{
  "best_metric": 0.0025608371943235397,
  "best_model_checkpoint": "finetuned-arsenic/checkpoint-1700",
  "epoch": 4.0,
  "eval_steps": 100,
  "global_step": 2164,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.018484288354898338,
      "grad_norm": 3.1622745990753174,
      "learning_rate": 0.0001990757855822551,
      "loss": 0.5015,
      "step": 10
    },
    {
      "epoch": 0.036968576709796676,
      "grad_norm": 3.4678640365600586,
      "learning_rate": 0.00019815157116451017,
      "loss": 0.4604,
      "step": 20
    },
    {
      "epoch": 0.05545286506469501,
      "grad_norm": 3.134012460708618,
      "learning_rate": 0.00019722735674676528,
      "loss": 0.2732,
      "step": 30
    },
    {
      "epoch": 0.07393715341959335,
      "grad_norm": 1.4681167602539062,
      "learning_rate": 0.00019630314232902034,
      "loss": 0.4454,
      "step": 40
    },
    {
      "epoch": 0.09242144177449169,
      "grad_norm": 2.2367947101593018,
      "learning_rate": 0.00019537892791127544,
      "loss": 0.3936,
      "step": 50
    },
    {
      "epoch": 0.11090573012939002,
      "grad_norm": 1.4795340299606323,
      "learning_rate": 0.0001944547134935305,
      "loss": 0.2588,
      "step": 60
    },
    {
      "epoch": 0.12939001848428835,
      "grad_norm": 1.862062692642212,
      "learning_rate": 0.0001935304990757856,
      "loss": 0.3119,
      "step": 70
    },
    {
      "epoch": 0.1478743068391867,
      "grad_norm": 2.609395980834961,
      "learning_rate": 0.00019260628465804066,
      "loss": 0.2267,
      "step": 80
    },
    {
      "epoch": 0.16635859519408502,
      "grad_norm": 1.176942229270935,
      "learning_rate": 0.00019168207024029577,
      "loss": 0.1786,
      "step": 90
    },
    {
      "epoch": 0.18484288354898337,
      "grad_norm": 3.820526599884033,
      "learning_rate": 0.00019075785582255082,
      "loss": 0.2214,
      "step": 100
    },
    {
      "epoch": 0.18484288354898337,
      "eval_accuracy": 0.9607072691552063,
      "eval_loss": 0.12430191040039062,
      "eval_runtime": 63.6085,
      "eval_samples_per_second": 24.006,
      "eval_steps_per_second": 3.003,
      "step": 100
    },
    {
      "epoch": 0.2033271719038817,
      "grad_norm": 1.79490065574646,
      "learning_rate": 0.00018983364140480593,
      "loss": 0.2468,
      "step": 110
    },
    {
      "epoch": 0.22181146025878004,
      "grad_norm": 0.4057539701461792,
      "learning_rate": 0.000188909426987061,
      "loss": 0.2852,
      "step": 120
    },
    {
      "epoch": 0.24029574861367836,
      "grad_norm": 0.1476714164018631,
      "learning_rate": 0.0001879852125693161,
      "loss": 0.1344,
      "step": 130
    },
    {
      "epoch": 0.2587800369685767,
      "grad_norm": 3.7497174739837646,
      "learning_rate": 0.00018706099815157118,
      "loss": 0.1937,
      "step": 140
    },
    {
      "epoch": 0.27726432532347506,
      "grad_norm": 1.4686886072158813,
      "learning_rate": 0.00018613678373382626,
      "loss": 0.14,
      "step": 150
    },
    {
      "epoch": 0.2957486136783734,
      "grad_norm": 2.834284782409668,
      "learning_rate": 0.00018521256931608134,
      "loss": 0.2202,
      "step": 160
    },
    {
      "epoch": 0.3142329020332717,
      "grad_norm": 3.5288290977478027,
      "learning_rate": 0.00018428835489833642,
      "loss": 0.1548,
      "step": 170
    },
    {
      "epoch": 0.33271719038817005,
      "grad_norm": 2.251044273376465,
      "learning_rate": 0.0001833641404805915,
      "loss": 0.2552,
      "step": 180
    },
    {
      "epoch": 0.3512014787430684,
      "grad_norm": 2.9413559436798096,
      "learning_rate": 0.00018243992606284658,
      "loss": 0.1838,
      "step": 190
    },
    {
      "epoch": 0.36968576709796674,
      "grad_norm": 1.2569198608398438,
      "learning_rate": 0.0001815157116451017,
      "loss": 0.1213,
      "step": 200
    },
    {
      "epoch": 0.36968576709796674,
      "eval_accuracy": 0.933857236411264,
      "eval_loss": 0.17631804943084717,
      "eval_runtime": 52.9442,
      "eval_samples_per_second": 28.842,
      "eval_steps_per_second": 3.608,
      "step": 200
    },
    {
      "epoch": 0.38817005545286504,
      "grad_norm": 6.763731002807617,
      "learning_rate": 0.00018059149722735675,
      "loss": 0.1987,
      "step": 210
    },
    {
      "epoch": 0.4066543438077634,
      "grad_norm": 1.2817621231079102,
      "learning_rate": 0.00017966728280961186,
      "loss": 0.2626,
      "step": 220
    },
    {
      "epoch": 0.42513863216266173,
      "grad_norm": 5.742928504943848,
      "learning_rate": 0.0001787430683918669,
      "loss": 0.1899,
      "step": 230
    },
    {
      "epoch": 0.4436229205175601,
      "grad_norm": 0.5674309134483337,
      "learning_rate": 0.00017781885397412202,
      "loss": 0.0699,
      "step": 240
    },
    {
      "epoch": 0.46210720887245843,
      "grad_norm": 5.822825908660889,
      "learning_rate": 0.00017689463955637707,
      "loss": 0.091,
      "step": 250
    },
    {
      "epoch": 0.4805914972273567,
      "grad_norm": 0.5449599027633667,
      "learning_rate": 0.00017597042513863218,
      "loss": 0.1584,
      "step": 260
    },
    {
      "epoch": 0.49907578558225507,
      "grad_norm": 0.09896630793809891,
      "learning_rate": 0.00017504621072088724,
      "loss": 0.2052,
      "step": 270
    },
    {
      "epoch": 0.5175600739371534,
      "grad_norm": 4.201374530792236,
      "learning_rate": 0.00017412199630314234,
      "loss": 0.1447,
      "step": 280
    },
    {
      "epoch": 0.5360443622920518,
      "grad_norm": 6.048889636993408,
      "learning_rate": 0.00017319778188539743,
      "loss": 0.1032,
      "step": 290
    },
    {
      "epoch": 0.5545286506469501,
      "grad_norm": 5.420243740081787,
      "learning_rate": 0.0001722735674676525,
      "loss": 0.1201,
      "step": 300
    },
    {
      "epoch": 0.5545286506469501,
      "eval_accuracy": 0.9607072691552063,
      "eval_loss": 0.10179698467254639,
      "eval_runtime": 50.5522,
      "eval_samples_per_second": 30.206,
      "eval_steps_per_second": 3.778,
      "step": 300
    },
    {
      "epoch": 0.5730129390018485,
      "grad_norm": 1.7872523069381714,
      "learning_rate": 0.0001713493530499076,
      "loss": 0.1725,
      "step": 310
    },
    {
      "epoch": 0.5914972273567468,
      "grad_norm": 3.7505619525909424,
      "learning_rate": 0.00017042513863216267,
      "loss": 0.0805,
      "step": 320
    },
    {
      "epoch": 0.609981515711645,
      "grad_norm": 3.3265647888183594,
      "learning_rate": 0.00016950092421441775,
      "loss": 0.0945,
      "step": 330
    },
    {
      "epoch": 0.6284658040665434,
      "grad_norm": 0.0822325125336647,
      "learning_rate": 0.00016857670979667283,
      "loss": 0.184,
      "step": 340
    },
    {
      "epoch": 0.6469500924214417,
      "grad_norm": 6.458297252655029,
      "learning_rate": 0.00016765249537892791,
      "loss": 0.2462,
      "step": 350
    },
    {
      "epoch": 0.6654343807763401,
      "grad_norm": 8.956634521484375,
      "learning_rate": 0.000166728280961183,
      "loss": 0.1549,
      "step": 360
    },
    {
      "epoch": 0.6839186691312384,
      "grad_norm": 8.221627235412598,
      "learning_rate": 0.00016580406654343808,
      "loss": 0.0573,
      "step": 370
    },
    {
      "epoch": 0.7024029574861368,
      "grad_norm": 1.3730543851852417,
      "learning_rate": 0.00016487985212569316,
      "loss": 0.1435,
      "step": 380
    },
    {
      "epoch": 0.7208872458410351,
      "grad_norm": 3.857621669769287,
      "learning_rate": 0.00016395563770794827,
      "loss": 0.1512,
      "step": 390
    },
    {
      "epoch": 0.7393715341959335,
      "grad_norm": 0.7978448867797852,
      "learning_rate": 0.00016303142329020332,
      "loss": 0.0991,
      "step": 400
    },
    {
      "epoch": 0.7393715341959335,
      "eval_accuracy": 0.9417157825802227,
      "eval_loss": 0.20708701014518738,
      "eval_runtime": 49.1709,
      "eval_samples_per_second": 31.055,
      "eval_steps_per_second": 3.884,
      "step": 400
    },
    {
      "epoch": 0.7578558225508318,
      "grad_norm": 0.13973641395568848,
      "learning_rate": 0.00016210720887245843,
      "loss": 0.2316,
      "step": 410
    },
    {
      "epoch": 0.7763401109057301,
      "grad_norm": 5.2767415046691895,
      "learning_rate": 0.00016118299445471348,
      "loss": 0.0945,
      "step": 420
    },
    {
      "epoch": 0.7948243992606284,
      "grad_norm": 1.6643435955047607,
      "learning_rate": 0.0001602587800369686,
      "loss": 0.1037,
      "step": 430
    },
    {
      "epoch": 0.8133086876155268,
      "grad_norm": 2.7701528072357178,
      "learning_rate": 0.00015933456561922367,
      "loss": 0.0687,
      "step": 440
    },
    {
      "epoch": 0.8317929759704251,
      "grad_norm": 0.07744336128234863,
      "learning_rate": 0.00015841035120147876,
      "loss": 0.1097,
      "step": 450
    },
    {
      "epoch": 0.8502772643253235,
      "grad_norm": 1.1288193464279175,
      "learning_rate": 0.00015748613678373384,
      "loss": 0.1874,
      "step": 460
    },
    {
      "epoch": 0.8687615526802218,
      "grad_norm": 7.898807525634766,
      "learning_rate": 0.00015656192236598892,
      "loss": 0.1038,
      "step": 470
    },
    {
      "epoch": 0.8872458410351202,
      "grad_norm": 2.908022403717041,
      "learning_rate": 0.000155637707948244,
      "loss": 0.0918,
      "step": 480
    },
    {
      "epoch": 0.9057301293900185,
      "grad_norm": 1.4139397144317627,
      "learning_rate": 0.00015471349353049908,
      "loss": 0.1258,
      "step": 490
    },
    {
      "epoch": 0.9242144177449169,
      "grad_norm": 0.09938838332891464,
      "learning_rate": 0.00015378927911275416,
      "loss": 0.1127,
      "step": 500
    },
    {
      "epoch": 0.9242144177449169,
      "eval_accuracy": 0.9666011787819253,
      "eval_loss": 0.08861085772514343,
      "eval_runtime": 48.3076,
      "eval_samples_per_second": 31.61,
      "eval_steps_per_second": 3.954,
      "step": 500
    },
    {
      "epoch": 0.9426987060998152,
      "grad_norm": 0.2065439671278,
      "learning_rate": 0.00015286506469500925,
      "loss": 0.0658,
      "step": 510
    },
    {
      "epoch": 0.9611829944547134,
      "grad_norm": 4.408117771148682,
      "learning_rate": 0.00015194085027726433,
      "loss": 0.0739,
      "step": 520
    },
    {
      "epoch": 0.9796672828096118,
      "grad_norm": 0.18956203758716583,
      "learning_rate": 0.0001510166358595194,
      "loss": 0.1474,
      "step": 530
    },
    {
      "epoch": 0.9981515711645101,
      "grad_norm": 2.8244571685791016,
      "learning_rate": 0.0001500924214417745,
      "loss": 0.0509,
      "step": 540
    },
    {
      "epoch": 1.0166358595194085,
      "grad_norm": 0.705922544002533,
      "learning_rate": 0.00014916820702402957,
      "loss": 0.0719,
      "step": 550
    },
    {
      "epoch": 1.0351201478743068,
      "grad_norm": 1.5707758665084839,
      "learning_rate": 0.00014824399260628468,
      "loss": 0.0954,
      "step": 560
    },
    {
      "epoch": 1.0536044362292052,
      "grad_norm": 1.2642130851745605,
      "learning_rate": 0.00014731977818853976,
      "loss": 0.0859,
      "step": 570
    },
    {
      "epoch": 1.0720887245841035,
      "grad_norm": 0.222818985581398,
      "learning_rate": 0.00014639556377079484,
      "loss": 0.0827,
      "step": 580
    },
    {
      "epoch": 1.0905730129390019,
      "grad_norm": 0.05685073137283325,
      "learning_rate": 0.00014547134935304992,
      "loss": 0.0628,
      "step": 590
    },
    {
      "epoch": 1.1090573012939002,
      "grad_norm": 0.06841599196195602,
      "learning_rate": 0.000144547134935305,
      "loss": 0.0314,
      "step": 600
    },
    {
      "epoch": 1.1090573012939002,
      "eval_accuracy": 0.9908316961362148,
      "eval_loss": 0.033296775072813034,
      "eval_runtime": 49.081,
      "eval_samples_per_second": 31.112,
      "eval_steps_per_second": 3.892,
      "step": 600
    },
    {
      "epoch": 1.1275415896487986,
      "grad_norm": 0.8085036873817444,
      "learning_rate": 0.0001436229205175601,
      "loss": 0.0295,
      "step": 610
    },
    {
      "epoch": 1.146025878003697,
      "grad_norm": 0.030383585020899773,
      "learning_rate": 0.00014269870609981517,
      "loss": 0.0168,
      "step": 620
    },
    {
      "epoch": 1.1645101663585953,
      "grad_norm": 2.179034948348999,
      "learning_rate": 0.00014177449168207025,
      "loss": 0.217,
      "step": 630
    },
    {
      "epoch": 1.1829944547134936,
      "grad_norm": 0.033746711909770966,
      "learning_rate": 0.00014085027726432533,
      "loss": 0.12,
      "step": 640
    },
    {
      "epoch": 1.201478743068392,
      "grad_norm": 0.06319836527109146,
      "learning_rate": 0.0001399260628465804,
      "loss": 0.01,
      "step": 650
    },
    {
      "epoch": 1.21996303142329,
      "grad_norm": 0.06434024125337601,
      "learning_rate": 0.0001390018484288355,
      "loss": 0.0755,
      "step": 660
    },
    {
      "epoch": 1.2384473197781884,
      "grad_norm": 3.2345378398895264,
      "learning_rate": 0.00013807763401109058,
      "loss": 0.1755,
      "step": 670
    },
    {
      "epoch": 1.2569316081330868,
      "grad_norm": 8.074490547180176,
      "learning_rate": 0.00013715341959334566,
      "loss": 0.1248,
      "step": 680
    },
    {
      "epoch": 1.2754158964879851,
      "grad_norm": 0.21135617792606354,
      "learning_rate": 0.00013622920517560074,
      "loss": 0.0662,
      "step": 690
    },
    {
      "epoch": 1.2939001848428835,
      "grad_norm": 0.0901883915066719,
      "learning_rate": 0.00013530499075785582,
      "loss": 0.0252,
      "step": 700
    },
    {
      "epoch": 1.2939001848428835,
      "eval_accuracy": 0.9973804846103471,
      "eval_loss": 0.010960910469293594,
      "eval_runtime": 49.6376,
      "eval_samples_per_second": 30.763,
      "eval_steps_per_second": 3.848,
      "step": 700
    },
    {
      "epoch": 1.3123844731977818,
      "grad_norm": 2.1113812923431396,
      "learning_rate": 0.0001343807763401109,
      "loss": 0.0685,
      "step": 710
    },
    {
      "epoch": 1.3308687615526802,
      "grad_norm": 0.062072403728961945,
      "learning_rate": 0.000133456561922366,
      "loss": 0.0747,
      "step": 720
    },
    {
      "epoch": 1.3493530499075785,
      "grad_norm": 4.824169158935547,
      "learning_rate": 0.00013253234750462106,
      "loss": 0.0791,
      "step": 730
    },
    {
      "epoch": 1.3678373382624769,
      "grad_norm": 0.12804123759269714,
      "learning_rate": 0.00013160813308687617,
      "loss": 0.045,
      "step": 740
    },
    {
      "epoch": 1.3863216266173752,
      "grad_norm": 0.3526214361190796,
      "learning_rate": 0.00013068391866913125,
      "loss": 0.0749,
      "step": 750
    },
    {
      "epoch": 1.4048059149722736,
      "grad_norm": 0.5252203345298767,
      "learning_rate": 0.00012975970425138634,
      "loss": 0.0459,
      "step": 760
    },
    {
      "epoch": 1.423290203327172,
      "grad_norm": 9.271531105041504,
      "learning_rate": 0.00012883548983364142,
      "loss": 0.1033,
      "step": 770
    },
    {
      "epoch": 1.4417744916820703,
      "grad_norm": 0.13968196511268616,
      "learning_rate": 0.0001279112754158965,
      "loss": 0.03,
      "step": 780
    },
    {
      "epoch": 1.4602587800369686,
      "grad_norm": 0.14128534495830536,
      "learning_rate": 0.00012698706099815158,
      "loss": 0.0402,
      "step": 790
    },
    {
      "epoch": 1.478743068391867,
      "grad_norm": 1.0647764205932617,
      "learning_rate": 0.00012606284658040666,
      "loss": 0.0582,
      "step": 800
    },
    {
      "epoch": 1.478743068391867,
      "eval_accuracy": 0.9986902423051736,
      "eval_loss": 0.010355145670473576,
      "eval_runtime": 48.0295,
      "eval_samples_per_second": 31.793,
      "eval_steps_per_second": 3.977,
      "step": 800
    },
    {
      "epoch": 1.4972273567467653,
      "grad_norm": 3.4088709354400635,
      "learning_rate": 0.00012513863216266174,
      "loss": 0.0495,
      "step": 810
    },
    {
      "epoch": 1.5157116451016637,
      "grad_norm": 0.2625955045223236,
      "learning_rate": 0.00012421441774491682,
      "loss": 0.0641,
      "step": 820
    },
    {
      "epoch": 1.5341959334565618,
      "grad_norm": 0.08498267084360123,
      "learning_rate": 0.0001232902033271719,
      "loss": 0.0331,
      "step": 830
    },
    {
      "epoch": 1.5526802218114604,
      "grad_norm": 10.900208473205566,
      "learning_rate": 0.000122365988909427,
      "loss": 0.0478,
      "step": 840
    },
    {
      "epoch": 1.5711645101663585,
      "grad_norm": 0.056279949843883514,
      "learning_rate": 0.00012144177449168208,
      "loss": 0.0754,
      "step": 850
    },
    {
      "epoch": 1.589648798521257,
      "grad_norm": 0.05696770176291466,
      "learning_rate": 0.00012051756007393715,
      "loss": 0.0906,
      "step": 860
    },
    {
      "epoch": 1.6081330868761552,
      "grad_norm": 0.022501414641737938,
      "learning_rate": 0.00011959334565619225,
      "loss": 0.0413,
      "step": 870
    },
    {
      "epoch": 1.6266173752310538,
      "grad_norm": 2.3988966941833496,
      "learning_rate": 0.00011866913123844731,
      "loss": 0.074,
      "step": 880
    },
    {
      "epoch": 1.645101663585952,
      "grad_norm": 0.03616836294531822,
      "learning_rate": 0.00011774491682070241,
      "loss": 0.0094,
      "step": 890
    },
    {
      "epoch": 1.6635859519408502,
      "grad_norm": 9.183988571166992,
      "learning_rate": 0.00011682070240295748,
      "loss": 0.0455,
      "step": 900
    },
    {
      "epoch": 1.6635859519408502,
      "eval_accuracy": 0.9954158480681073,
      "eval_loss": 0.01981273666024208,
      "eval_runtime": 48.1294,
      "eval_samples_per_second": 31.727,
      "eval_steps_per_second": 3.968,
      "step": 900
    },
    {
      "epoch": 1.6820702402957486,
      "grad_norm": 20.508174896240234,
      "learning_rate": 0.00011589648798521257,
      "loss": 0.1063,
      "step": 910
    },
    {
      "epoch": 1.700554528650647,
      "grad_norm": 3.3739655017852783,
      "learning_rate": 0.00011497227356746765,
      "loss": 0.0711,
      "step": 920
    },
    {
      "epoch": 1.7190388170055453,
      "grad_norm": 0.03310403227806091,
      "learning_rate": 0.00011404805914972275,
      "loss": 0.0216,
      "step": 930
    },
    {
      "epoch": 1.7375231053604436,
      "grad_norm": 0.027661019936203957,
      "learning_rate": 0.00011312384473197783,
      "loss": 0.05,
      "step": 940
    },
    {
      "epoch": 1.756007393715342,
      "grad_norm": 0.02968147210776806,
      "learning_rate": 0.00011219963031423291,
      "loss": 0.0368,
      "step": 950
    },
    {
      "epoch": 1.7744916820702403,
      "grad_norm": 0.01928904838860035,
      "learning_rate": 0.000111275415896488,
      "loss": 0.0083,
      "step": 960
    },
    {
      "epoch": 1.7929759704251387,
      "grad_norm": 0.048537448048591614,
      "learning_rate": 0.00011035120147874307,
      "loss": 0.0421,
      "step": 970
    },
    {
      "epoch": 1.8114602587800368,
      "grad_norm": 0.01963023841381073,
      "learning_rate": 0.00010942698706099817,
      "loss": 0.0059,
      "step": 980
    },
    {
      "epoch": 1.8299445471349354,
      "grad_norm": 0.035353146493434906,
      "learning_rate": 0.00010850277264325324,
      "loss": 0.0446,
      "step": 990
    },
    {
      "epoch": 1.8484288354898335,
      "grad_norm": 0.015135395340621471,
      "learning_rate": 0.00010757855822550833,
      "loss": 0.0569,
      "step": 1000
    },
    {
      "epoch": 1.8484288354898335,
      "eval_accuracy": 0.9960707269155207,
      "eval_loss": 0.017953284084796906,
      "eval_runtime": 49.2028,
      "eval_samples_per_second": 31.035,
      "eval_steps_per_second": 3.882,
      "step": 1000
    },
    {
      "epoch": 1.866913123844732,
      "grad_norm": 0.02287949249148369,
      "learning_rate": 0.0001066543438077634,
      "loss": 0.0033,
      "step": 1010
    },
    {
      "epoch": 1.8853974121996302,
      "grad_norm": 0.024308230727910995,
      "learning_rate": 0.0001057301293900185,
      "loss": 0.0028,
      "step": 1020
    },
    {
      "epoch": 1.9038817005545288,
      "grad_norm": 8.346418380737305,
      "learning_rate": 0.00010480591497227356,
      "loss": 0.0468,
      "step": 1030
    },
    {
      "epoch": 1.922365988909427,
      "grad_norm": 0.015379060991108418,
      "learning_rate": 0.00010388170055452866,
      "loss": 0.1142,
      "step": 1040
    },
    {
      "epoch": 1.9408502772643255,
      "grad_norm": 0.021259386092424393,
      "learning_rate": 0.00010295748613678373,
      "loss": 0.057,
      "step": 1050
    },
    {
      "epoch": 1.9593345656192236,
      "grad_norm": 4.173085689544678,
      "learning_rate": 0.00010203327171903882,
      "loss": 0.0333,
      "step": 1060
    },
    {
      "epoch": 1.9778188539741222,
      "grad_norm": 0.013770132325589657,
      "learning_rate": 0.0001011090573012939,
      "loss": 0.0248,
      "step": 1070
    },
    {
      "epoch": 1.9963031423290203,
      "grad_norm": 0.029461657628417015,
      "learning_rate": 0.000100184842883549,
      "loss": 0.1002,
      "step": 1080
    },
    {
      "epoch": 2.014787430683919,
      "grad_norm": 0.042121052742004395,
      "learning_rate": 9.926062846580408e-05,
      "loss": 0.0499,
      "step": 1090
    },
    {
      "epoch": 2.033271719038817,
      "grad_norm": 6.605316162109375,
      "learning_rate": 9.833641404805916e-05,
      "loss": 0.0627,
      "step": 1100
    },
    {
      "epoch": 2.033271719038817,
      "eval_accuracy": 0.9947609692206941,
      "eval_loss": 0.024407994002103806,
      "eval_runtime": 48.147,
      "eval_samples_per_second": 31.715,
      "eval_steps_per_second": 3.967,
      "step": 1100
    },
    {
      "epoch": 2.0517560073937156,
      "grad_norm": 0.03806009516119957,
      "learning_rate": 9.741219963031424e-05,
      "loss": 0.0162,
      "step": 1110
    },
    {
      "epoch": 2.0702402957486137,
      "grad_norm": 0.04799278452992439,
      "learning_rate": 9.648798521256932e-05,
      "loss": 0.0083,
      "step": 1120
    },
    {
      "epoch": 2.088724584103512,
      "grad_norm": 3.807424545288086,
      "learning_rate": 9.55637707948244e-05,
      "loss": 0.0349,
      "step": 1130
    },
    {
      "epoch": 2.1072088724584104,
      "grad_norm": 0.013640133664011955,
      "learning_rate": 9.463955637707949e-05,
      "loss": 0.0079,
      "step": 1140
    },
    {
      "epoch": 2.1256931608133085,
      "grad_norm": 0.9286332726478577,
      "learning_rate": 9.371534195933457e-05,
      "loss": 0.0675,
      "step": 1150
    },
    {
      "epoch": 2.144177449168207,
      "grad_norm": 7.584729194641113,
      "learning_rate": 9.279112754158965e-05,
      "loss": 0.0223,
      "step": 1160
    },
    {
      "epoch": 2.162661737523105,
      "grad_norm": 0.02116883546113968,
      "learning_rate": 9.186691312384473e-05,
      "loss": 0.0046,
      "step": 1170
    },
    {
      "epoch": 2.1811460258780038,
      "grad_norm": 0.014812180772423744,
      "learning_rate": 9.094269870609981e-05,
      "loss": 0.0194,
      "step": 1180
    },
    {
      "epoch": 2.199630314232902,
      "grad_norm": 0.011532890610396862,
      "learning_rate": 9.001848428835489e-05,
      "loss": 0.0236,
      "step": 1190
    },
    {
      "epoch": 2.2181146025878005,
      "grad_norm": 0.01616760343313217,
      "learning_rate": 8.909426987060999e-05,
      "loss": 0.0328,
      "step": 1200
    },
    {
      "epoch": 2.2181146025878005,
      "eval_accuracy": 0.9986902423051736,
      "eval_loss": 0.0054244487546384335,
      "eval_runtime": 49.4505,
      "eval_samples_per_second": 30.879,
      "eval_steps_per_second": 3.862,
      "step": 1200
    },
    {
      "epoch": 2.2365988909426986,
      "grad_norm": 0.04048744961619377,
      "learning_rate": 8.817005545286507e-05,
      "loss": 0.0329,
      "step": 1210
    },
    {
      "epoch": 2.255083179297597,
      "grad_norm": 0.0121224420145154,
      "learning_rate": 8.724584103512015e-05,
      "loss": 0.0229,
      "step": 1220
    },
    {
      "epoch": 2.2735674676524953,
      "grad_norm": 3.312351942062378,
      "learning_rate": 8.632162661737525e-05,
      "loss": 0.0337,
      "step": 1230
    },
    {
      "epoch": 2.292051756007394,
      "grad_norm": 0.04162617400288582,
      "learning_rate": 8.539741219963033e-05,
      "loss": 0.0562,
      "step": 1240
    },
    {
      "epoch": 2.310536044362292,
      "grad_norm": 0.0170124601572752,
      "learning_rate": 8.447319778188541e-05,
      "loss": 0.0048,
      "step": 1250
    },
    {
      "epoch": 2.3290203327171906,
      "grad_norm": 0.03438381850719452,
      "learning_rate": 8.354898336414049e-05,
      "loss": 0.0305,
      "step": 1260
    },
    {
      "epoch": 2.3475046210720887,
      "grad_norm": 4.124629974365234,
      "learning_rate": 8.262476894639557e-05,
      "loss": 0.0413,
      "step": 1270
    },
    {
      "epoch": 2.3659889094269873,
      "grad_norm": 0.009506890550255775,
      "learning_rate": 8.170055452865065e-05,
      "loss": 0.0775,
      "step": 1280
    },
    {
      "epoch": 2.3844731977818854,
      "grad_norm": 0.009906689636409283,
      "learning_rate": 8.077634011090573e-05,
      "loss": 0.0432,
      "step": 1290
    },
    {
      "epoch": 2.402957486136784,
      "grad_norm": 0.0475531667470932,
      "learning_rate": 7.985212569316082e-05,
      "loss": 0.0156,
      "step": 1300
    },
    {
      "epoch": 2.402957486136784,
      "eval_accuracy": 0.9947609692206941,
      "eval_loss": 0.0192621648311615,
      "eval_runtime": 48.3339,
      "eval_samples_per_second": 31.593,
      "eval_steps_per_second": 3.952,
      "step": 1300
    },
    {
      "epoch": 2.421441774491682,
      "grad_norm": 0.05756278708577156,
      "learning_rate": 7.89279112754159e-05,
      "loss": 0.0771,
      "step": 1310
    },
    {
      "epoch": 2.43992606284658,
      "grad_norm": 0.021766463294625282,
      "learning_rate": 7.800369685767098e-05,
      "loss": 0.0661,
      "step": 1320
    },
    {
      "epoch": 2.4584103512014788,
      "grad_norm": 0.02948659099638462,
      "learning_rate": 7.707948243992606e-05,
      "loss": 0.0166,
      "step": 1330
    },
    {
      "epoch": 2.476894639556377,
      "grad_norm": 9.3042573928833,
      "learning_rate": 7.615526802218114e-05,
      "loss": 0.0309,
      "step": 1340
    },
    {
      "epoch": 2.4953789279112755,
      "grad_norm": 0.023814814165234566,
      "learning_rate": 7.523105360443624e-05,
      "loss": 0.0783,
      "step": 1350
    },
    {
      "epoch": 2.5138632162661736,
      "grad_norm": 3.380216598510742,
      "learning_rate": 7.430683918669132e-05,
      "loss": 0.0556,
      "step": 1360
    },
    {
      "epoch": 2.532347504621072,
      "grad_norm": 0.0165497325360775,
      "learning_rate": 7.33826247689464e-05,
      "loss": 0.0221,
      "step": 1370
    },
    {
      "epoch": 2.5508317929759703,
      "grad_norm": 0.19721367955207825,
      "learning_rate": 7.245841035120148e-05,
      "loss": 0.0147,
      "step": 1380
    },
    {
      "epoch": 2.569316081330869,
      "grad_norm": 7.320788383483887,
      "learning_rate": 7.153419593345656e-05,
      "loss": 0.0251,
      "step": 1390
    },
    {
      "epoch": 2.587800369685767,
      "grad_norm": 0.009770031087100506,
      "learning_rate": 7.060998151571166e-05,
      "loss": 0.0016,
      "step": 1400
    },
    {
      "epoch": 2.587800369685767,
      "eval_accuracy": 0.9973804846103471,
      "eval_loss": 0.007392051629722118,
      "eval_runtime": 49.7571,
      "eval_samples_per_second": 30.689,
      "eval_steps_per_second": 3.839,
      "step": 1400
    },
    {
      "epoch": 2.6062846580406656,
      "grad_norm": 0.009464599192142487,
      "learning_rate": 6.968576709796674e-05,
      "loss": 0.0235,
      "step": 1410
    },
    {
      "epoch": 2.6247689463955637,
      "grad_norm": 3.2390127182006836,
      "learning_rate": 6.876155268022182e-05,
      "loss": 0.0419,
      "step": 1420
    },
    {
      "epoch": 2.6432532347504623,
      "grad_norm": 0.015608754009008408,
      "learning_rate": 6.78373382624769e-05,
      "loss": 0.0043,
      "step": 1430
    },
    {
      "epoch": 2.6617375231053604,
      "grad_norm": 0.35069820284843445,
      "learning_rate": 6.691312384473198e-05,
      "loss": 0.0079,
      "step": 1440
    },
    {
      "epoch": 2.6802218114602585,
      "grad_norm": 0.012105106376111507,
      "learning_rate": 6.598890942698706e-05,
      "loss": 0.0501,
      "step": 1450
    },
    {
      "epoch": 2.698706099815157,
      "grad_norm": 0.06180037185549736,
      "learning_rate": 6.506469500924215e-05,
      "loss": 0.0282,
      "step": 1460
    },
    {
      "epoch": 2.7171903881700556,
      "grad_norm": 0.008026616647839546,
      "learning_rate": 6.414048059149723e-05,
      "loss": 0.0033,
      "step": 1470
    },
    {
      "epoch": 2.7356746765249538,
      "grad_norm": 0.4511905610561371,
      "learning_rate": 6.321626617375231e-05,
      "loss": 0.0269,
      "step": 1480
    },
    {
      "epoch": 2.754158964879852,
      "grad_norm": 0.05588386207818985,
      "learning_rate": 6.229205175600739e-05,
      "loss": 0.0051,
      "step": 1490
    },
    {
      "epoch": 2.7726432532347505,
      "grad_norm": 0.025905737653374672,
      "learning_rate": 6.136783733826249e-05,
      "loss": 0.0032,
      "step": 1500
    },
    {
      "epoch": 2.7726432532347505,
      "eval_accuracy": 0.9980353634577603,
      "eval_loss": 0.004458332899957895,
      "eval_runtime": 48.7133,
      "eval_samples_per_second": 31.347,
      "eval_steps_per_second": 3.921,
      "step": 1500
    },
    {
      "epoch": 2.791127541589649,
      "grad_norm": 0.007365287281572819,
      "learning_rate": 6.044362292051756e-05,
      "loss": 0.0238,
      "step": 1510
    },
    {
      "epoch": 2.809611829944547,
      "grad_norm": 0.0069998023100197315,
      "learning_rate": 5.951940850277264e-05,
      "loss": 0.0382,
      "step": 1520
    },
    {
      "epoch": 2.8280961182994453,
      "grad_norm": 2.074650287628174,
      "learning_rate": 5.859519408502773e-05,
      "loss": 0.052,
      "step": 1530
    },
    {
      "epoch": 2.846580406654344,
      "grad_norm": 0.02161816880106926,
      "learning_rate": 5.767097966728281e-05,
      "loss": 0.0027,
      "step": 1540
    },
    {
      "epoch": 2.865064695009242,
      "grad_norm": 0.008178635500371456,
      "learning_rate": 5.674676524953789e-05,
      "loss": 0.0529,
      "step": 1550
    },
    {
      "epoch": 2.8835489833641406,
      "grad_norm": 0.010601267218589783,
      "learning_rate": 5.5822550831792974e-05,
      "loss": 0.0018,
      "step": 1560
    },
    {
      "epoch": 2.9020332717190387,
      "grad_norm": 0.009498382918536663,
      "learning_rate": 5.4898336414048056e-05,
      "loss": 0.0232,
      "step": 1570
    },
    {
      "epoch": 2.9205175600739373,
      "grad_norm": 0.012483976781368256,
      "learning_rate": 5.397412199630314e-05,
      "loss": 0.0016,
      "step": 1580
    },
    {
      "epoch": 2.9390018484288354,
      "grad_norm": 0.09681292623281479,
      "learning_rate": 5.304990757855823e-05,
      "loss": 0.0156,
      "step": 1590
    },
    {
      "epoch": 2.957486136783734,
      "grad_norm": 0.08324334770441055,
      "learning_rate": 5.2125693160813314e-05,
      "loss": 0.0233,
      "step": 1600
    },
    {
      "epoch": 2.957486136783734,
      "eval_accuracy": 0.9993451211525868,
      "eval_loss": 0.002875428879633546,
      "eval_runtime": 49.5419,
      "eval_samples_per_second": 30.822,
      "eval_steps_per_second": 3.855,
      "step": 1600
    },
    {
      "epoch": 2.975970425138632,
      "grad_norm": 0.00964986253529787,
      "learning_rate": 5.1201478743068395e-05,
      "loss": 0.0189,
      "step": 1610
    },
    {
      "epoch": 2.9944547134935307,
      "grad_norm": 0.023254286497831345,
      "learning_rate": 5.027726432532348e-05,
      "loss": 0.0016,
      "step": 1620
    },
    {
      "epoch": 3.0129390018484288,
      "grad_norm": 0.03303303197026253,
      "learning_rate": 4.935304990757856e-05,
      "loss": 0.0421,
      "step": 1630
    },
    {
      "epoch": 3.0314232902033273,
      "grad_norm": 0.05670865997672081,
      "learning_rate": 4.8428835489833646e-05,
      "loss": 0.0164,
      "step": 1640
    },
    {
      "epoch": 3.0499075785582255,
      "grad_norm": 0.6677367091178894,
      "learning_rate": 4.750462107208873e-05,
      "loss": 0.0021,
      "step": 1650
    },
    {
      "epoch": 3.068391866913124,
      "grad_norm": 0.07437073439359665,
      "learning_rate": 4.658040665434381e-05,
      "loss": 0.0311,
      "step": 1660
    },
    {
      "epoch": 3.086876155268022,
      "grad_norm": 0.011145181953907013,
      "learning_rate": 4.565619223659889e-05,
      "loss": 0.0021,
      "step": 1670
    },
    {
      "epoch": 3.1053604436229207,
      "grad_norm": 0.3542933464050293,
      "learning_rate": 4.473197781885398e-05,
      "loss": 0.0024,
      "step": 1680
    },
    {
      "epoch": 3.123844731977819,
      "grad_norm": 0.012382660992443562,
      "learning_rate": 4.380776340110906e-05,
      "loss": 0.0046,
      "step": 1690
    },
    {
      "epoch": 3.142329020332717,
      "grad_norm": 0.0069410833530128,
      "learning_rate": 4.288354898336414e-05,
      "loss": 0.0434,
      "step": 1700
    },
    {
      "epoch": 3.142329020332717,
      "eval_accuracy": 0.9993451211525868,
      "eval_loss": 0.0025608371943235397,
      "eval_runtime": 48.733,
      "eval_samples_per_second": 31.334,
      "eval_steps_per_second": 3.919,
      "step": 1700
    },
    {
      "epoch": 3.1608133086876156,
      "grad_norm": 0.3573768436908722,
      "learning_rate": 4.195933456561922e-05,
      "loss": 0.0014,
      "step": 1710
    },
    {
      "epoch": 3.1792975970425137,
      "grad_norm": 0.009119544178247452,
      "learning_rate": 4.1035120147874305e-05,
      "loss": 0.0009,
      "step": 1720
    },
    {
      "epoch": 3.1977818853974123,
      "grad_norm": 0.02414746955037117,
      "learning_rate": 4.011090573012939e-05,
      "loss": 0.0009,
      "step": 1730
    },
    {
      "epoch": 3.2162661737523104,
      "grad_norm": 0.006854603998363018,
      "learning_rate": 3.9186691312384474e-05,
      "loss": 0.0396,
      "step": 1740
    },
    {
      "epoch": 3.234750462107209,
      "grad_norm": 0.010423395782709122,
      "learning_rate": 3.826247689463956e-05,
      "loss": 0.0375,
      "step": 1750
    },
    {
      "epoch": 3.253234750462107,
      "grad_norm": 0.010691582225263119,
      "learning_rate": 3.7338262476894644e-05,
      "loss": 0.0121,
      "step": 1760
    },
    {
      "epoch": 3.2717190388170057,
      "grad_norm": 0.1874719113111496,
      "learning_rate": 3.6414048059149726e-05,
      "loss": 0.0019,
      "step": 1770
    },
    {
      "epoch": 3.290203327171904,
      "grad_norm": 0.14755026996135712,
      "learning_rate": 3.548983364140481e-05,
      "loss": 0.0115,
      "step": 1780
    },
    {
      "epoch": 3.3086876155268024,
      "grad_norm": 0.024332020431756973,
      "learning_rate": 3.456561922365989e-05,
      "loss": 0.0062,
      "step": 1790
    },
    {
      "epoch": 3.3271719038817005,
      "grad_norm": 0.007643954362720251,
      "learning_rate": 3.364140480591497e-05,
      "loss": 0.0079,
      "step": 1800
    },
    {
      "epoch": 3.3271719038817005,
      "eval_accuracy": 0.9980353634577603,
      "eval_loss": 0.009474328719079494,
      "eval_runtime": 48.3759,
      "eval_samples_per_second": 31.565,
      "eval_steps_per_second": 3.948,
      "step": 1800
    },
    {
      "epoch": 3.345656192236599,
      "grad_norm": 0.008659069426357746,
      "learning_rate": 3.271719038817006e-05,
      "loss": 0.0012,
      "step": 1810
    },
    {
      "epoch": 3.364140480591497,
      "grad_norm": 0.11470120400190353,
      "learning_rate": 3.179297597042514e-05,
      "loss": 0.0026,
      "step": 1820
    },
    {
      "epoch": 3.3826247689463957,
      "grad_norm": 9.137892723083496,
      "learning_rate": 3.086876155268023e-05,
      "loss": 0.0298,
      "step": 1830
    },
    {
      "epoch": 3.401109057301294,
      "grad_norm": 0.01612740568816662,
      "learning_rate": 2.994454713493531e-05,
      "loss": 0.0405,
      "step": 1840
    },
    {
      "epoch": 3.4195933456561924,
      "grad_norm": 0.01397150382399559,
      "learning_rate": 2.902033271719039e-05,
      "loss": 0.0018,
      "step": 1850
    },
    {
      "epoch": 3.4380776340110906,
      "grad_norm": 0.0072495522908866405,
      "learning_rate": 2.8096118299445472e-05,
      "loss": 0.0008,
      "step": 1860
    },
    {
      "epoch": 3.4565619223659887,
      "grad_norm": 0.009662143886089325,
      "learning_rate": 2.7171903881700557e-05,
      "loss": 0.0484,
      "step": 1870
    },
    {
      "epoch": 3.4750462107208873,
      "grad_norm": 0.005946171935647726,
      "learning_rate": 2.624768946395564e-05,
      "loss": 0.0014,
      "step": 1880
    },
    {
      "epoch": 3.4935304990757854,
      "grad_norm": 0.010195981711149216,
      "learning_rate": 2.532347504621072e-05,
      "loss": 0.0479,
      "step": 1890
    },
    {
      "epoch": 3.512014787430684,
      "grad_norm": 0.02414027974009514,
      "learning_rate": 2.4399260628465805e-05,
      "loss": 0.0175,
      "step": 1900
    },
    {
      "epoch": 3.512014787430684,
      "eval_accuracy": 0.9973804846103471,
      "eval_loss": 0.011128582060337067,
      "eval_runtime": 48.1238,
      "eval_samples_per_second": 31.731,
      "eval_steps_per_second": 3.969,
      "step": 1900
    },
    {
      "epoch": 3.530499075785582,
      "grad_norm": 0.028868084773421288,
      "learning_rate": 2.347504621072089e-05,
      "loss": 0.001,
      "step": 1910
    },
    {
      "epoch": 3.5489833641404807,
      "grad_norm": 0.005811081267893314,
      "learning_rate": 2.255083179297597e-05,
      "loss": 0.0204,
      "step": 1920
    },
    {
      "epoch": 3.567467652495379,
      "grad_norm": 0.6555341482162476,
      "learning_rate": 2.1626617375231053e-05,
      "loss": 0.04,
      "step": 1930
    },
    {
      "epoch": 3.5859519408502774,
      "grad_norm": 0.006032618228346109,
      "learning_rate": 2.0702402957486137e-05,
      "loss": 0.0096,
      "step": 1940
    },
    {
      "epoch": 3.6044362292051755,
      "grad_norm": 0.006055204197764397,
      "learning_rate": 1.9778188539741222e-05,
      "loss": 0.0075,
      "step": 1950
    },
    {
      "epoch": 3.622920517560074,
      "grad_norm": 0.010445200838148594,
      "learning_rate": 1.8853974121996304e-05,
      "loss": 0.0042,
      "step": 1960
    },
    {
      "epoch": 3.641404805914972,
      "grad_norm": 0.02863822504878044,
      "learning_rate": 1.7929759704251385e-05,
      "loss": 0.001,
      "step": 1970
    },
    {
      "epoch": 3.6598890942698707,
      "grad_norm": 0.007255155127495527,
      "learning_rate": 1.700554528650647e-05,
      "loss": 0.0015,
      "step": 1980
    },
    {
      "epoch": 3.678373382624769,
      "grad_norm": 0.012167246080935001,
      "learning_rate": 1.6081330868761555e-05,
      "loss": 0.0185,
      "step": 1990
    },
    {
      "epoch": 3.6968576709796674,
      "grad_norm": 0.008204938843846321,
      "learning_rate": 1.5157116451016636e-05,
      "loss": 0.0013,
      "step": 2000
    },
    {
      "epoch": 3.6968576709796674,
      "eval_accuracy": 0.9973804846103471,
      "eval_loss": 0.010935045778751373,
      "eval_runtime": 47.9657,
      "eval_samples_per_second": 31.835,
      "eval_steps_per_second": 3.982,
      "step": 2000
    },
    {
      "epoch": 3.7153419593345656,
      "grad_norm": 0.0491810217499733,
      "learning_rate": 1.423290203327172e-05,
      "loss": 0.0011,
      "step": 2010
    },
    {
      "epoch": 3.733826247689464,
      "grad_norm": 0.006103880237787962,
      "learning_rate": 1.3308687615526803e-05,
      "loss": 0.0008,
      "step": 2020
    },
    {
      "epoch": 3.7523105360443623,
      "grad_norm": 0.008025890216231346,
      "learning_rate": 1.2384473197781886e-05,
      "loss": 0.0102,
      "step": 2030
    },
    {
      "epoch": 3.7707948243992604,
      "grad_norm": 0.005999017972499132,
      "learning_rate": 1.1460258780036969e-05,
      "loss": 0.0007,
      "step": 2040
    },
    {
      "epoch": 3.789279112754159,
      "grad_norm": 0.030818577855825424,
      "learning_rate": 1.0536044362292052e-05,
      "loss": 0.0088,
      "step": 2050
    },
    {
      "epoch": 3.8077634011090575,
      "grad_norm": 0.005371921230107546,
      "learning_rate": 9.611829944547135e-06,
      "loss": 0.0007,
      "step": 2060
    },
    {
      "epoch": 3.8262476894639557,
      "grad_norm": 0.005443415604531765,
      "learning_rate": 8.687615526802218e-06,
      "loss": 0.0015,
      "step": 2070
    },
    {
      "epoch": 3.844731977818854,
      "grad_norm": 0.005951932165771723,
      "learning_rate": 7.763401109057302e-06,
      "loss": 0.0008,
      "step": 2080
    },
    {
      "epoch": 3.8632162661737524,
      "grad_norm": 0.006055652163922787,
      "learning_rate": 6.8391866913123855e-06,
      "loss": 0.0009,
      "step": 2090
    },
    {
      "epoch": 3.8817005545286505,
      "grad_norm": 0.005791305564343929,
      "learning_rate": 5.914972273567468e-06,
      "loss": 0.0008,
      "step": 2100
    },
    {
      "epoch": 3.8817005545286505,
      "eval_accuracy": 0.9986902423051736,
      "eval_loss": 0.005253881681710482,
      "eval_runtime": 48.3848,
      "eval_samples_per_second": 31.56,
      "eval_steps_per_second": 3.948,
      "step": 2100
    },
    {
      "epoch": 3.900184842883549,
      "grad_norm": 0.011120261624455452,
      "learning_rate": 4.990757855822551e-06,
      "loss": 0.0023,
      "step": 2110
    },
    {
      "epoch": 3.918669131238447,
      "grad_norm": 0.005558283068239689,
      "learning_rate": 4.066543438077634e-06,
      "loss": 0.0143,
      "step": 2120
    },
    {
      "epoch": 3.9371534195933457,
      "grad_norm": 0.005820298567414284,
      "learning_rate": 3.1423290203327177e-06,
      "loss": 0.0016,
      "step": 2130
    },
    {
      "epoch": 3.955637707948244,
      "grad_norm": 0.006517091300338507,
      "learning_rate": 2.2181146025878004e-06,
      "loss": 0.0015,
      "step": 2140
    },
    {
      "epoch": 3.9741219963031424,
      "grad_norm": 0.08097829669713974,
      "learning_rate": 1.2939001848428836e-06,
      "loss": 0.0379,
      "step": 2150
    },
    {
      "epoch": 3.9926062846580406,
      "grad_norm": 0.0057843709364533424,
      "learning_rate": 3.6968576709796674e-07,
      "loss": 0.0008,
      "step": 2160
    },
    {
      "epoch": 4.0,
      "step": 2164,
      "total_flos": 2.6818427765818e+18,
      "train_loss": 0.07006576062802418,
      "train_runtime": 2504.6639,
      "train_samples_per_second": 13.817,
      "train_steps_per_second": 0.864
    }
  ],
  "logging_steps": 10,
  "max_steps": 2164,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.6818427765818e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}