{
  "best_metric": 0.9593754410743713,
  "best_model_checkpoint": "/scratch/czm5kz/NEW_finetuned_llama27b32_1_0.0003_alternate_no_output/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": 1.9491676092147827,
      "learning_rate": 0.0002989308624376336,
      "loss": 3.393,
      "step": 5
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.7517892122268677,
      "learning_rate": 0.00029786172487526725,
      "loss": 2.5606,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.6203854084014893,
      "learning_rate": 0.0002967925873129009,
      "loss": 1.987,
      "step": 15
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8757699728012085,
      "learning_rate": 0.00029572344975053457,
      "loss": 1.7622,
      "step": 20
    },
    {
      "epoch": 0.01,
      "eval_loss": 1.6077924966812134,
      "eval_runtime": 227.5054,
      "eval_samples_per_second": 49.357,
      "eval_steps_per_second": 6.171,
      "step": 20
    },
    {
      "epoch": 0.02,
      "grad_norm": 0.9948850870132446,
      "learning_rate": 0.00029465431218816815,
      "loss": 1.5741,
      "step": 25
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.1261812448501587,
      "learning_rate": 0.00029358517462580184,
      "loss": 1.4732,
      "step": 30
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.7787383794784546,
      "learning_rate": 0.0002925160370634355,
      "loss": 1.3471,
      "step": 35
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.935620129108429,
      "learning_rate": 0.0002914468995010691,
      "loss": 1.2991,
      "step": 40
    },
    {
      "epoch": 0.03,
      "eval_loss": 1.243576169013977,
      "eval_runtime": 227.5867,
      "eval_samples_per_second": 49.339,
      "eval_steps_per_second": 6.169,
      "step": 40
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.9946210980415344,
      "learning_rate": 0.00029037776193870275,
      "loss": 1.2733,
      "step": 45
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.828574538230896,
      "learning_rate": 0.0002893086243763364,
      "loss": 1.203,
      "step": 50
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.7947638630867004,
      "learning_rate": 0.00028823948681397,
      "loss": 1.1877,
      "step": 55
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.826960027217865,
      "learning_rate": 0.0002871703492516037,
      "loss": 1.1397,
      "step": 60
    },
    {
      "epoch": 0.04,
      "eval_loss": 1.1289246082305908,
      "eval_runtime": 227.0882,
      "eval_samples_per_second": 49.448,
      "eval_steps_per_second": 6.183,
      "step": 60
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.8590031862258911,
      "learning_rate": 0.0002861012116892373,
      "loss": 1.1081,
      "step": 65
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.6775720715522766,
      "learning_rate": 0.000285032074126871,
      "loss": 1.1283,
      "step": 70
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.059350609779358,
      "learning_rate": 0.0002839629365645046,
      "loss": 1.1091,
      "step": 75
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.754294216632843,
      "learning_rate": 0.00028289379900213826,
      "loss": 1.1062,
      "step": 80
    },
    {
      "epoch": 0.06,
      "eval_loss": 1.0860652923583984,
      "eval_runtime": 227.2483,
      "eval_samples_per_second": 49.413,
      "eval_steps_per_second": 6.178,
      "step": 80
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.7335007786750793,
      "learning_rate": 0.0002818246614397719,
      "loss": 1.0554,
      "step": 85
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.6339726448059082,
      "learning_rate": 0.00028075552387740553,
      "loss": 1.0523,
      "step": 90
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.1034719944000244,
      "learning_rate": 0.00027968638631503917,
      "loss": 1.0769,
      "step": 95
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.6975623369216919,
      "learning_rate": 0.0002786172487526728,
      "loss": 1.0596,
      "step": 100
    },
    {
      "epoch": 0.07,
      "eval_loss": 1.0658228397369385,
      "eval_runtime": 227.3685,
      "eval_samples_per_second": 49.387,
      "eval_steps_per_second": 6.175,
      "step": 100
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.6593163013458252,
      "learning_rate": 0.00027754811119030644,
      "loss": 1.079,
      "step": 105
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.6725738048553467,
      "learning_rate": 0.0002764789736279401,
      "loss": 1.0656,
      "step": 110
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.5702206492424011,
      "learning_rate": 0.00027540983606557377,
      "loss": 1.0733,
      "step": 115
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.7566413283348083,
      "learning_rate": 0.0002743406985032074,
      "loss": 1.0589,
      "step": 120
    },
    {
      "epoch": 0.09,
      "eval_loss": 1.0495774745941162,
      "eval_runtime": 227.0521,
      "eval_samples_per_second": 49.456,
      "eval_steps_per_second": 6.184,
      "step": 120
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.6189225912094116,
      "learning_rate": 0.00027327156094084104,
      "loss": 1.0778,
      "step": 125
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.6081388592720032,
      "learning_rate": 0.0002722024233784747,
      "loss": 1.0163,
      "step": 130
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.2115389108657837,
      "learning_rate": 0.0002711332858161083,
      "loss": 1.0035,
      "step": 135
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.6294423341751099,
      "learning_rate": 0.00027006414825374195,
      "loss": 1.0234,
      "step": 140
    },
    {
      "epoch": 0.1,
      "eval_loss": 1.0386356115341187,
      "eval_runtime": 227.6468,
      "eval_samples_per_second": 49.326,
      "eval_steps_per_second": 6.167,
      "step": 140
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.6168947219848633,
      "learning_rate": 0.00026899501069137564,
      "loss": 0.9842,
      "step": 145
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.5433680415153503,
      "learning_rate": 0.0002679258731290092,
      "loss": 1.0196,
      "step": 150
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.5716556906700134,
      "learning_rate": 0.0002668567355666429,
      "loss": 1.0261,
      "step": 155
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.7313571572303772,
      "learning_rate": 0.00026578759800427654,
      "loss": 1.0426,
      "step": 160
    },
    {
      "epoch": 0.11,
      "eval_loss": 1.0302481651306152,
      "eval_runtime": 227.5439,
      "eval_samples_per_second": 49.349,
      "eval_steps_per_second": 6.17,
      "step": 160
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.5894333124160767,
      "learning_rate": 0.0002647184604419102,
      "loss": 1.0848,
      "step": 165
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.5585373640060425,
      "learning_rate": 0.0002636493228795438,
      "loss": 1.0557,
      "step": 170
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.5632246136665344,
      "learning_rate": 0.00026258018531717745,
      "loss": 1.0309,
      "step": 175
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.49537792801856995,
      "learning_rate": 0.0002615110477548111,
      "loss": 1.0356,
      "step": 180
    },
    {
      "epoch": 0.13,
      "eval_loss": 1.0230164527893066,
      "eval_runtime": 227.8202,
      "eval_samples_per_second": 49.289,
      "eval_steps_per_second": 6.163,
      "step": 180
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.5201237201690674,
      "learning_rate": 0.0002604419101924447,
      "loss": 1.0166,
      "step": 185
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.7654304504394531,
      "learning_rate": 0.00025937277263007836,
      "loss": 1.0645,
      "step": 190
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.5226192474365234,
      "learning_rate": 0.000258303635067712,
      "loss": 1.0297,
      "step": 195
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.5896185040473938,
      "learning_rate": 0.0002572344975053457,
      "loss": 1.0265,
      "step": 200
    },
    {
      "epoch": 0.14,
      "eval_loss": 1.0194729566574097,
      "eval_runtime": 227.6113,
      "eval_samples_per_second": 49.334,
      "eval_steps_per_second": 6.168,
      "step": 200
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.5336751937866211,
      "learning_rate": 0.0002561653599429793,
      "loss": 1.0388,
      "step": 205
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.6317204833030701,
      "learning_rate": 0.00025509622238061296,
      "loss": 1.0365,
      "step": 210
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.5186619758605957,
      "learning_rate": 0.0002540270848182466,
      "loss": 0.9862,
      "step": 215
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.5502139925956726,
      "learning_rate": 0.00025295794725588023,
      "loss": 1.018,
      "step": 220
    },
    {
      "epoch": 0.16,
      "eval_loss": 1.015502691268921,
      "eval_runtime": 227.1348,
      "eval_samples_per_second": 49.438,
      "eval_steps_per_second": 6.181,
      "step": 220
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.4881739318370819,
      "learning_rate": 0.00025188880969351387,
      "loss": 1.0419,
      "step": 225
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.5655686855316162,
      "learning_rate": 0.00025081967213114756,
      "loss": 0.9913,
      "step": 230
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.5858956575393677,
      "learning_rate": 0.00024975053456878114,
      "loss": 1.0137,
      "step": 235
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.46040791273117065,
      "learning_rate": 0.00024868139700641483,
      "loss": 0.9942,
      "step": 240
    },
    {
      "epoch": 0.17,
      "eval_loss": 1.0137138366699219,
      "eval_runtime": 227.2315,
      "eval_samples_per_second": 49.417,
      "eval_steps_per_second": 6.179,
      "step": 240
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.6350478529930115,
      "learning_rate": 0.00024761225944404847,
      "loss": 1.0222,
      "step": 245
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.5849551558494568,
      "learning_rate": 0.0002465431218816821,
      "loss": 1.0154,
      "step": 250
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.42272746562957764,
      "learning_rate": 0.00024547398431931574,
      "loss": 1.0189,
      "step": 255
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.4941234886646271,
      "learning_rate": 0.0002444048467569494,
      "loss": 1.0114,
      "step": 260
    },
    {
      "epoch": 0.19,
      "eval_loss": 1.0054244995117188,
      "eval_runtime": 227.5101,
      "eval_samples_per_second": 49.356,
      "eval_steps_per_second": 6.171,
      "step": 260
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.4647514522075653,
      "learning_rate": 0.000243335709194583,
      "loss": 1.0339,
      "step": 265
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.4477405548095703,
      "learning_rate": 0.00024226657163221665,
      "loss": 1.0285,
      "step": 270
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.4374540448188782,
      "learning_rate": 0.0002411974340698503,
      "loss": 1.0008,
      "step": 275
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.39951157569885254,
      "learning_rate": 0.00024012829650748392,
      "loss": 0.9829,
      "step": 280
    },
    {
      "epoch": 0.2,
      "eval_loss": 1.0040351152420044,
      "eval_runtime": 227.8754,
      "eval_samples_per_second": 49.277,
      "eval_steps_per_second": 6.161,
      "step": 280
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.557092547416687,
      "learning_rate": 0.00023905915894511758,
      "loss": 0.9946,
      "step": 285
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.48575475811958313,
      "learning_rate": 0.00023799002138275122,
      "loss": 1.0176,
      "step": 290
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.4411045014858246,
      "learning_rate": 0.00023692088382038488,
      "loss": 1.0372,
      "step": 295
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.46225762367248535,
      "learning_rate": 0.0002358517462580185,
      "loss": 1.0448,
      "step": 300
    },
    {
      "epoch": 0.21,
      "eval_loss": 1.0029548406600952,
      "eval_runtime": 227.9278,
      "eval_samples_per_second": 49.266,
      "eval_steps_per_second": 6.16,
      "step": 300
    },
    {
      "epoch": 0.22,
      "grad_norm": 1.3242034912109375,
      "learning_rate": 0.00023478260869565215,
      "loss": 1.0389,
      "step": 305
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.6357295513153076,
      "learning_rate": 0.0002337134711332858,
      "loss": 0.97,
      "step": 310
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.6022586226463318,
      "learning_rate": 0.00023264433357091945,
      "loss": 0.9643,
      "step": 315
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.36969560384750366,
      "learning_rate": 0.0002315751960085531,
      "loss": 1.0139,
      "step": 320
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.9994259476661682,
      "eval_runtime": 227.6329,
      "eval_samples_per_second": 49.329,
      "eval_steps_per_second": 6.168,
      "step": 320
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.4300327003002167,
      "learning_rate": 0.00023050605844618672,
      "loss": 0.9854,
      "step": 325
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.5460650324821472,
      "learning_rate": 0.00022943692088382036,
      "loss": 1.0384,
      "step": 330
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.5851055383682251,
      "learning_rate": 0.00022836778332145402,
      "loss": 1.0378,
      "step": 335
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.48823001980781555,
      "learning_rate": 0.00022729864575908766,
      "loss": 0.995,
      "step": 340
    },
    {
      "epoch": 0.24,
      "eval_loss": 0.995897650718689,
      "eval_runtime": 227.1636,
      "eval_samples_per_second": 49.431,
      "eval_steps_per_second": 6.181,
      "step": 340
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.49934253096580505,
      "learning_rate": 0.00022622950819672127,
      "loss": 1.0244,
      "step": 345
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.4276964068412781,
      "learning_rate": 0.00022516037063435493,
      "loss": 0.9814,
      "step": 350
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.4696836471557617,
      "learning_rate": 0.00022409123307198857,
      "loss": 1.011,
      "step": 355
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.8325011730194092,
      "learning_rate": 0.00022302209550962223,
      "loss": 1.0064,
      "step": 360
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.9960118532180786,
      "eval_runtime": 227.2098,
      "eval_samples_per_second": 49.421,
      "eval_steps_per_second": 6.179,
      "step": 360
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.45505791902542114,
      "learning_rate": 0.00022195295794725584,
      "loss": 0.9677,
      "step": 365
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.5271835327148438,
      "learning_rate": 0.0002208838203848895,
      "loss": 1.0595,
      "step": 370
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.4093138873577118,
      "learning_rate": 0.00021981468282252314,
      "loss": 0.9966,
      "step": 375
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.44212207198143005,
      "learning_rate": 0.0002187455452601568,
      "loss": 0.9898,
      "step": 380
    },
    {
      "epoch": 0.27,
      "eval_loss": 0.9923149943351746,
      "eval_runtime": 227.4616,
      "eval_samples_per_second": 49.367,
      "eval_steps_per_second": 6.172,
      "step": 380
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.4539279341697693,
      "learning_rate": 0.0002176764076977904,
      "loss": 0.9675,
      "step": 385
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.4204266369342804,
      "learning_rate": 0.00021660727013542407,
      "loss": 1.0235,
      "step": 390
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.7492608428001404,
      "learning_rate": 0.0002155381325730577,
      "loss": 0.9584,
      "step": 395
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.5321412086486816,
      "learning_rate": 0.00021446899501069137,
      "loss": 0.9828,
      "step": 400
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.9924358129501343,
      "eval_runtime": 227.9389,
      "eval_samples_per_second": 49.263,
      "eval_steps_per_second": 6.16,
      "step": 400
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.5250059962272644,
      "learning_rate": 0.00021339985744832498,
      "loss": 1.0105,
      "step": 405
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.45952174067497253,
      "learning_rate": 0.00021233071988595865,
      "loss": 0.9956,
      "step": 410
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.45821675658226013,
      "learning_rate": 0.00021126158232359228,
      "loss": 1.0208,
      "step": 415
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.4415878355503082,
      "learning_rate": 0.00021019244476122595,
      "loss": 1.0286,
      "step": 420
    },
    {
      "epoch": 0.3,
      "eval_loss": 0.9904425740242004,
      "eval_runtime": 227.2284,
      "eval_samples_per_second": 49.417,
      "eval_steps_per_second": 6.179,
      "step": 420
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.4214901626110077,
      "learning_rate": 0.00020912330719885958,
      "loss": 0.9945,
      "step": 425
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.4757830500602722,
      "learning_rate": 0.0002080541696364932,
      "loss": 1.0125,
      "step": 430
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.48218998312950134,
      "learning_rate": 0.00020698503207412685,
      "loss": 0.987,
      "step": 435
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.5465131998062134,
      "learning_rate": 0.0002059158945117605,
      "loss": 0.9945,
      "step": 440
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.9884626865386963,
      "eval_runtime": 227.9678,
      "eval_samples_per_second": 49.257,
      "eval_steps_per_second": 6.159,
      "step": 440
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.6086324453353882,
      "learning_rate": 0.00020484675694939415,
      "loss": 0.9763,
      "step": 445
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.4712287187576294,
      "learning_rate": 0.00020377761938702776,
      "loss": 1.0055,
      "step": 450
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.46066540479660034,
      "learning_rate": 0.00020270848182466143,
      "loss": 1.0283,
      "step": 455
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.412824809551239,
      "learning_rate": 0.00020163934426229506,
      "loss": 0.9841,
      "step": 460
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.9868392944335938,
      "eval_runtime": 227.472,
      "eval_samples_per_second": 49.364,
      "eval_steps_per_second": 6.172,
      "step": 460
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.4112266004085541,
      "learning_rate": 0.00020057020669992872,
      "loss": 0.9619,
      "step": 465
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.41810086369514465,
      "learning_rate": 0.00019950106913756233,
      "loss": 0.9915,
      "step": 470
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.4301850497722626,
      "learning_rate": 0.000198431931575196,
      "loss": 0.9889,
      "step": 475
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.452982634305954,
      "learning_rate": 0.00019736279401282963,
      "loss": 1.0032,
      "step": 480
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.9845861792564392,
      "eval_runtime": 227.9942,
      "eval_samples_per_second": 49.251,
      "eval_steps_per_second": 6.158,
      "step": 480
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.5018645524978638,
      "learning_rate": 0.0001962936564504633,
      "loss": 0.9595,
      "step": 485
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.44759634137153625,
      "learning_rate": 0.0001952245188880969,
      "loss": 0.9759,
      "step": 490
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.5419363379478455,
      "learning_rate": 0.00019415538132573057,
      "loss": 0.9756,
      "step": 495
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.5175866484642029,
      "learning_rate": 0.0001930862437633642,
      "loss": 0.9977,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.9829397797584534,
      "eval_runtime": 227.4734,
      "eval_samples_per_second": 49.364,
      "eval_steps_per_second": 6.172,
      "step": 500
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.7049946188926697,
      "learning_rate": 0.00019201710620099787,
      "loss": 0.9869,
      "step": 505
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.4388104975223541,
      "learning_rate": 0.00019094796863863148,
      "loss": 1.0134,
      "step": 510
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.488154798746109,
      "learning_rate": 0.0001898788310762651,
      "loss": 1.0117,
      "step": 515
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.39673909544944763,
      "learning_rate": 0.00018880969351389878,
      "loss": 0.9987,
      "step": 520
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.9813041687011719,
      "eval_runtime": 227.269,
      "eval_samples_per_second": 49.408,
      "eval_steps_per_second": 6.178,
      "step": 520
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.4031273424625397,
      "learning_rate": 0.0001877405559515324,
      "loss": 0.9711,
      "step": 525
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.5331501364707947,
      "learning_rate": 0.00018667141838916605,
      "loss": 1.0479,
      "step": 530
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.4277609586715698,
      "learning_rate": 0.00018560228082679968,
      "loss": 1.0113,
      "step": 535
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.5301382541656494,
      "learning_rate": 0.00018453314326443335,
      "loss": 1.029,
      "step": 540
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.98179030418396,
      "eval_runtime": 227.6248,
      "eval_samples_per_second": 49.331,
      "eval_steps_per_second": 6.168,
      "step": 540
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.43908432126045227,
      "learning_rate": 0.00018346400570206698,
      "loss": 0.976,
      "step": 545
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.45188236236572266,
      "learning_rate": 0.00018239486813970065,
      "loss": 1.0183,
      "step": 550
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.3613215386867523,
      "learning_rate": 0.00018132573057733425,
      "loss": 0.9799,
      "step": 555
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.5476358532905579,
      "learning_rate": 0.00018025659301496792,
      "loss": 0.999,
      "step": 560
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.981633186340332,
      "eval_runtime": 227.815,
      "eval_samples_per_second": 49.29,
      "eval_steps_per_second": 6.163,
      "step": 560
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.5802431106567383,
      "learning_rate": 0.00017918745545260155,
      "loss": 1.0214,
      "step": 565
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.45726877450942993,
      "learning_rate": 0.00017811831789023522,
      "loss": 0.9812,
      "step": 570
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.4250318706035614,
      "learning_rate": 0.00017704918032786883,
      "loss": 0.9646,
      "step": 575
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.462782621383667,
      "learning_rate": 0.0001759800427655025,
      "loss": 1.0018,
      "step": 580
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.9794904589653015,
      "eval_runtime": 227.5588,
      "eval_samples_per_second": 49.345,
      "eval_steps_per_second": 6.17,
      "step": 580
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.4206041693687439,
      "learning_rate": 0.00017491090520313613,
      "loss": 1.0011,
      "step": 585
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.38388872146606445,
      "learning_rate": 0.0001738417676407698,
      "loss": 0.9974,
      "step": 590
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.4569980204105377,
      "learning_rate": 0.0001727726300784034,
      "loss": 0.9873,
      "step": 595
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.44570407271385193,
      "learning_rate": 0.00017170349251603703,
      "loss": 0.9819,
      "step": 600
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.9782843589782715,
      "eval_runtime": 227.3655,
      "eval_samples_per_second": 49.387,
      "eval_steps_per_second": 6.175,
      "step": 600
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.4842546284198761,
      "learning_rate": 0.0001706343549536707,
      "loss": 1.022,
      "step": 605
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.3910045921802521,
      "learning_rate": 0.00016956521739130433,
      "loss": 0.9678,
      "step": 610
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.4159318804740906,
      "learning_rate": 0.00016849607982893797,
      "loss": 0.9746,
      "step": 615
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.4413444995880127,
      "learning_rate": 0.0001674269422665716,
      "loss": 0.9651,
      "step": 620
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.9774540066719055,
      "eval_runtime": 227.3235,
      "eval_samples_per_second": 49.397,
      "eval_steps_per_second": 6.176,
      "step": 620
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.593654453754425,
      "learning_rate": 0.00016635780470420527,
      "loss": 0.9888,
      "step": 625
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.4957660734653473,
      "learning_rate": 0.0001652886671418389,
      "loss": 1.0392,
      "step": 630
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.4909263849258423,
      "learning_rate": 0.00016421952957947254,
      "loss": 0.9973,
      "step": 635
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.40913325548171997,
      "learning_rate": 0.00016315039201710618,
      "loss": 0.9688,
      "step": 640
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.9764226675033569,
      "eval_runtime": 228.0151,
      "eval_samples_per_second": 49.247,
      "eval_steps_per_second": 6.157,
      "step": 640
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.3790026605129242,
      "learning_rate": 0.00016208125445473984,
      "loss": 0.9521,
      "step": 645
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.3385336399078369,
      "learning_rate": 0.00016101211689237348,
      "loss": 0.9764,
      "step": 650
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.43496620655059814,
      "learning_rate": 0.00015994297933000714,
      "loss": 0.9575,
      "step": 655
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.47056716680526733,
      "learning_rate": 0.00015887384176764075,
      "loss": 0.985,
      "step": 660
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.9764449000358582,
      "eval_runtime": 227.4751,
      "eval_samples_per_second": 49.364,
      "eval_steps_per_second": 6.172,
      "step": 660
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.4433695673942566,
      "learning_rate": 0.0001578047042052744,
      "loss": 0.9669,
      "step": 665
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.40916725993156433,
      "learning_rate": 0.00015673556664290805,
      "loss": 1.0049,
      "step": 670
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.4507070779800415,
      "learning_rate": 0.0001556664290805417,
      "loss": 1.01,
      "step": 675
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.46380680799484253,
      "learning_rate": 0.00015459729151817532,
      "loss": 1.0002,
      "step": 680
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.9757907390594482,
      "eval_runtime": 227.4869,
      "eval_samples_per_second": 49.361,
      "eval_steps_per_second": 6.172,
      "step": 680
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.42961764335632324,
      "learning_rate": 0.00015352815395580896,
      "loss": 0.965,
      "step": 685
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.3826574683189392,
      "learning_rate": 0.00015245901639344262,
      "loss": 0.9412,
      "step": 690
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.4790586829185486,
      "learning_rate": 0.00015138987883107623,
      "loss": 0.9941,
      "step": 695
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.39779818058013916,
      "learning_rate": 0.0001503207412687099,
      "loss": 0.9649,
      "step": 700
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.9745392799377441,
      "eval_runtime": 227.4768,
      "eval_samples_per_second": 49.363,
      "eval_steps_per_second": 6.172,
      "step": 700
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.49817249178886414,
      "learning_rate": 0.00014925160370634355,
      "loss": 0.9573,
      "step": 705
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.373766154050827,
      "learning_rate": 0.0001481824661439772,
      "loss": 0.9868,
      "step": 710
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.3962858021259308,
      "learning_rate": 0.00014711332858161083,
      "loss": 0.9495,
      "step": 715
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.4002228081226349,
      "learning_rate": 0.00014604419101924446,
      "loss": 0.9574,
      "step": 720
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.972939133644104,
      "eval_runtime": 227.5154,
      "eval_samples_per_second": 49.355,
      "eval_steps_per_second": 6.171,
      "step": 720
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.7709822058677673,
      "learning_rate": 0.0001449750534568781,
      "loss": 1.007,
      "step": 725
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.42385196685791016,
      "learning_rate": 0.00014390591589451173,
      "loss": 1.0022,
      "step": 730
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.4056071937084198,
      "learning_rate": 0.0001428367783321454,
      "loss": 0.9409,
      "step": 735
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.4253057539463043,
      "learning_rate": 0.00014176764076977903,
      "loss": 0.9587,
      "step": 740
    },
    {
      "epoch": 0.53,
      "eval_loss": 0.9732517004013062,
      "eval_runtime": 227.3891,
      "eval_samples_per_second": 49.382,
      "eval_steps_per_second": 6.174,
      "step": 740
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.4211929142475128,
      "learning_rate": 0.00014069850320741267,
      "loss": 1.0021,
      "step": 745
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.4413307011127472,
      "learning_rate": 0.0001396293656450463,
      "loss": 1.0075,
      "step": 750
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.42298370599746704,
      "learning_rate": 0.00013856022808267997,
      "loss": 0.9788,
      "step": 755
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.3264484703540802,
      "learning_rate": 0.0001374910905203136,
      "loss": 0.9904,
      "step": 760
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.9725102186203003,
      "eval_runtime": 227.1577,
      "eval_samples_per_second": 49.433,
      "eval_steps_per_second": 6.181,
      "step": 760
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.3268262445926666,
      "learning_rate": 0.00013642195295794724,
      "loss": 0.9549,
      "step": 765
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.3310069143772125,
      "learning_rate": 0.00013535281539558088,
      "loss": 0.9848,
      "step": 770
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.38187867403030396,
      "learning_rate": 0.00013428367783321454,
      "loss": 0.9802,
      "step": 775
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.37355056405067444,
      "learning_rate": 0.00013321454027084818,
      "loss": 0.9959,
      "step": 780
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.9710414409637451,
      "eval_runtime": 227.6662,
      "eval_samples_per_second": 49.322,
      "eval_steps_per_second": 6.167,
      "step": 780
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.48310089111328125,
      "learning_rate": 0.0001321454027084818,
      "loss": 1.0199,
      "step": 785
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.3701138198375702,
      "learning_rate": 0.00013107626514611545,
      "loss": 0.9983,
      "step": 790
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.359737366437912,
      "learning_rate": 0.0001300071275837491,
      "loss": 0.9747,
      "step": 795
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.5300225615501404,
      "learning_rate": 0.00012893799002138275,
      "loss": 1.0007,
      "step": 800
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.9719007015228271,
      "eval_runtime": 227.4887,
      "eval_samples_per_second": 49.361,
      "eval_steps_per_second": 6.172,
      "step": 800
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.43651485443115234,
      "learning_rate": 0.00012786885245901638,
      "loss": 0.9885,
      "step": 805
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.4850105047225952,
      "learning_rate": 0.00012679971489665002,
      "loss": 1.0109,
      "step": 810
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.45454463362693787,
      "learning_rate": 0.00012573057733428366,
      "loss": 0.9987,
      "step": 815
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.40239349007606506,
      "learning_rate": 0.0001246614397719173,
      "loss": 0.9995,
      "step": 820
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.9698851108551025,
      "eval_runtime": 227.4941,
      "eval_samples_per_second": 49.36,
      "eval_steps_per_second": 6.172,
      "step": 820
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.4248133897781372,
      "learning_rate": 0.00012359230220955095,
      "loss": 1.0366,
      "step": 825
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.4477006494998932,
      "learning_rate": 0.0001225231646471846,
      "loss": 0.9866,
      "step": 830
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.45568859577178955,
      "learning_rate": 0.00012145402708481824,
      "loss": 0.9754,
      "step": 835
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.4142056107521057,
      "learning_rate": 0.00012038488952245188,
      "loss": 0.9394,
      "step": 840
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.9695276021957397,
      "eval_runtime": 227.2433,
      "eval_samples_per_second": 49.414,
      "eval_steps_per_second": 6.178,
      "step": 840
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.44050607085227966,
      "learning_rate": 0.00011931575196008553,
      "loss": 0.9709,
      "step": 845
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.45126873254776,
      "learning_rate": 0.00011824661439771916,
      "loss": 1.0009,
      "step": 850
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.47706660628318787,
      "learning_rate": 0.00011717747683535281,
      "loss": 0.9789,
      "step": 855
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.43811845779418945,
      "learning_rate": 0.00011610833927298645,
      "loss": 0.9558,
      "step": 860
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.9690415263175964,
      "eval_runtime": 227.2613,
      "eval_samples_per_second": 49.41,
      "eval_steps_per_second": 6.178,
      "step": 860
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.5149155855178833,
      "learning_rate": 0.0001150392017106201,
      "loss": 0.9848,
      "step": 865
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.482454389333725,
      "learning_rate": 0.00011397006414825373,
      "loss": 0.9798,
      "step": 870
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.3987056612968445,
      "learning_rate": 0.00011290092658588738,
      "loss": 0.9388,
      "step": 875
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.4491436183452606,
      "learning_rate": 0.00011183178902352102,
      "loss": 0.9512,
      "step": 880
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.9684551954269409,
      "eval_runtime": 227.6007,
      "eval_samples_per_second": 49.336,
      "eval_steps_per_second": 6.169,
      "step": 880
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.4160451889038086,
      "learning_rate": 0.00011076265146115467,
      "loss": 0.9869,
      "step": 885
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.5239384770393372,
      "learning_rate": 0.00010969351389878829,
      "loss": 0.9676,
      "step": 890
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.42906704545021057,
      "learning_rate": 0.00010862437633642194,
      "loss": 0.9611,
      "step": 895
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.3256273567676544,
      "learning_rate": 0.00010755523877405558,
      "loss": 0.9908,
      "step": 900
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.967640221118927,
      "eval_runtime": 227.6932,
      "eval_samples_per_second": 49.316,
      "eval_steps_per_second": 6.166,
      "step": 900
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.42826366424560547,
      "learning_rate": 0.00010648610121168923,
      "loss": 0.9655,
      "step": 905
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.46543028950691223,
      "learning_rate": 0.00010541696364932286,
      "loss": 0.9607,
      "step": 910
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.5507635474205017,
      "learning_rate": 0.00010434782608695651,
      "loss": 0.9864,
      "step": 915
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.42310115694999695,
      "learning_rate": 0.00010327868852459015,
      "loss": 0.9611,
      "step": 920
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.967984676361084,
      "eval_runtime": 227.9969,
      "eval_samples_per_second": 49.251,
      "eval_steps_per_second": 6.158,
      "step": 920
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.4114033281803131,
      "learning_rate": 0.0001022095509622238,
      "loss": 0.9503,
      "step": 925
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.33842840790748596,
      "learning_rate": 0.00010114041339985743,
      "loss": 0.939,
      "step": 930
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.35206928849220276,
      "learning_rate": 0.00010007127583749108,
      "loss": 1.0102,
      "step": 935
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.45287784934043884,
      "learning_rate": 9.900213827512472e-05,
      "loss": 0.9444,
      "step": 940
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.9666356444358826,
      "eval_runtime": 227.3836,
      "eval_samples_per_second": 49.384,
      "eval_steps_per_second": 6.175,
      "step": 940
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.5042274594306946,
      "learning_rate": 9.793300071275837e-05,
      "loss": 0.9986,
      "step": 945
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.4130496084690094,
      "learning_rate": 9.686386315039202e-05,
      "loss": 0.9481,
      "step": 950
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.4442286491394043,
      "learning_rate": 9.579472558802566e-05,
      "loss": 0.9832,
      "step": 955
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.5206668972969055,
      "learning_rate": 9.47255880256593e-05,
      "loss": 0.9668,
      "step": 960
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.9656959772109985,
      "eval_runtime": 228.0783,
      "eval_samples_per_second": 49.233,
      "eval_steps_per_second": 6.156,
      "step": 960
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.4653392434120178,
      "learning_rate": 9.365645046329294e-05,
      "loss": 0.9966,
      "step": 965
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.44000840187072754,
      "learning_rate": 9.258731290092659e-05,
      "loss": 1.0002,
      "step": 970
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.4088016450405121,
      "learning_rate": 9.151817533856021e-05,
      "loss": 0.9868,
      "step": 975
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.4259507656097412,
      "learning_rate": 9.044903777619385e-05,
      "loss": 0.9887,
      "step": 980
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.965515673160553,
      "eval_runtime": 228.1355,
      "eval_samples_per_second": 49.221,
      "eval_steps_per_second": 6.154,
      "step": 980
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.42707499861717224,
      "learning_rate": 8.93799002138275e-05,
      "loss": 1.0194,
      "step": 985
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.38201627135276794,
      "learning_rate": 8.831076265146115e-05,
      "loss": 0.9835,
      "step": 990
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.4495198428630829,
      "learning_rate": 8.724162508909478e-05,
      "loss": 0.9911,
      "step": 995
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.4431660771369934,
      "learning_rate": 8.617248752672843e-05,
      "loss": 0.9759,
      "step": 1000
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.964968204498291,
      "eval_runtime": 227.5676,
      "eval_samples_per_second": 49.344,
      "eval_steps_per_second": 6.17,
      "step": 1000
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.43940964341163635,
      "learning_rate": 8.510334996436207e-05,
      "loss": 0.958,
      "step": 1005
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.4617341458797455,
      "learning_rate": 8.403421240199572e-05,
      "loss": 0.9383,
      "step": 1010
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.33945685625076294,
      "learning_rate": 8.296507483962936e-05,
      "loss": 0.9541,
      "step": 1015
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.45421719551086426,
      "learning_rate": 8.1895937277263e-05,
      "loss": 0.9567,
      "step": 1020
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.9648857712745667,
      "eval_runtime": 227.4739,
      "eval_samples_per_second": 49.364,
      "eval_steps_per_second": 6.172,
      "step": 1020
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.3823011815547943,
      "learning_rate": 8.082679971489664e-05,
      "loss": 0.9496,
      "step": 1025
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.39452335238456726,
      "learning_rate": 7.975766215253029e-05,
      "loss": 0.9633,
      "step": 1030
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.3680623769760132,
      "learning_rate": 7.868852459016393e-05,
      "loss": 0.988,
      "step": 1035
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.5220345854759216,
      "learning_rate": 7.761938702779758e-05,
      "loss": 0.9858,
      "step": 1040
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.9640631079673767,
      "eval_runtime": 227.6642,
      "eval_samples_per_second": 49.323,
      "eval_steps_per_second": 6.167,
      "step": 1040
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.39960744976997375,
      "learning_rate": 7.655024946543121e-05,
      "loss": 0.9936,
      "step": 1045
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.43278273940086365,
      "learning_rate": 7.548111190306486e-05,
      "loss": 0.9618,
      "step": 1050
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.3960849642753601,
      "learning_rate": 7.44119743406985e-05,
      "loss": 0.999,
      "step": 1055
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.4232223629951477,
      "learning_rate": 7.334283677833213e-05,
      "loss": 0.9702,
      "step": 1060
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.9640046954154968,
      "eval_runtime": 227.4939,
      "eval_samples_per_second": 49.36,
      "eval_steps_per_second": 6.172,
      "step": 1060
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.3899974822998047,
      "learning_rate": 7.227369921596578e-05,
      "loss": 0.9744,
      "step": 1065
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.38672515749931335,
      "learning_rate": 7.120456165359942e-05,
      "loss": 0.9424,
      "step": 1070
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.33859318494796753,
      "learning_rate": 7.013542409123307e-05,
      "loss": 0.9721,
      "step": 1075
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.4227427542209625,
      "learning_rate": 6.90662865288667e-05,
      "loss": 0.9731,
      "step": 1080
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.9635033011436462,
      "eval_runtime": 227.4766,
      "eval_samples_per_second": 49.363,
      "eval_steps_per_second": 6.172,
      "step": 1080
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.6176052689552307,
      "learning_rate": 6.799714896650034e-05,
      "loss": 0.99,
      "step": 1085
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.3367667496204376,
      "learning_rate": 6.692801140413399e-05,
      "loss": 0.965,
      "step": 1090
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.41608142852783203,
      "learning_rate": 6.585887384176763e-05,
      "loss": 0.9678,
      "step": 1095
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.3237856328487396,
      "learning_rate": 6.478973627940128e-05,
      "loss": 0.927,
      "step": 1100
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.9633656144142151,
      "eval_runtime": 227.8317,
      "eval_samples_per_second": 49.286,
      "eval_steps_per_second": 6.162,
      "step": 1100
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.3576589822769165,
      "learning_rate": 6.372059871703493e-05,
      "loss": 0.9729,
      "step": 1105
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.47324052453041077,
      "learning_rate": 6.265146115466856e-05,
      "loss": 0.989,
      "step": 1110
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.42953863739967346,
      "learning_rate": 6.158232359230221e-05,
      "loss": 0.9566,
      "step": 1115
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.4253891110420227,
      "learning_rate": 6.051318602993584e-05,
      "loss": 0.9878,
      "step": 1120
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.9624494314193726,
      "eval_runtime": 227.67,
      "eval_samples_per_second": 49.321,
      "eval_steps_per_second": 6.167,
      "step": 1120
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.4608158767223358,
      "learning_rate": 5.9444048467569485e-05,
      "loss": 0.9787,
      "step": 1125
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.4022761285305023,
      "learning_rate": 5.837491090520313e-05,
      "loss": 0.9738,
      "step": 1130
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.3572002053260803,
      "learning_rate": 5.730577334283677e-05,
      "loss": 0.9919,
      "step": 1135
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.4154810905456543,
      "learning_rate": 5.6236635780470413e-05,
      "loss": 0.9861,
      "step": 1140
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.962546169757843,
      "eval_runtime": 227.8477,
      "eval_samples_per_second": 49.283,
      "eval_steps_per_second": 6.162,
      "step": 1140
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.36054152250289917,
      "learning_rate": 5.5167498218104056e-05,
      "loss": 0.9834,
      "step": 1145
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.3918026089668274,
      "learning_rate": 5.40983606557377e-05,
      "loss": 0.9978,
      "step": 1150
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.4808182120323181,
      "learning_rate": 5.302922309337134e-05,
      "loss": 0.9481,
      "step": 1155
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.37957093119621277,
      "learning_rate": 5.196008553100499e-05,
      "loss": 0.9703,
      "step": 1160
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.9620444178581238,
      "eval_runtime": 228.027,
      "eval_samples_per_second": 49.244,
      "eval_steps_per_second": 6.157,
      "step": 1160
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.3866080939769745,
      "learning_rate": 5.089094796863862e-05,
      "loss": 1.0014,
      "step": 1165
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.4046897888183594,
      "learning_rate": 4.9821810406272264e-05,
      "loss": 0.9911,
      "step": 1170
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.4034290909767151,
      "learning_rate": 4.875267284390591e-05,
      "loss": 0.908,
      "step": 1175
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.39161401987075806,
      "learning_rate": 4.7683535281539556e-05,
      "loss": 0.9699,
      "step": 1180
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.9614489674568176,
      "eval_runtime": 227.849,
      "eval_samples_per_second": 49.283,
      "eval_steps_per_second": 6.162,
      "step": 1180
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.43111854791641235,
      "learning_rate": 4.66143977191732e-05,
      "loss": 1.0008,
      "step": 1185
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.3769904673099518,
      "learning_rate": 4.554526015680684e-05,
      "loss": 1.0144,
      "step": 1190
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.466468870639801,
      "learning_rate": 4.4476122594440485e-05,
      "loss": 0.9589,
      "step": 1195
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.41613641381263733,
      "learning_rate": 4.340698503207413e-05,
      "loss": 1.0043,
      "step": 1200
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.9611164927482605,
      "eval_runtime": 227.5578,
      "eval_samples_per_second": 49.346,
      "eval_steps_per_second": 6.17,
      "step": 1200
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.5018350481987,
      "learning_rate": 4.2337847469707764e-05,
      "loss": 0.9734,
      "step": 1205
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.5452864766120911,
      "learning_rate": 4.1268709907341407e-05,
      "loss": 0.953,
      "step": 1210
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.4102606177330017,
      "learning_rate": 4.019957234497505e-05,
      "loss": 0.9841,
      "step": 1215
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.4345894753932953,
      "learning_rate": 3.913043478260869e-05,
      "loss": 1.0105,
      "step": 1220
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.9609247446060181,
      "eval_runtime": 227.5088,
      "eval_samples_per_second": 49.356,
      "eval_steps_per_second": 6.171,
      "step": 1220
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.34100887179374695,
      "learning_rate": 3.8061297220242335e-05,
      "loss": 0.9571,
      "step": 1225
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.3881888687610626,
      "learning_rate": 3.699215965787598e-05,
      "loss": 0.9768,
      "step": 1230
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.36524873971939087,
      "learning_rate": 3.592302209550962e-05,
      "loss": 0.9543,
      "step": 1235
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.39331433176994324,
      "learning_rate": 3.485388453314326e-05,
      "loss": 0.9705,
      "step": 1240
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.9608638286590576,
      "eval_runtime": 227.7591,
      "eval_samples_per_second": 49.302,
      "eval_steps_per_second": 6.164,
      "step": 1240
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.4295852780342102,
      "learning_rate": 3.3784746970776906e-05,
      "loss": 0.9623,
      "step": 1245
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.4223569333553314,
      "learning_rate": 3.271560940841055e-05,
      "loss": 0.964,
      "step": 1250
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.3776945173740387,
      "learning_rate": 3.164647184604419e-05,
      "loss": 0.9981,
      "step": 1255
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.3786326050758362,
      "learning_rate": 3.057733428367783e-05,
      "loss": 0.9738,
      "step": 1260
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.9605663418769836,
      "eval_runtime": 227.6759,
      "eval_samples_per_second": 49.32,
      "eval_steps_per_second": 6.167,
      "step": 1260
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.5197612643241882,
      "learning_rate": 2.950819672131147e-05,
      "loss": 0.9834,
      "step": 1265
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.42392510175704956,
      "learning_rate": 2.8439059158945114e-05,
      "loss": 1.0068,
      "step": 1270
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.3506150543689728,
      "learning_rate": 2.736992159657876e-05,
      "loss": 0.9518,
      "step": 1275
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.3928501605987549,
      "learning_rate": 2.63007840342124e-05,
      "loss": 1.0032,
      "step": 1280
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.9605409502983093,
      "eval_runtime": 227.9108,
      "eval_samples_per_second": 49.269,
      "eval_steps_per_second": 6.16,
      "step": 1280
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.3545394241809845,
      "learning_rate": 2.5231646471846042e-05,
      "loss": 0.9809,
      "step": 1285
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.3934232294559479,
      "learning_rate": 2.4162508909479685e-05,
      "loss": 0.9812,
      "step": 1290
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.4914129674434662,
      "learning_rate": 2.3093371347113328e-05,
      "loss": 1.041,
      "step": 1295
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.45427605509757996,
      "learning_rate": 2.2024233784746968e-05,
      "loss": 0.9473,
      "step": 1300
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.9605054259300232,
      "eval_runtime": 228.4903,
      "eval_samples_per_second": 49.144,
      "eval_steps_per_second": 6.145,
      "step": 1300
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.3592887818813324,
      "learning_rate": 2.095509622238061e-05,
      "loss": 0.9616,
      "step": 1305
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.47144436836242676,
      "learning_rate": 1.9885958660014253e-05,
      "loss": 0.9862,
      "step": 1310
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.45374444127082825,
      "learning_rate": 1.8816821097647896e-05,
      "loss": 1.006,
      "step": 1315
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.49605900049209595,
      "learning_rate": 1.774768353528154e-05,
      "loss": 0.9037,
      "step": 1320
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.9600586295127869,
      "eval_runtime": 227.3914,
      "eval_samples_per_second": 49.382,
      "eval_steps_per_second": 6.174,
      "step": 1320
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.388639360666275,
      "learning_rate": 1.6678545972915182e-05,
      "loss": 0.9601,
      "step": 1325
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.47878703474998474,
      "learning_rate": 1.560940841054882e-05,
      "loss": 1.0,
      "step": 1330
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.42967456579208374,
      "learning_rate": 1.4540270848182466e-05,
      "loss": 0.996,
      "step": 1335
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.3042117953300476,
      "learning_rate": 1.3471133285816107e-05,
      "loss": 0.9258,
      "step": 1340
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.9597083926200867,
      "eval_runtime": 227.7172,
      "eval_samples_per_second": 49.311,
      "eval_steps_per_second": 6.166,
      "step": 1340
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.44863736629486084,
      "learning_rate": 1.240199572344975e-05,
      "loss": 0.9422,
      "step": 1345
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.47229892015457153,
      "learning_rate": 1.1332858161083391e-05,
      "loss": 0.9497,
      "step": 1350
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.4254206120967865,
      "learning_rate": 1.0263720598717034e-05,
      "loss": 0.993,
      "step": 1355
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.3855699896812439,
      "learning_rate": 9.194583036350677e-06,
      "loss": 0.9965,
      "step": 1360
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.9595157504081726,
      "eval_runtime": 227.4432,
      "eval_samples_per_second": 49.371,
      "eval_steps_per_second": 6.173,
      "step": 1360
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.3840392231941223,
      "learning_rate": 8.12544547398432e-06,
      "loss": 0.9326,
      "step": 1365
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.36666616797447205,
      "learning_rate": 7.0563079116179615e-06,
      "loss": 0.9729,
      "step": 1370
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.379976749420166,
      "learning_rate": 5.9871703492516035e-06,
      "loss": 0.9771,
      "step": 1375
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.3656114935874939,
      "learning_rate": 4.9180327868852455e-06,
      "loss": 0.9165,
      "step": 1380
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.959417998790741,
      "eval_runtime": 227.5694,
      "eval_samples_per_second": 49.343,
      "eval_steps_per_second": 6.17,
      "step": 1380
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.4715186059474945,
      "learning_rate": 3.848895224518888e-06,
      "loss": 0.9531,
      "step": 1385
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.4237213730812073,
      "learning_rate": 2.7797576621525303e-06,
      "loss": 0.9883,
      "step": 1390
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.43288546800613403,
      "learning_rate": 1.7106200997861725e-06,
      "loss": 0.9835,
      "step": 1395
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.3952523171901703,
      "learning_rate": 6.414825374198146e-07,
      "loss": 1.022,
      "step": 1400
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.9593754410743713,
      "eval_runtime": 227.652,
      "eval_samples_per_second": 49.325,
      "eval_steps_per_second": 6.167,
      "step": 1400
    }
  ],
  "logging_steps": 5,
  "max_steps": 1403,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 20,
  "total_flos": 2.634861827260416e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}