993905934931e-05,
      "loss": 0.0274,
      "step": 1800
    },
    {
      "epoch": 2.750190985485103,
      "eval_loss": 0.03465921804308891,
      "eval_runtime": 185.6419,
      "eval_samples_per_second": 6.27,
      "eval_steps_per_second": 6.27,
      "step": 1800
    },
    {
      "epoch": 2.765469824293354,
      "grad_norm": 0.28768208622932434,
      "learning_rate": 4.938621460929639e-05,
      "loss": 0.0187,
      "step": 1810
    },
    {
      "epoch": 2.7807486631016043,
      "grad_norm": 0.24558648467063904,
      "learning_rate": 4.885256010021233e-05,
      "loss": 0.0185,
      "step": 1820
    },
    {
      "epoch": 2.7960275019098546,
      "grad_norm": 0.27898794412612915,
      "learning_rate": 4.831903634213599e-05,
      "loss": 0.0193,
      "step": 1830
    },
    {
      "epoch": 2.8113063407181054,
      "grad_norm": 0.06646784394979477,
      "learning_rate": 4.778570413020702e-05,
      "loss": 0.021,
      "step": 1840
    },
    {
      "epoch": 2.826585179526356,
      "grad_norm": 0.43815556168556213,
      "learning_rate": 4.725262423773838e-05,
      "loss": 0.0215,
      "step": 1850
    },
    {
      "epoch": 2.826585179526356,
      "eval_loss": 0.03624486178159714,
      "eval_runtime": 185.8089,
      "eval_samples_per_second": 6.265,
      "eval_steps_per_second": 6.265,
      "step": 1850
    },
    {
      "epoch": 2.8418640183346064,
      "grad_norm": 0.22074271738529205,
      "learning_rate": 4.671985740929123e-05,
      "loss": 0.0138,
      "step": 1860
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 0.05057336017489433,
      "learning_rate": 4.618746435375295e-05,
      "loss": 0.0212,
      "step": 1870
    },
    {
      "epoch": 2.8724216959511075,
      "grad_norm": 0.21193920075893402,
      "learning_rate": 4.565550573741942e-05,
      "loss": 0.0175,
      "step": 1880
    },
    {
      "epoch": 2.8877005347593583,
      "grad_norm": 0.5380321741104126,
      "learning_rate": 4.512404217708217e-05,
      "loss": 0.0242,
      "step": 1890
    },
    {
      "epoch": 2.902979373567609,
      "grad_norm": 0.33445772528648376,
      "learning_rate": 4.45931342331209e-05,
      "loss": 0.0201,
      "step": 1900
    },
    {
      "epoch": 2.902979373567609,
      "eval_loss": 0.03776334598660469,
      "eval_runtime": 185.4273,
      "eval_samples_per_second": 6.277,
      "eval_steps_per_second": 6.277,
      "step": 1900
    },
    {
      "epoch": 2.9182582123758594,
      "grad_norm": 0.5133573412895203,
      "learning_rate": 4.406284240260278e-05,
      "loss": 0.0282,
      "step": 1910
    },
    {
      "epoch": 2.93353705118411,
      "grad_norm": 0.1490231305360794,
      "learning_rate": 4.3533227112388694e-05,
      "loss": 0.0184,
      "step": 1920
    },
    {
      "epoch": 2.9488158899923604,
      "grad_norm": 0.25799980759620667,
      "learning_rate": 4.300434871224763e-05,
      "loss": 0.0256,
      "step": 1930
    },
    {
      "epoch": 2.964094728800611,
      "grad_norm": 0.3950966000556946,
      "learning_rate": 4.247626746797983e-05,
      "loss": 0.0231,
      "step": 1940
    },
    {
      "epoch": 2.979373567608862,
      "grad_norm": 0.24436689913272858,
      "learning_rate": 4.1949043554549406e-05,
      "loss": 0.016,
      "step": 1950
    },
    {
      "epoch": 2.979373567608862,
      "eval_loss": 0.034271180629730225,
      "eval_runtime": 185.359,
      "eval_samples_per_second": 6.28,
      "eval_steps_per_second": 6.28,
      "step": 1950
    },
    {
      "epoch": 2.9946524064171123,
      "grad_norm": 0.43147724866867065,
      "learning_rate": 4.14227370492275e-05,
      "loss": 0.0254,
      "step": 1960
    },
    {
      "epoch": 3.009931245225363,
      "grad_norm": 0.0808895006775856,
      "learning_rate": 4.08974079247464e-05,
      "loss": 0.0153,
      "step": 1970
    },
    {
      "epoch": 3.0252100840336134,
      "grad_norm": 0.024843959137797356,
      "learning_rate": 4.037311604246565e-05,
      "loss": 0.0091,
      "step": 1980
    },
    {
      "epoch": 3.040488922841864,
      "grad_norm": 0.2611240744590759,
      "learning_rate": 3.9849921145550805e-05,
      "loss": 0.01,
      "step": 1990
    },
    {
      "epoch": 3.0557677616501144,
      "grad_norm": 0.09150966256856918,
      "learning_rate": 3.9327882852165795e-05,
      "loss": 0.009,
      "step": 2000
    },
    {
      "epoch": 3.0557677616501144,
      "eval_loss": 0.03721309453248978,
      "eval_runtime": 185.6193,
      "eval_samples_per_second": 6.271,
      "eval_steps_per_second": 6.271,
      "step": 2000
    },
    {
      "epoch": 3.071046600458365,
      "grad_norm": 0.10902375727891922,
      "learning_rate": 3.880706064867926e-05,
      "loss": 0.0094,
      "step": 2010
    },
    {
      "epoch": 3.0863254392666155,
      "grad_norm": 0.3028370440006256,
      "learning_rate": 3.8287513882886196e-05,
      "loss": 0.0062,
      "step": 2020
    },
    {
      "epoch": 3.1016042780748663,
      "grad_norm": 0.11812803894281387,
      "learning_rate": 3.776930175724521e-05,
      "loss": 0.0102,
      "step": 2030
    },
    {
      "epoch": 3.116883116883117,
      "grad_norm": 0.21782122552394867,
      "learning_rate": 3.7252483322132386e-05,
      "loss": 0.0129,
      "step": 2040
    },
    {
      "epoch": 3.1321619556913673,
      "grad_norm": 0.09384758770465851,
      "learning_rate": 3.673711746911252e-05,
      "loss": 0.0106,
      "step": 2050
    },
    {
      "epoch": 3.1321619556913673,
      "eval_loss": 0.03892069309949875,
      "eval_runtime": 185.889,
      "eval_samples_per_second": 6.262,
      "eval_steps_per_second": 6.262,
      "step": 2050
    },
    {
      "epoch": 3.147440794499618,
      "grad_norm": 0.16040179133415222,
      "learning_rate": 3.6223262924228344e-05,
      "loss": 0.0053,
      "step": 2060
    },
    {
      "epoch": 3.1627196333078684,
      "grad_norm": 0.04083335027098656,
      "learning_rate": 3.5710978241308733e-05,
      "loss": 0.0097,
      "step": 2070
    },
    {
      "epoch": 3.177998472116119,
      "grad_norm": 0.311753511428833,
      "learning_rate": 3.520032179529652e-05,
      "loss": 0.0101,
      "step": 2080
    },
    {
      "epoch": 3.19327731092437,
      "grad_norm": 0.36532482504844666,
      "learning_rate": 3.4691351775596564e-05,
      "loss": 0.0086,
      "step": 2090
    },
    {
      "epoch": 3.2085561497326203,
      "grad_norm": 0.25262030959129333,
      "learning_rate": 3.41841261794451e-05,
      "loss": 0.0061,
      "step": 2100
    },
    {
      "epoch": 3.2085561497326203,
      "eval_loss": 0.0431986041367054,
      "eval_runtime": 186.9913,
      "eval_samples_per_second": 6.225,
      "eval_steps_per_second": 6.225,
      "step": 2100
    },
    {
      "epoch": 3.223834988540871,
      "grad_norm": 0.12558381259441376,
      "learning_rate": 3.367870280530101e-05,
      "loss": 0.0124,
      "step": 2110
    },
    {
      "epoch": 3.2391138273491213,
      "grad_norm": 0.3111749589443207,
      "learning_rate": 3.3175139246259536e-05,
      "loss": 0.0084,
      "step": 2120
    },
    {
      "epoch": 3.254392666157372,
      "grad_norm": 0.04653264582157135,
      "learning_rate": 3.2673492883489696e-05,
      "loss": 0.0061,
      "step": 2130
    },
    {
      "epoch": 3.2696715049656224,
      "grad_norm": 0.6107545495033264,
      "learning_rate": 3.2173820879695535e-05,
      "loss": 0.0102,
      "step": 2140
    },
    {
      "epoch": 3.284950343773873,
      "grad_norm": 0.4089570641517639,
      "learning_rate": 3.1676180172602525e-05,
      "loss": 0.0075,
      "step": 2150
    },
    {
      "epoch": 3.284950343773873,
      "eval_loss": 0.04343220591545105,
      "eval_runtime": 188.264,
      "eval_samples_per_second": 6.183,
      "eval_steps_per_second": 6.183,
      "step": 2150
    },
    {
      "epoch": 3.300229182582124,
      "grad_norm": 0.12825438380241394,
      "learning_rate": 3.11806274684695e-05,
      "loss": 0.0081,
      "step": 2160
    },
    {
      "epoch": 3.3155080213903743,
      "grad_norm": 0.13207189738750458,
      "learning_rate": 3.068721923562688e-05,
      "loss": 0.0044,
      "step": 2170
    },
    {
      "epoch": 3.330786860198625,
      "grad_norm": 0.1463756561279297,
      "learning_rate": 3.019601169804216e-05,
      "loss": 0.0133,
      "step": 2180
    },
    {
      "epoch": 3.3460656990068753,
      "grad_norm": 0.39609789848327637,
      "learning_rate": 2.9707060828913225e-05,
      "loss": 0.0067,
      "step": 2190
    },
    {
      "epoch": 3.361344537815126,
      "grad_norm": 0.35121870040893555,
      "learning_rate": 2.9220422344290056e-05,
      "loss": 0.0089,
      "step": 2200
    },
    {
      "epoch": 3.361344537815126,
      "eval_loss": 0.04335245117545128,
      "eval_runtime": 189.2873,
      "eval_samples_per_second": 6.149,
      "eval_steps_per_second": 6.149,
      "step": 2200
    },
    {
      "epoch": 3.3766233766233764,
      "grad_norm": 0.3331177532672882,
      "learning_rate": 2.873615169672601e-05,
      "loss": 0.0056,
      "step": 2210
    },
    {
      "epoch": 3.391902215431627,
      "grad_norm": 0.21099725365638733,
      "learning_rate": 2.8254304068958927e-05,
      "loss": 0.0081,
      "step": 2220
    },
    {
      "epoch": 3.407181054239878,
      "grad_norm": 0.0633828192949295,
      "learning_rate": 2.7774934367622996e-05,
      "loss": 0.0106,
      "step": 2230
    },
    {
      "epoch": 3.4224598930481283,
      "grad_norm": 0.3705494999885559,
      "learning_rate": 2.7298097216992284e-05,
      "loss": 0.0072,
      "step": 2240
    },
    {
      "epoch": 3.437738731856379,
      "grad_norm": 0.2332771122455597,
      "learning_rate": 2.6823846952756125e-05,
      "loss": 0.0102,
      "step": 2250
    },
    {
      "epoch": 3.437738731856379,
      "eval_loss": 0.04621849209070206,
      "eval_runtime": 190.621,
      "eval_samples_per_second": 6.106,
      "eval_steps_per_second": 6.106,
      "step": 2250
    },
    {
      "epoch": 3.4530175706646293,
      "grad_norm": 0.26837480068206787,
      "learning_rate": 2.6352237615827636e-05,
      "loss": 0.005,
      "step": 2260
    },
    {
      "epoch": 3.46829640947288,
      "grad_norm": 0.20386861264705658,
      "learning_rate": 2.5883322946185777e-05,
      "loss": 0.0079,
      "step": 2270
    },
    {
      "epoch": 3.483575248281131,
      "grad_norm": 0.6309589743614197,
      "learning_rate": 2.5417156376751562e-05,
      "loss": 0.0146,
      "step": 2280
    },
    {
      "epoch": 3.498854087089381,
      "grad_norm": 0.10873563587665558,
      "learning_rate": 2.4953791027299506e-05,
      "loss": 0.0042,
      "step": 2290
    },
    {
      "epoch": 3.514132925897632,
      "grad_norm": 0.11322731524705887,
      "learning_rate": 2.4493279698404493e-05,
      "loss": 0.0083,
      "step": 2300
    },
    {
      "epoch": 3.514132925897632,
      "eval_loss": 0.04650338739156723,
      "eval_runtime": 192.2547,
      "eval_samples_per_second": 6.054,
      "eval_steps_per_second": 6.054,
      "step": 2300
    },
    {
      "epoch": 3.5294117647058822,
      "grad_norm": 0.12000257521867752,
      "learning_rate": 2.403567486542518e-05,
      "loss": 0.0083,
      "step": 2310
    },
    {
      "epoch": 3.544690603514133,
      "grad_norm": 0.27765700221061707,
      "learning_rate": 2.3581028672524485e-05,
      "loss": 0.0096,
      "step": 2320
    },
    {
      "epoch": 3.5599694423223838,
      "grad_norm": 0.1320173442363739,
      "learning_rate": 2.312939292672765e-05,
      "loss": 0.0069,
      "step": 2330
    },
    {
      "epoch": 3.575248281130634,
      "grad_norm": 0.26741477847099304,
      "learning_rate": 2.268081909201885e-05,
      "loss": 0.0039,
      "step": 2340
    },
    {
      "epoch": 3.5905271199388844,
      "grad_norm": 0.016012145206332207,
      "learning_rate": 2.2235358283476936e-05,
      "loss": 0.0131,
      "step": 2350
    },
    {
      "epoch": 3.5905271199388844,
      "eval_loss": 0.0442974716424942,
      "eval_runtime": 193.688,
      "eval_samples_per_second": 6.01,
      "eval_steps_per_second": 6.01,
      "step": 2350
    },
    {
      "epoch": 3.605805958747135,
      "grad_norm": 0.04935047775506973,
      "learning_rate": 2.179306126145075e-05,
      "loss": 0.0117,
      "step": 2360
    },
    {
      "epoch": 3.621084797555386,
      "grad_norm": 0.3039514124393463,
      "learning_rate": 2.1353978425775008e-05,
      "loss": 0.0081,
      "step": 2370
    },
    {
      "epoch": 3.6363636363636362,
      "grad_norm": 0.23148193955421448,
      "learning_rate": 2.091815981002731e-05,
      "loss": 0.0081,
      "step": 2380
    },
    {
      "epoch": 3.651642475171887,
      "grad_norm": 0.3215958774089813,
      "learning_rate": 2.0485655075826667e-05,
      "loss": 0.0116,
      "step": 2390
    },
    {
      "epoch": 3.6669213139801373,
      "grad_norm": 0.5242221355438232,
      "learning_rate": 2.0056513507174685e-05,
      "loss": 0.0054,
      "step": 2400
    },
    {
      "epoch": 3.6669213139801373,
      "eval_loss": 0.04235563799738884,
      "eval_runtime": 194.4097,
      "eval_samples_per_second": 5.987,
      "eval_steps_per_second": 5.987,
      "step": 2400
    },
    {
      "epoch": 3.682200152788388,
      "grad_norm": 0.20493106544017792,
      "learning_rate": 1.963078400483953e-05,
      "loss": 0.0115,
      "step": 2410
    },
    {
      "epoch": 3.697478991596639,
      "grad_norm": 0.5069310069084167,
      "learning_rate": 1.9208515080783723e-05,
      "loss": 0.0084,
      "step": 2420
    },
    {
      "epoch": 3.712757830404889,
      "grad_norm": 0.2113790065050125,
      "learning_rate": 1.8789754852636245e-05,
      "loss": 0.0043,
      "step": 2430
    },
    {
      "epoch": 3.72803666921314,
      "grad_norm": 0.04292771592736244,
      "learning_rate": 1.837455103820942e-05,
      "loss": 0.014,
      "step": 2440
    },
    {
      "epoch": 3.7433155080213902,
      "grad_norm": 0.11845701187849045,
      "learning_rate": 1.7962950950061502e-05,
      "loss": 0.0038,
      "step": 2450
    },
    {
      "epoch": 3.7433155080213902,
      "eval_loss": 0.04276303946971893,
      "eval_runtime": 194.915,
      "eval_samples_per_second": 5.972,
      "eval_steps_per_second": 5.972,
      "step": 2450
    },
    {
      "epoch": 3.758594346829641,
      "grad_norm": 0.08765993267297745,
      "learning_rate": 1.7555001490105488e-05,
      "loss": 0.0072,
      "step": 2460
    },
    {
      "epoch": 3.7738731856378918,
      "grad_norm": 0.17220765352249146,
      "learning_rate": 1.7150749144264462e-05,
      "loss": 0.0088,
      "step": 2470
    },
    {
      "epoch": 3.789152024446142,
      "grad_norm": 0.1879175752401352,
      "learning_rate": 1.6750239977174682e-05,
      "loss": 0.0067,
      "step": 2480
    },
    {
      "epoch": 3.8044308632543924,
      "grad_norm": 0.38570868968963623,
      "learning_rate": 1.6353519626936397e-05,
      "loss": 0.012,
      "step": 2490
    },
    {
      "epoch": 3.819709702062643,
      "grad_norm": 0.08858808875083923,
      "learning_rate": 1.596063329991341e-05,
      "loss": 0.0074,
      "step": 2500
    },
    {
      "epoch": 3.819709702062643,
      "eval_loss": 0.04288734495639801,
      "eval_runtime": 196.135,
      "eval_samples_per_second": 5.935,
      "eval_steps_per_second": 5.935,
      "step": 2500
    },
    {
      "epoch": 3.834988540870894,
      "grad_norm": 0.09536246210336685,
      "learning_rate": 1.5571625765581832e-05,
      "loss": 0.0071,
      "step": 2510
    },
    {
      "epoch": 3.8502673796791442,
      "grad_norm": 0.1967010349035263,
      "learning_rate": 1.5186541351428545e-05,
      "loss": 0.0085,
      "step": 2520
    },
    {
      "epoch": 3.865546218487395,
      "grad_norm": 0.7389079332351685,
      "learning_rate": 1.4805423937900087e-05,
      "loss": 0.0111,
      "step": 2530
    },
    {
      "epoch": 3.8808250572956453,
      "grad_norm": 0.21783044934272766,
      "learning_rate": 1.4428316953402526e-05,
      "loss": 0.0074,
      "step": 2540
    },
    {
      "epoch": 3.896103896103896,
      "grad_norm": 0.18299157917499542,
      "learning_rate": 1.4055263369352672e-05,
      "loss": 0.0056,
      "step": 2550
    },
    {
      "epoch": 3.896103896103896,
      "eval_loss": 0.04263545572757721,
      "eval_runtime": 197.0054,
      "eval_samples_per_second": 5.908,
      "eval_steps_per_second": 5.908,
      "step": 2550
    },
    {
      "epoch": 3.911382734912147,
      "grad_norm": 0.2797110378742218,
      "learning_rate": 1.3686305695281559e-05,
      "loss": 0.0119,
      "step": 2560
    },
    {
      "epoch": 3.926661573720397,
      "grad_norm": 0.4075024724006653,
      "learning_rate": 1.3321485973990494e-05,
      "loss": 0.0072,
      "step": 2570
    },
    {
      "epoch": 3.941940412528648,
      "grad_norm": 0.21931466460227966,
      "learning_rate": 1.2960845776760156e-05,
      "loss": 0.0137,
      "step": 2580
    },
    {
      "epoch": 3.9572192513368982,
      "grad_norm": 0.13403423130512238,
      "learning_rate": 1.2604426198613688e-05,
      "loss": 0.0057,
      "step": 2590
    },
    {
      "epoch": 3.972498090145149,
      "grad_norm": 0.42918461561203003,
      "learning_rate": 1.2252267853633798e-05,
      "loss": 0.007,
      "step": 2600
    },
    {
      "epoch": 3.972498090145149,
      "eval_loss": 0.042820464819669724,
      "eval_runtime": 197.5342,
      "eval_samples_per_second": 5.893,
      "eval_steps_per_second": 5.893,
      "step": 2600
    },
    {
      "epoch": 3.9877769289533997,
      "grad_norm": 0.05887259915471077,
      "learning_rate": 1.1904410870334803e-05,
      "loss": 0.0122,
      "step": 2610
    },
    {
      "epoch": 4.0030557677616505,
      "grad_norm": 0.09298639744520187,
      "learning_rate": 1.1560894887090052e-05,
      "loss": 0.01,
      "step": 2620
    },
    {
      "epoch": 4.0183346065699,
      "grad_norm": 0.38059186935424805,
      "learning_rate": 1.1221759047615004e-05,
      "loss": 0.0032,
      "step": 2630
    },
    {
      "epoch": 4.033613445378151,
      "grad_norm": 0.04407782480120659,
      "learning_rate": 1.0887041996506859e-05,
      "loss": 0.0018,
      "step": 2640
    },
    {
      "epoch": 4.048892284186402,
      "grad_norm": 0.03980774059891701,
      "learning_rate": 1.0556781874841027e-05,
      "loss": 0.0034,
      "step": 2650
    },
    {
      "epoch": 4.048892284186402,
      "eval_loss": 0.043419938534498215,
      "eval_runtime": 197.6882,
      "eval_samples_per_second": 5.888,
      "eval_steps_per_second": 5.888,
      "step": 2650
    },
    {
      "epoch": 4.064171122994653,
      "grad_norm": 0.21608670055866241,
      "learning_rate": 1.0231016315824875e-05,
      "loss": 0.0042,
      "step": 2660
    },
    {
      "epoch": 4.0794499618029025,
      "grad_norm": 0.03518800064921379,
      "learning_rate": 9.909782440509491e-06,
      "loss": 0.004,
      "step": 2670
    },
    {
      "epoch": 4.094728800611153,
      "grad_norm": 0.46719732880592346,
      "learning_rate": 9.593116853559648e-06,
      "loss": 0.0048,
      "step": 2680
    },
    {
      "epoch": 4.110007639419404,
      "grad_norm": 0.05882187932729721,
      "learning_rate": 9.281055639082747e-06,
      "loss": 0.0048,
      "step": 2690
    },
    {
      "epoch": 4.125286478227655,
      "grad_norm": 0.0705062597990036,
      "learning_rate": 8.973634356517063e-06,
      "loss": 0.0051,
      "step": 2700
    },
    {
      "epoch": 4.125286478227655,
      "eval_loss": 0.04598911106586456,
      "eval_runtime": 198.719,
      "eval_samples_per_second": 5.858,
      "eval_steps_per_second": 5.858,
      "step": 2700
    },
    {
      "epoch": 4.140565317035906,
      "grad_norm": 0.21090030670166016,
      "learning_rate": 8.670888036579639e-06,
      "loss": 0.0043,
      "step": 2710
    },
    {
      "epoch": 4.1558441558441555,
      "grad_norm": 0.11752012372016907,
      "learning_rate": 8.372851177274604e-06,
      "loss": 0.0045,
      "step": 2720
    },
    {
      "epoch": 4.171122994652406,
      "grad_norm": 0.18993452191352844,
      "learning_rate": 8.079557739962128e-06,
      "loss": 0.0065,
      "step": 2730
    },
    {
      "epoch": 4.186401833460657,
      "grad_norm": 0.07488386332988739,
      "learning_rate": 7.791041145488454e-06,
      "loss": 0.0033,
      "step": 2740
    },
    {
      "epoch": 4.201680672268908,
      "grad_norm": 0.06048087030649185,
      "learning_rate": 7.507334270377619e-06,
      "loss": 0.0043,
      "step": 2750
    },
    {
      "epoch": 4.201680672268908,
      "eval_loss": 0.046420708298683167,
      "eval_runtime": 199.5972,
      "eval_samples_per_second": 5.832,
      "eval_steps_per_second": 5.832,
      "step": 2750
    },
    {
      "epoch": 4.2169595110771585,
      "grad_norm": 0.08631011098623276,
      "learning_rate": 7.228469443085206e-06,
      "loss": 0.0029,
      "step": 2760
    },
    {
      "epoch": 4.232238349885408,
      "grad_norm": 0.43059056997299194,
      "learning_rate": 6.954478440314427e-06,
      "loss": 0.0059,
      "step": 2770
    },
    {
      "epoch": 4.247517188693659,
      "grad_norm": 0.026996580883860588,
      "learning_rate": 6.685392483395259e-06,
      "loss": 0.0018,
      "step": 2780
    },
    {
      "epoch": 4.26279602750191,
      "grad_norm": 0.3803030252456665,
      "learning_rate": 6.421242234726682e-06,
      "loss": 0.0036,
      "step": 2790
    },
    {
      "epoch": 4.278074866310161,
      "grad_norm": 0.32429173588752747,
      "learning_rate": 6.1620577942827166e-06,
      "loss": 0.0023,
      "step": 2800
    },
    {
      "epoch": 4.278074866310161,
      "eval_loss": 0.04719507321715355,
      "eval_runtime": 199.3439,
      "eval_samples_per_second": 5.839,
      "eval_steps_per_second": 5.839,
      "step": 2800
    },
    {
      "epoch": 4.293353705118411,
      "grad_norm": 0.23215115070343018,
      "learning_rate": 5.907868696182584e-06,
      "loss": 0.0042,
      "step": 2810
    },
    {
      "epoch": 4.308632543926661,
      "grad_norm": 0.020175212994217873,
      "learning_rate": 5.658703905325186e-06,
      "loss": 0.0038,
      "step": 2820
    },
    {
      "epoch": 4.323911382734912,
      "grad_norm": 0.16418121755123138,
      "learning_rate": 5.414591814088627e-06,
      "loss": 0.0047,
      "step": 2830
    },
    {
      "epoch": 4.339190221543163,
      "grad_norm": 0.04667774215340614,
      "learning_rate": 5.17556023909489e-06,
      "loss": 0.0035,
      "step": 2840
    },
    {
      "epoch": 4.354469060351414,
      "grad_norm": 0.21239246428012848,
      "learning_rate": 4.941636418040058e-06,
      "loss": 0.0035,
      "step": 2850
    },
    {
      "epoch": 4.354469060351414,
      "eval_loss": 0.047712646424770355,
      "eval_runtime": 199.277,
      "eval_samples_per_second": 5.841,
      "eval_steps_per_second": 5.841,
      "step": 2850
    },
    {
      "epoch": 4.369747899159664,
      "grad_norm": 0.2187272608280182,
      "learning_rate": 4.7128470065906925e-06,
      "loss": 0.0033,
      "step": 2860
    },
    {
      "epoch": 4.385026737967914,
      "grad_norm": 0.1950971633195877,
      "learning_rate": 4.4892180753462744e-06,
      "loss": 0.0028,
      "step": 2870
    },
    {
      "epoch": 4.400305576776165,
      "grad_norm": 0.13193565607070923,
      "learning_rate": 4.270775106868586e-06,
      "loss": 0.0028,
      "step": 2880
    },
    {
      "epoch": 4.415584415584416,
      "grad_norm": 0.05055893212556839,
      "learning_rate": 4.057542992777868e-06,
      "loss": 0.0019,
      "step": 2890
    },
    {
      "epoch": 4.4308632543926665,
      "grad_norm": 0.12795984745025635,
      "learning_rate": 3.849546030916473e-06,
      "loss": 0.0021,
      "step": 2900
    },
    {
      "epoch": 4.4308632543926665,
      "eval_loss": 0.04882192611694336,
      "eval_runtime": 200.4256,
      "eval_samples_per_second": 5.808,
      "eval_steps_per_second": 5.808,
      "step": 2900
    },
    {
      "epoch": 4.446142093200916,
      "grad_norm": 0.13322241604328156,
      "learning_rate": 3.646807922580098e-06,
      "loss": 0.0017,
      "step": 2910
    },
    {
      "epoch": 4.461420932009167,
      "grad_norm": 0.1583353877067566,
      "learning_rate": 3.4493517698170164e-06,
      "loss": 0.0051,
      "step": 2920
    },
    {
      "epoch": 4.476699770817418,
      "grad_norm": 0.0266915000975132,
      "learning_rate": 3.2572000727956186e-06,
      "loss": 0.004,
      "step": 2930
    },
    {
      "epoch": 4.491978609625669,
      "grad_norm": 0.41155147552490234,
      "learning_rate": 3.070374727240466e-06,
      "loss": 0.0019,
      "step": 2940
    },
    {
      "epoch": 4.507257448433919,
      "grad_norm": 0.09082096815109253,
      "learning_rate": 2.8888970219373314e-06,
      "loss": 0.0021,
      "step": 2950
    },
    {
      "epoch": 4.507257448433919,
      "eval_loss": 0.0496671088039875,
      "eval_runtime": 200.9403,
      "eval_samples_per_second": 5.793,
      "eval_steps_per_second": 5.793,
      "step": 2950
    },
    {
      "epoch": 4.522536287242169,
      "grad_norm": 0.03273924067616463,
      "learning_rate": 2.7127876363072736e-06,
      "loss": 0.0018,
      "step": 2960
    },
    {
      "epoch": 4.53781512605042,
      "grad_norm": 0.3910587430000305,
      "learning_rate": 2.54206663805025e-06,
      "loss": 0.0026,
      "step": 2970
    },
    {
      "epoch": 4.553093964858671,
      "grad_norm": 0.07621679455041885,
      "learning_rate": 2.3767534808584125e-06,
      "loss": 0.001,
      "step": 2980
    },
    {
      "epoch": 4.5683728036669216,
      "grad_norm": 0.0307341106235981,
      "learning_rate": 2.2168670021993075e-06,
      "loss": 0.0023,
      "step": 2990
    },
    {
      "epoch": 4.583651642475171,
      "grad_norm": 0.25612154603004456,
      "learning_rate": 2.0624254211693894e-06,
      "loss": 0.0024,
      "step": 3000
    },
    {
      "epoch": 4.583651642475171,
      "eval_loss": 0.05014451965689659,
      "eval_runtime": 200.6569,
      "eval_samples_per_second": 5.801,
      "eval_steps_per_second": 5.801,
      "step": 3000
    },
    {
      "epoch": 4.598930481283422,
      "grad_norm": 0.05506656691431999,
      "learning_rate": 1.9134463364179177e-06,
      "loss": 0.0031,
      "step": 3010
    },
    {
      "epoch": 4.614209320091673,
      "grad_norm": 0.2952725887298584,
      "learning_rate": 1.7699467241416024e-06,
      "loss": 0.0038,
      "step": 3020
    },
    {
      "epoch": 4.629488158899924,
      "grad_norm": 0.23653660714626312,
      "learning_rate": 1.6319429361501714e-06,
      "loss": 0.0023,
      "step": 3030
    },
    {
      "epoch": 4.6447669977081745,
      "grad_norm": 0.014740560203790665,
      "learning_rate": 1.4994506980030577e-06,
      "loss": 0.0012,
      "step": 3040
    },
    {
      "epoch": 4.660045836516424,
      "grad_norm": 0.38012057542800903,
      "learning_rate": 1.3724851072174917e-06,
      "loss": 0.0013,
      "step": 3050
    },
    {
      "epoch": 4.660045836516424,
      "eval_loss": 0.050505463033914566,
      "eval_runtime": 200.7331,
      "eval_samples_per_second": 5.799,
      "eval_steps_per_second": 5.799,
      "step": 3050
    },
    {
      "epoch": 4.675324675324675,
      "grad_norm": 0.1817777156829834,
      "learning_rate": 1.251060631548112e-06,
      "loss": 0.0057,
      "step": 3060
    },
    {
      "epoch": 4.690603514132926,
      "grad_norm": 0.08571702241897583,
      "learning_rate": 1.135191107338368e-06,
      "loss": 0.0013,
      "step": 3070
    },
    {
      "epoch": 4.705882352941177,
      "grad_norm": 0.49542587995529175,
      "learning_rate": 1.0248897379438904e-06,
      "loss": 0.0028,
      "step": 3080
    },
    {
      "epoch": 4.721161191749427,
      "grad_norm": 0.037737760692834854,
      "learning_rate": 9.201690922279405e-07,
      "loss": 0.0024,
      "step": 3090
    },
    {
      "epoch": 4.736440030557677,
      "grad_norm": 0.04411200433969498,
      "learning_rate": 8.210411031291776e-07,
      "loss": 0.0031,
      "step": 3100
    },
    {
      "epoch": 4.736440030557677,
      "eval_loss": 0.051058944314718246,
      "eval_runtime": 201.9356,
      "eval_samples_per_second": 5.764,
      "eval_steps_per_second": 5.764,
      "step": 3100
    },
    {
      "epoch": 4.751718869365928,
      "grad_norm": 0.032969068735837936,
      "learning_rate": 7.275170663019415e-07,
      "loss": 0.002,
      "step": 3110
    },
    {
      "epoch": 4.766997708174179,
      "grad_norm": 0.40801408886909485,
      "learning_rate": 6.396076388290484e-07,
      "loss": 0.0036,
      "step": 3120
    },
    {
      "epoch": 4.7822765469824295,
      "grad_norm": 0.027664145454764366,
      "learning_rate": 5.573228380074736e-07,
      "loss": 0.0012,
      "step": 3130
    },
    {
      "epoch": 4.79755538579068,
      "grad_norm": 0.012874705716967583,
      "learning_rate": 4.806720402068477e-07,
      "loss": 0.0019,
      "step": 3140
    },
    {
      "epoch": 4.81283422459893,
      "grad_norm": 0.03708149120211601,
      "learning_rate": 4.0966397980100604e-07,
      "loss": 0.0031,
      "step": 3150
    },
    {
      "epoch": 4.81283422459893,
      "eval_loss": 0.05105143412947655,
      "eval_runtime": 202.4138,
      "eval_samples_per_second": 5.751,
      "eval_steps_per_second": 5.751,
      "step": 3150
    },
    {
      "epoch": 4.828113063407181,
      "grad_norm": 0.02528611198067665,
      "learning_rate": 3.4430674817274575e-07,
      "loss": 0.0018,
      "step": 3160
    },
    {
      "epoch": 4.843391902215432,
      "grad_norm": 0.05252419784665108,
      "learning_rate": 2.8460779279176896e-07,
      "loss": 0.002,
      "step": 3170
    },
    {
      "epoch": 4.8586707410236825,
      "grad_norm": 0.026198415085673332,
      "learning_rate": 2.3057391636606696e-07,
      "loss": 0.0012,
      "step": 3180
    },
    {
      "epoch": 4.873949579831933,
      "grad_norm": 0.46497654914855957,
      "learning_rate": 1.8221127606674605e-07,
      "loss": 0.0044,
      "step": 3190
    },
    {
      "epoch": 4.889228418640183,
      "grad_norm": 0.11807846277952194,
      "learning_rate": 1.3952538282639982e-07,
      "loss": 0.0023,
      "step": 3200
    },
    {
      "epoch": 4.889228418640183,
      "eval_loss": 0.051153309643268585,
      "eval_runtime": 201.9276,
      "eval_samples_per_second": 5.764,
      "eval_steps_per_second": 5.764,
      "step": 3200
    },
    {
      "epoch": 4.904507257448434,
      "grad_norm": 0.03819113224744797,
      "learning_rate": 1.025211007111615e-07,
      "loss": 0.0009,
      "step": 3210
    },
    {
      "epoch": 4.919786096256685,
      "grad_norm": 0.014340424910187721,
      "learning_rate": 7.120264636643615e-08,
      "loss": 0.0008,
      "step": 3220
    },
    {
      "epoch": 4.935064935064935,
      "grad_norm": 0.02623898908495903,
      "learning_rate": 4.5573588536407254e-08,
      "loss": 0.0034,
      "step": 3230
    },
    {
      "epoch": 4.950343773873186,
      "grad_norm": 0.42800572514533997,
      "learning_rate": 2.5636847657367623e-08,
      "loss": 0.0039,
      "step": 3240
    },
    {
      "epoch": 4.965622612681436,
      "grad_norm": 0.08229000121355057,
      "learning_rate": 1.1394695524963306e-08,
      "loss": 0.0023,
      "step": 3250
    },
    {
      "epoch": 4.965622612681436,
      "eval_loss": 0.051108263432979584,
      "eval_runtime": 202.0569,
      "eval_samples_per_second": 5.761,
      "eval_steps_per_second": 5.761,
      "step": 3250
    },
    {
      "epoch": 4.980901451489687,
      "grad_norm": 0.08490585535764694,
      "learning_rate": 2.8487550352951363e-09,
      "loss": 0.004,
      "step": 3260
    },
    {
      "epoch": 4.9961802902979375,
      "grad_norm": 0.07834521681070328,
      "learning_rate": 0.0,
      "loss": 0.0033,
      "step": 3270
    },
    {
      "epoch": 4.9961802902979375,
      "step": 3270,
      "total_flos": 8.213792003390177e+17,
      "train_loss": 0.04069158185845873,
      "train_runtime": 42808.6048,
      "train_samples_per_second": 1.223,
      "train_steps_per_second": 0.076
    }
  ],
  "logging_steps": 10,
  "max_steps": 3270,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.213792003390177e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}