{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 9.760827407886232,
  "eval_steps": 25,
  "global_step": 3775,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06464124111182935,
      "grad_norm": 0.4801495671272278,
      "learning_rate": 0.0001987561544441565,
      "loss": 0.6331,
      "step": 25
    },
    {
      "epoch": 0.06464124111182935,
      "eval_loss": 0.614825427532196,
      "eval_runtime": 13.3629,
      "eval_samples_per_second": 14.518,
      "eval_steps_per_second": 1.871,
      "step": 25
    },
    {
      "epoch": 0.1292824822236587,
      "grad_norm": 0.5615572333335876,
      "learning_rate": 0.0001974604819901529,
      "loss": 0.4891,
      "step": 50
    },
    {
      "epoch": 0.1292824822236587,
      "eval_loss": 0.6148936748504639,
      "eval_runtime": 13.3296,
      "eval_samples_per_second": 14.554,
      "eval_steps_per_second": 1.876,
      "step": 50
    },
    {
      "epoch": 0.19392372333548805,
      "grad_norm": 0.5219907164573669,
      "learning_rate": 0.00019616480953614927,
      "loss": 0.6005,
      "step": 75
    },
    {
      "epoch": 0.19392372333548805,
      "eval_loss": 0.5868657827377319,
      "eval_runtime": 13.3229,
      "eval_samples_per_second": 14.561,
      "eval_steps_per_second": 1.876,
      "step": 75
    },
    {
      "epoch": 0.2585649644473174,
      "grad_norm": 0.38672706484794617,
      "learning_rate": 0.00019486913708214565,
      "loss": 0.4458,
      "step": 100
    },
    {
      "epoch": 0.2585649644473174,
      "eval_loss": 0.5840005874633789,
      "eval_runtime": 13.3283,
      "eval_samples_per_second": 14.556,
      "eval_steps_per_second": 1.876,
      "step": 100
    },
    {
      "epoch": 0.32320620555914675,
      "grad_norm": 0.28140130639076233,
      "learning_rate": 0.000193573464628142,
      "loss": 0.5595,
      "step": 125
    },
    {
      "epoch": 0.32320620555914675,
      "eval_loss": 0.5389835834503174,
      "eval_runtime": 13.3413,
      "eval_samples_per_second": 14.541,
      "eval_steps_per_second": 1.874,
      "step": 125
    },
    {
      "epoch": 0.3878474466709761,
      "grad_norm": 0.4100258946418762,
      "learning_rate": 0.00019227779217413838,
      "loss": 0.3954,
      "step": 150
    },
    {
      "epoch": 0.3878474466709761,
      "eval_loss": 0.5461440086364746,
      "eval_runtime": 13.3424,
      "eval_samples_per_second": 14.54,
      "eval_steps_per_second": 1.874,
      "step": 150
    },
    {
      "epoch": 0.45248868778280543,
      "grad_norm": 0.443906307220459,
      "learning_rate": 0.00019098211972013476,
      "loss": 0.5271,
      "step": 175
    },
    {
      "epoch": 0.45248868778280543,
      "eval_loss": 0.511994481086731,
      "eval_runtime": 13.3406,
      "eval_samples_per_second": 14.542,
      "eval_steps_per_second": 1.874,
      "step": 175
    },
    {
      "epoch": 0.5171299288946348,
      "grad_norm": 0.45473217964172363,
      "learning_rate": 0.00018968644726613114,
      "loss": 0.3711,
      "step": 200
    },
    {
      "epoch": 0.5171299288946348,
      "eval_loss": 0.5170156359672546,
      "eval_runtime": 13.3485,
      "eval_samples_per_second": 14.533,
      "eval_steps_per_second": 1.873,
      "step": 200
    },
    {
      "epoch": 0.5817711700064642,
      "grad_norm": 0.2725232243537903,
      "learning_rate": 0.0001883907748121275,
      "loss": 0.485,
      "step": 225
    },
    {
      "epoch": 0.5817711700064642,
      "eval_loss": 0.48935818672180176,
      "eval_runtime": 13.3325,
      "eval_samples_per_second": 14.551,
      "eval_steps_per_second": 1.875,
      "step": 225
    },
    {
      "epoch": 0.6464124111182935,
      "grad_norm": 0.37662985920906067,
      "learning_rate": 0.00018709510235812387,
      "loss": 0.3509,
      "step": 250
    },
    {
      "epoch": 0.6464124111182935,
      "eval_loss": 0.49390536546707153,
      "eval_runtime": 13.3404,
      "eval_samples_per_second": 14.542,
      "eval_steps_per_second": 1.874,
      "step": 250
    },
    {
      "epoch": 0.7110536522301228,
      "grad_norm": 0.37969282269477844,
      "learning_rate": 0.00018579942990412025,
      "loss": 0.4957,
      "step": 275
    },
    {
      "epoch": 0.7110536522301228,
      "eval_loss": 0.4695137143135071,
      "eval_runtime": 13.3432,
      "eval_samples_per_second": 14.539,
      "eval_steps_per_second": 1.874,
      "step": 275
    },
    {
      "epoch": 0.7756948933419522,
      "grad_norm": 0.47988465428352356,
      "learning_rate": 0.0001845037574501166,
      "loss": 0.3239,
      "step": 300
    },
    {
      "epoch": 0.7756948933419522,
      "eval_loss": 0.4713852107524872,
      "eval_runtime": 13.3412,
      "eval_samples_per_second": 14.541,
      "eval_steps_per_second": 1.874,
      "step": 300
    },
    {
      "epoch": 0.8403361344537815,
      "grad_norm": 0.48248979449272156,
      "learning_rate": 0.00018320808499611298,
      "loss": 0.4371,
      "step": 325
    },
    {
      "epoch": 0.8403361344537815,
      "eval_loss": 0.4440518915653229,
      "eval_runtime": 13.3429,
      "eval_samples_per_second": 14.54,
      "eval_steps_per_second": 1.874,
      "step": 325
    },
    {
      "epoch": 0.9049773755656109,
      "grad_norm": 0.7585775256156921,
      "learning_rate": 0.00018191241254210936,
      "loss": 0.2987,
      "step": 350
    },
    {
      "epoch": 0.9049773755656109,
      "eval_loss": 0.45601731538772583,
      "eval_runtime": 13.3413,
      "eval_samples_per_second": 14.541,
      "eval_steps_per_second": 1.874,
      "step": 350
    },
    {
      "epoch": 0.9696186166774402,
      "grad_norm": 0.35234636068344116,
      "learning_rate": 0.00018061674008810574,
      "loss": 0.4198,
      "step": 375
    },
    {
      "epoch": 0.9696186166774402,
      "eval_loss": 0.4415298402309418,
      "eval_runtime": 13.3427,
      "eval_samples_per_second": 14.54,
      "eval_steps_per_second": 1.874,
      "step": 375
    },
    {
      "epoch": 1.0342598577892697,
      "grad_norm": 0.4931798577308655,
      "learning_rate": 0.0001793210676341021,
      "loss": 0.4064,
      "step": 400
    },
    {
      "epoch": 1.0342598577892697,
      "eval_loss": 0.42522725462913513,
      "eval_runtime": 13.3548,
      "eval_samples_per_second": 14.527,
      "eval_steps_per_second": 1.872,
      "step": 400
    },
    {
      "epoch": 1.098901098901099,
      "grad_norm": 0.5019311308860779,
      "learning_rate": 0.00017802539518009847,
      "loss": 0.2946,
      "step": 425
    },
    {
      "epoch": 1.098901098901099,
      "eval_loss": 0.4273912012577057,
      "eval_runtime": 13.3417,
      "eval_samples_per_second": 14.541,
      "eval_steps_per_second": 1.874,
      "step": 425
    },
    {
      "epoch": 1.1635423400129283,
      "grad_norm": 0.5820428133010864,
      "learning_rate": 0.00017672972272609485,
      "loss": 0.3152,
      "step": 450
    },
    {
      "epoch": 1.1635423400129283,
      "eval_loss": 0.4081675410270691,
      "eval_runtime": 13.3377,
      "eval_samples_per_second": 14.545,
      "eval_steps_per_second": 1.874,
      "step": 450
    },
    {
      "epoch": 1.2281835811247577,
      "grad_norm": 0.5499022006988525,
      "learning_rate": 0.00017543405027209123,
      "loss": 0.2964,
      "step": 475
    },
    {
      "epoch": 1.2281835811247577,
      "eval_loss": 0.41611921787261963,
      "eval_runtime": 13.3496,
      "eval_samples_per_second": 14.532,
      "eval_steps_per_second": 1.873,
      "step": 475
    },
    {
      "epoch": 1.292824822236587,
      "grad_norm": 0.5704877376556396,
      "learning_rate": 0.00017413837781808759,
      "loss": 0.3104,
      "step": 500
    },
    {
      "epoch": 1.292824822236587,
      "eval_loss": 0.4043827950954437,
      "eval_runtime": 13.3487,
      "eval_samples_per_second": 14.533,
      "eval_steps_per_second": 1.873,
      "step": 500
    },
    {
      "epoch": 1.3574660633484164,
      "grad_norm": 0.410047322511673,
      "learning_rate": 0.000172842705364084,
      "loss": 0.3023,
      "step": 525
    },
    {
      "epoch": 1.3574660633484164,
      "eval_loss": 0.40202653408050537,
      "eval_runtime": 13.351,
      "eval_samples_per_second": 14.531,
      "eval_steps_per_second": 1.873,
      "step": 525
    },
    {
      "epoch": 1.4221073044602457,
      "grad_norm": 0.609170138835907,
      "learning_rate": 0.00017154703291008034,
      "loss": 0.298,
      "step": 550
    },
    {
      "epoch": 1.4221073044602457,
      "eval_loss": 0.3982485234737396,
      "eval_runtime": 13.3436,
      "eval_samples_per_second": 14.539,
      "eval_steps_per_second": 1.874,
      "step": 550
    },
    {
      "epoch": 1.486748545572075,
      "grad_norm": 0.5322830677032471,
      "learning_rate": 0.0001702513604560767,
      "loss": 0.2675,
      "step": 575
    },
    {
      "epoch": 1.486748545572075,
      "eval_loss": 0.3925786018371582,
      "eval_runtime": 13.3528,
      "eval_samples_per_second": 14.529,
      "eval_steps_per_second": 1.872,
      "step": 575
    },
    {
      "epoch": 1.5513897866839044,
      "grad_norm": 0.4065910279750824,
      "learning_rate": 0.00016895568800207308,
      "loss": 0.3029,
      "step": 600
    },
    {
      "epoch": 1.5513897866839044,
      "eval_loss": 0.3790633976459503,
      "eval_runtime": 13.361,
      "eval_samples_per_second": 14.52,
      "eval_steps_per_second": 1.871,
      "step": 600
    },
    {
      "epoch": 1.6160310277957337,
      "grad_norm": 0.5289133787155151,
      "learning_rate": 0.00016766001554806945,
      "loss": 0.2705,
      "step": 625
    },
    {
      "epoch": 1.6160310277957337,
      "eval_loss": 0.3800542652606964,
      "eval_runtime": 13.3478,
      "eval_samples_per_second": 14.534,
      "eval_steps_per_second": 1.873,
      "step": 625
    },
    {
      "epoch": 1.680672268907563,
      "grad_norm": 0.33712488412857056,
      "learning_rate": 0.00016636434309406583,
      "loss": 0.2841,
      "step": 650
    },
    {
      "epoch": 1.680672268907563,
      "eval_loss": 0.3804551362991333,
      "eval_runtime": 13.3456,
      "eval_samples_per_second": 14.537,
      "eval_steps_per_second": 1.873,
      "step": 650
    },
    {
      "epoch": 1.7453135100193924,
      "grad_norm": 0.31102249026298523,
      "learning_rate": 0.0001650686706400622,
      "loss": 0.28,
      "step": 675
    },
    {
      "epoch": 1.7453135100193924,
      "eval_loss": 0.3744698762893677,
      "eval_runtime": 13.3601,
      "eval_samples_per_second": 14.521,
      "eval_steps_per_second": 1.871,
      "step": 675
    },
    {
      "epoch": 1.8099547511312217,
      "grad_norm": 0.5691291689872742,
      "learning_rate": 0.00016377299818605857,
      "loss": 0.2652,
      "step": 700
    },
    {
      "epoch": 1.8099547511312217,
      "eval_loss": 0.3774305582046509,
      "eval_runtime": 13.3431,
      "eval_samples_per_second": 14.539,
      "eval_steps_per_second": 1.874,
      "step": 700
    },
    {
      "epoch": 1.874595992243051,
      "grad_norm": 0.4610290825366974,
      "learning_rate": 0.00016247732573205495,
      "loss": 0.2628,
      "step": 725
    },
    {
      "epoch": 1.874595992243051,
      "eval_loss": 0.37386658787727356,
      "eval_runtime": 13.3514,
      "eval_samples_per_second": 14.53,
      "eval_steps_per_second": 1.872,
      "step": 725
    },
    {
      "epoch": 1.9392372333548804,
      "grad_norm": 0.4986567497253418,
      "learning_rate": 0.00016118165327805132,
      "loss": 0.2622,
      "step": 750
    },
    {
      "epoch": 1.9392372333548804,
      "eval_loss": 0.37703824043273926,
      "eval_runtime": 13.3525,
      "eval_samples_per_second": 14.529,
      "eval_steps_per_second": 1.872,
      "step": 750
    },
    {
      "epoch": 2.0038784744667097,
      "grad_norm": 0.3808733820915222,
      "learning_rate": 0.00015988598082404768,
      "loss": 0.2308,
      "step": 775
    },
    {
      "epoch": 2.0038784744667097,
      "eval_loss": 0.3765748143196106,
      "eval_runtime": 13.3632,
      "eval_samples_per_second": 14.518,
      "eval_steps_per_second": 1.871,
      "step": 775
    },
    {
      "epoch": 2.0685197155785393,
      "grad_norm": 0.37369033694267273,
      "learning_rate": 0.00015859030837004406,
      "loss": 0.2115,
      "step": 800
    },
    {
      "epoch": 2.0685197155785393,
      "eval_loss": 0.3865245580673218,
      "eval_runtime": 13.356,
      "eval_samples_per_second": 14.525,
      "eval_steps_per_second": 1.872,
      "step": 800
    },
    {
      "epoch": 2.1331609566903684,
      "grad_norm": 0.41955631971359253,
      "learning_rate": 0.00015729463591604044,
      "loss": 0.1749,
      "step": 825
    },
    {
      "epoch": 2.1331609566903684,
      "eval_loss": 0.38168737292289734,
      "eval_runtime": 13.3562,
      "eval_samples_per_second": 14.525,
      "eval_steps_per_second": 1.872,
      "step": 825
    },
    {
      "epoch": 2.197802197802198,
      "grad_norm": 0.39571812748908997,
      "learning_rate": 0.0001559989634620368,
      "loss": 0.2131,
      "step": 850
    },
    {
      "epoch": 2.197802197802198,
      "eval_loss": 0.36951744556427,
      "eval_runtime": 13.355,
      "eval_samples_per_second": 14.526,
      "eval_steps_per_second": 1.872,
      "step": 850
    },
    {
      "epoch": 2.262443438914027,
      "grad_norm": 0.6356109380722046,
      "learning_rate": 0.00015470329100803317,
      "loss": 0.1882,
      "step": 875
    },
    {
      "epoch": 2.262443438914027,
      "eval_loss": 0.37231186032295227,
      "eval_runtime": 13.3614,
      "eval_samples_per_second": 14.519,
      "eval_steps_per_second": 1.871,
      "step": 875
    },
    {
      "epoch": 2.3270846800258567,
      "grad_norm": 0.3918738067150116,
      "learning_rate": 0.00015340761855402955,
      "loss": 0.2094,
      "step": 900
    },
    {
      "epoch": 2.3270846800258567,
      "eval_loss": 0.38670334219932556,
      "eval_runtime": 13.3624,
      "eval_samples_per_second": 14.518,
      "eval_steps_per_second": 1.871,
      "step": 900
    },
    {
      "epoch": 2.391725921137686,
      "grad_norm": 0.6116267442703247,
      "learning_rate": 0.00015211194610002593,
      "loss": 0.1987,
      "step": 925
    },
    {
      "epoch": 2.391725921137686,
      "eval_loss": 0.3871639668941498,
      "eval_runtime": 13.3618,
      "eval_samples_per_second": 14.519,
      "eval_steps_per_second": 1.871,
      "step": 925
    },
    {
      "epoch": 2.4563671622495153,
      "grad_norm": 0.49433189630508423,
      "learning_rate": 0.00015081627364602228,
      "loss": 0.2305,
      "step": 950
    },
    {
      "epoch": 2.4563671622495153,
      "eval_loss": 0.366474986076355,
      "eval_runtime": 13.37,
      "eval_samples_per_second": 14.51,
      "eval_steps_per_second": 1.87,
      "step": 950
    },
    {
      "epoch": 2.5210084033613445,
      "grad_norm": 0.43418142199516296,
      "learning_rate": 0.00014952060119201869,
      "loss": 0.1755,
      "step": 975
    },
    {
      "epoch": 2.5210084033613445,
      "eval_loss": 0.3646507263183594,
      "eval_runtime": 13.3532,
      "eval_samples_per_second": 14.528,
      "eval_steps_per_second": 1.872,
      "step": 975
    },
    {
      "epoch": 2.585649644473174,
      "grad_norm": 0.2716585397720337,
      "learning_rate": 0.00014822492873801504,
      "loss": 0.2384,
      "step": 1000
    },
    {
      "epoch": 2.585649644473174,
      "eval_loss": 0.35182201862335205,
      "eval_runtime": 13.3693,
      "eval_samples_per_second": 14.511,
      "eval_steps_per_second": 1.87,
      "step": 1000
    },
    {
      "epoch": 2.650290885585003,
      "grad_norm": 0.4708922207355499,
      "learning_rate": 0.00014692925628401142,
      "loss": 0.1689,
      "step": 1025
    },
    {
      "epoch": 2.650290885585003,
      "eval_loss": 0.3655441999435425,
      "eval_runtime": 13.371,
      "eval_samples_per_second": 14.509,
      "eval_steps_per_second": 1.87,
      "step": 1025
    },
    {
      "epoch": 2.7149321266968327,
      "grad_norm": 0.4261806607246399,
      "learning_rate": 0.00014563358383000777,
      "loss": 0.2181,
      "step": 1050
    },
    {
      "epoch": 2.7149321266968327,
      "eval_loss": 0.3593634366989136,
      "eval_runtime": 13.3548,
      "eval_samples_per_second": 14.527,
      "eval_steps_per_second": 1.872,
      "step": 1050
    },
    {
      "epoch": 2.779573367808662,
      "grad_norm": 0.5506019592285156,
      "learning_rate": 0.00014433791137600415,
      "loss": 0.1778,
      "step": 1075
    },
    {
      "epoch": 2.779573367808662,
      "eval_loss": 0.35731247067451477,
      "eval_runtime": 13.357,
      "eval_samples_per_second": 14.524,
      "eval_steps_per_second": 1.872,
      "step": 1075
    },
    {
      "epoch": 2.8442146089204914,
      "grad_norm": 0.5495656132698059,
      "learning_rate": 0.00014304223892200053,
      "loss": 0.2239,
      "step": 1100
    },
    {
      "epoch": 2.8442146089204914,
      "eval_loss": 0.35565704107284546,
      "eval_runtime": 13.3623,
      "eval_samples_per_second": 14.518,
      "eval_steps_per_second": 1.871,
      "step": 1100
    },
    {
      "epoch": 2.9088558500323205,
      "grad_norm": 0.7420383095741272,
      "learning_rate": 0.00014174656646799688,
      "loss": 0.1647,
      "step": 1125
    },
    {
      "epoch": 2.9088558500323205,
      "eval_loss": 0.36469554901123047,
      "eval_runtime": 13.3662,
      "eval_samples_per_second": 14.514,
      "eval_steps_per_second": 1.87,
      "step": 1125
    },
    {
      "epoch": 2.97349709114415,
      "grad_norm": 0.2853482961654663,
      "learning_rate": 0.00014045089401399326,
      "loss": 0.1909,
      "step": 1150
    },
    {
      "epoch": 2.97349709114415,
      "eval_loss": 0.3539246916770935,
      "eval_runtime": 13.3625,
      "eval_samples_per_second": 14.518,
      "eval_steps_per_second": 1.871,
      "step": 1150
    },
    {
      "epoch": 3.038138332255979,
      "grad_norm": 0.4978826940059662,
      "learning_rate": 0.00013915522155998964,
      "loss": 0.1684,
      "step": 1175
    },
    {
      "epoch": 3.038138332255979,
      "eval_loss": 0.3665294945240021,
      "eval_runtime": 13.3706,
      "eval_samples_per_second": 14.509,
      "eval_steps_per_second": 1.87,
      "step": 1175
    },
    {
      "epoch": 3.1027795733678087,
      "grad_norm": 0.40166956186294556,
      "learning_rate": 0.00013785954910598602,
      "loss": 0.1398,
      "step": 1200
    },
    {
      "epoch": 3.1027795733678087,
      "eval_loss": 0.37494078278541565,
      "eval_runtime": 13.3667,
      "eval_samples_per_second": 14.514,
      "eval_steps_per_second": 1.87,
      "step": 1200
    },
    {
      "epoch": 3.167420814479638,
      "grad_norm": 0.507607638835907,
      "learning_rate": 0.00013656387665198237,
      "loss": 0.1326,
      "step": 1225
    },
    {
      "epoch": 3.167420814479638,
      "eval_loss": 0.3721146583557129,
      "eval_runtime": 13.3726,
      "eval_samples_per_second": 14.507,
      "eval_steps_per_second": 1.869,
      "step": 1225
    },
    {
      "epoch": 3.2320620555914674,
      "grad_norm": 0.38957518339157104,
      "learning_rate": 0.00013526820419797878,
      "loss": 0.1328,
      "step": 1250
    },
    {
      "epoch": 3.2320620555914674,
      "eval_loss": 0.37096062302589417,
      "eval_runtime": 13.364,
      "eval_samples_per_second": 14.517,
      "eval_steps_per_second": 1.871,
      "step": 1250
    },
    {
      "epoch": 3.2967032967032965,
      "grad_norm": 0.7261212468147278,
      "learning_rate": 0.00013397253174397513,
      "loss": 0.1634,
      "step": 1275
    },
    {
      "epoch": 3.2967032967032965,
      "eval_loss": 0.36481988430023193,
      "eval_runtime": 13.3645,
      "eval_samples_per_second": 14.516,
      "eval_steps_per_second": 1.871,
      "step": 1275
    },
    {
      "epoch": 3.361344537815126,
      "grad_norm": 0.3340177834033966,
      "learning_rate": 0.0001326768592899715,
      "loss": 0.1403,
      "step": 1300
    },
    {
      "epoch": 3.361344537815126,
      "eval_loss": 0.36701083183288574,
      "eval_runtime": 13.3687,
      "eval_samples_per_second": 14.512,
      "eval_steps_per_second": 1.87,
      "step": 1300
    },
    {
      "epoch": 3.425985778926955,
      "grad_norm": 0.5969800353050232,
      "learning_rate": 0.00013138118683596786,
      "loss": 0.1549,
      "step": 1325
    },
    {
      "epoch": 3.425985778926955,
      "eval_loss": 0.3541284203529358,
      "eval_runtime": 13.3607,
      "eval_samples_per_second": 14.52,
      "eval_steps_per_second": 1.871,
      "step": 1325
    },
    {
      "epoch": 3.490627020038785,
      "grad_norm": 0.30678942799568176,
      "learning_rate": 0.00013008551438196424,
      "loss": 0.1492,
      "step": 1350
    },
    {
      "epoch": 3.490627020038785,
      "eval_loss": 0.35420697927474976,
      "eval_runtime": 13.3674,
      "eval_samples_per_second": 14.513,
      "eval_steps_per_second": 1.87,
      "step": 1350
    },
    {
      "epoch": 3.555268261150614,
      "grad_norm": 0.7490562796592712,
      "learning_rate": 0.00012878984192796062,
      "loss": 0.1514,
      "step": 1375
    },
    {
      "epoch": 3.555268261150614,
      "eval_loss": 0.35634493827819824,
      "eval_runtime": 13.3824,
      "eval_samples_per_second": 14.497,
      "eval_steps_per_second": 1.868,
      "step": 1375
    },
    {
      "epoch": 3.6199095022624435,
      "grad_norm": 0.3132079541683197,
      "learning_rate": 0.00012749416947395697,
      "loss": 0.1493,
      "step": 1400
    },
    {
      "epoch": 3.6199095022624435,
      "eval_loss": 0.3662796914577484,
      "eval_runtime": 13.3685,
      "eval_samples_per_second": 14.512,
      "eval_steps_per_second": 1.87,
      "step": 1400
    },
    {
      "epoch": 3.684550743374273,
      "grad_norm": 0.4222179651260376,
      "learning_rate": 0.00012619849701995338,
      "loss": 0.1403,
      "step": 1425
    },
    {
      "epoch": 3.684550743374273,
      "eval_loss": 0.36451268196105957,
      "eval_runtime": 13.3664,
      "eval_samples_per_second": 14.514,
      "eval_steps_per_second": 1.87,
      "step": 1425
    },
    {
      "epoch": 3.749191984486102,
      "grad_norm": 0.4552185833454132,
      "learning_rate": 0.00012490282456594973,
      "loss": 0.15,
      "step": 1450
    },
    {
      "epoch": 3.749191984486102,
      "eval_loss": 0.3632330298423767,
      "eval_runtime": 13.374,
      "eval_samples_per_second": 14.506,
      "eval_steps_per_second": 1.869,
      "step": 1450
    },
    {
      "epoch": 3.8138332255979313,
      "grad_norm": 0.39688166975975037,
      "learning_rate": 0.0001236071521119461,
      "loss": 0.1648,
      "step": 1475
    },
    {
      "epoch": 3.8138332255979313,
      "eval_loss": 0.35391995310783386,
      "eval_runtime": 13.3842,
      "eval_samples_per_second": 14.495,
      "eval_steps_per_second": 1.868,
      "step": 1475
    },
    {
      "epoch": 3.878474466709761,
      "grad_norm": 0.2690788209438324,
      "learning_rate": 0.00012231147965794246,
      "loss": 0.1486,
      "step": 1500
    },
    {
      "epoch": 3.878474466709761,
      "eval_loss": 0.36058586835861206,
      "eval_runtime": 13.3733,
      "eval_samples_per_second": 14.506,
      "eval_steps_per_second": 1.869,
      "step": 1500
    },
    {
      "epoch": 3.9431157078215904,
      "grad_norm": 0.3785063922405243,
      "learning_rate": 0.00012101580720393886,
      "loss": 0.1523,
      "step": 1525
    },
    {
      "epoch": 3.9431157078215904,
      "eval_loss": 0.3479475975036621,
      "eval_runtime": 13.3703,
      "eval_samples_per_second": 14.51,
      "eval_steps_per_second": 1.87,
      "step": 1525
    },
    {
      "epoch": 4.0077569489334195,
      "grad_norm": 0.39190673828125,
      "learning_rate": 0.00011972013474993522,
      "loss": 0.1408,
      "step": 1550
    },
    {
      "epoch": 4.0077569489334195,
      "eval_loss": 0.3482886254787445,
      "eval_runtime": 13.3744,
      "eval_samples_per_second": 14.505,
      "eval_steps_per_second": 1.869,
      "step": 1550
    },
    {
      "epoch": 4.072398190045249,
      "grad_norm": 0.33796021342277527,
      "learning_rate": 0.0001184244622959316,
      "loss": 0.1163,
      "step": 1575
    },
    {
      "epoch": 4.072398190045249,
      "eval_loss": 0.36559316515922546,
      "eval_runtime": 13.3799,
      "eval_samples_per_second": 14.499,
      "eval_steps_per_second": 1.868,
      "step": 1575
    },
    {
      "epoch": 4.137039431157079,
      "grad_norm": 0.843379020690918,
      "learning_rate": 0.00011712878984192797,
      "loss": 0.1118,
      "step": 1600
    },
    {
      "epoch": 4.137039431157079,
      "eval_loss": 0.36994925141334534,
      "eval_runtime": 13.3773,
      "eval_samples_per_second": 14.502,
      "eval_steps_per_second": 1.869,
      "step": 1600
    },
    {
      "epoch": 4.201680672268908,
      "grad_norm": 0.30512169003486633,
      "learning_rate": 0.00011583311738792433,
      "loss": 0.1249,
      "step": 1625
    },
    {
      "epoch": 4.201680672268908,
      "eval_loss": 0.36630260944366455,
      "eval_runtime": 13.3863,
      "eval_samples_per_second": 14.492,
      "eval_steps_per_second": 1.868,
      "step": 1625
    },
    {
      "epoch": 4.266321913380737,
      "grad_norm": 0.35797879099845886,
      "learning_rate": 0.00011453744493392071,
      "loss": 0.1154,
      "step": 1650
    },
    {
      "epoch": 4.266321913380737,
      "eval_loss": 0.36872294545173645,
      "eval_runtime": 13.3799,
      "eval_samples_per_second": 14.499,
      "eval_steps_per_second": 1.868,
      "step": 1650
    },
    {
      "epoch": 4.330963154492566,
      "grad_norm": 0.3784850239753723,
      "learning_rate": 0.00011324177247991708,
      "loss": 0.1237,
      "step": 1675
    },
    {
      "epoch": 4.330963154492566,
      "eval_loss": 0.3683045506477356,
      "eval_runtime": 13.3855,
      "eval_samples_per_second": 14.493,
      "eval_steps_per_second": 1.868,
      "step": 1675
    },
    {
      "epoch": 4.395604395604396,
      "grad_norm": 0.5391280651092529,
      "learning_rate": 0.00011194610002591346,
      "loss": 0.1268,
      "step": 1700
    },
    {
      "epoch": 4.395604395604396,
      "eval_loss": 0.366862028837204,
      "eval_runtime": 13.3849,
      "eval_samples_per_second": 14.494,
      "eval_steps_per_second": 1.868,
      "step": 1700
    },
    {
      "epoch": 4.460245636716225,
      "grad_norm": 0.3441055417060852,
      "learning_rate": 0.00011065042757190982,
      "loss": 0.1154,
      "step": 1725
    },
    {
      "epoch": 4.460245636716225,
      "eval_loss": 0.36934134364128113,
      "eval_runtime": 13.3807,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 1725
    },
    {
      "epoch": 4.524886877828054,
      "grad_norm": 0.38485729694366455,
      "learning_rate": 0.0001093547551179062,
      "loss": 0.1196,
      "step": 1750
    },
    {
      "epoch": 4.524886877828054,
      "eval_loss": 0.37078526616096497,
      "eval_runtime": 13.3642,
      "eval_samples_per_second": 14.516,
      "eval_steps_per_second": 1.871,
      "step": 1750
    },
    {
      "epoch": 4.589528118939883,
      "grad_norm": 0.27951779961586,
      "learning_rate": 0.00010805908266390257,
      "loss": 0.1271,
      "step": 1775
    },
    {
      "epoch": 4.589528118939883,
      "eval_loss": 0.36370983719825745,
      "eval_runtime": 13.3815,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 1775
    },
    {
      "epoch": 4.654169360051713,
      "grad_norm": 0.41593897342681885,
      "learning_rate": 0.00010676341020989895,
      "loss": 0.1227,
      "step": 1800
    },
    {
      "epoch": 4.654169360051713,
      "eval_loss": 0.37005409598350525,
      "eval_runtime": 13.3667,
      "eval_samples_per_second": 14.514,
      "eval_steps_per_second": 1.87,
      "step": 1800
    },
    {
      "epoch": 4.7188106011635425,
      "grad_norm": 0.44129708409309387,
      "learning_rate": 0.00010546773775589531,
      "loss": 0.1177,
      "step": 1825
    },
    {
      "epoch": 4.7188106011635425,
      "eval_loss": 0.36843162775039673,
      "eval_runtime": 13.3752,
      "eval_samples_per_second": 14.504,
      "eval_steps_per_second": 1.869,
      "step": 1825
    },
    {
      "epoch": 4.783451842275372,
      "grad_norm": 0.6084820032119751,
      "learning_rate": 0.00010417206530189169,
      "loss": 0.1201,
      "step": 1850
    },
    {
      "epoch": 4.783451842275372,
      "eval_loss": 0.37733200192451477,
      "eval_runtime": 13.3814,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 1850
    },
    {
      "epoch": 4.848093083387201,
      "grad_norm": 0.4541929364204407,
      "learning_rate": 0.00010287639284788806,
      "loss": 0.1159,
      "step": 1875
    },
    {
      "epoch": 4.848093083387201,
      "eval_loss": 0.3665323853492737,
      "eval_runtime": 13.3818,
      "eval_samples_per_second": 14.497,
      "eval_steps_per_second": 1.868,
      "step": 1875
    },
    {
      "epoch": 4.912734324499031,
      "grad_norm": 0.5170966982841492,
      "learning_rate": 0.00010158072039388442,
      "loss": 0.1215,
      "step": 1900
    },
    {
      "epoch": 4.912734324499031,
      "eval_loss": 0.37353450059890747,
      "eval_runtime": 13.3783,
      "eval_samples_per_second": 14.501,
      "eval_steps_per_second": 1.869,
      "step": 1900
    },
    {
      "epoch": 4.97737556561086,
      "grad_norm": 0.3494817912578583,
      "learning_rate": 0.0001002850479398808,
      "loss": 0.1203,
      "step": 1925
    },
    {
      "epoch": 4.97737556561086,
      "eval_loss": 0.3610128164291382,
      "eval_runtime": 13.3811,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 1925
    },
    {
      "epoch": 5.042016806722689,
      "grad_norm": 0.713656485080719,
      "learning_rate": 9.898937548587718e-05,
      "loss": 0.1036,
      "step": 1950
    },
    {
      "epoch": 5.042016806722689,
      "eval_loss": 0.3713524341583252,
      "eval_runtime": 13.3758,
      "eval_samples_per_second": 14.504,
      "eval_steps_per_second": 1.869,
      "step": 1950
    },
    {
      "epoch": 5.106658047834518,
      "grad_norm": 0.564629316329956,
      "learning_rate": 9.769370303187355e-05,
      "loss": 0.0999,
      "step": 1975
    },
    {
      "epoch": 5.106658047834518,
      "eval_loss": 0.39409345388412476,
      "eval_runtime": 13.3765,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 1975
    },
    {
      "epoch": 5.171299288946348,
      "grad_norm": 0.3473380208015442,
      "learning_rate": 9.639803057786993e-05,
      "loss": 0.1026,
      "step": 2000
    },
    {
      "epoch": 5.171299288946348,
      "eval_loss": 0.3814166486263275,
      "eval_runtime": 13.3768,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 2000
    },
    {
      "epoch": 5.235940530058177,
      "grad_norm": 0.2592593729496002,
      "learning_rate": 9.510235812386628e-05,
      "loss": 0.1072,
      "step": 2025
    },
    {
      "epoch": 5.235940530058177,
      "eval_loss": 0.3801679313182831,
      "eval_runtime": 13.3829,
      "eval_samples_per_second": 14.496,
      "eval_steps_per_second": 1.868,
      "step": 2025
    },
    {
      "epoch": 5.300581771170006,
      "grad_norm": 0.5560858249664307,
      "learning_rate": 9.380668566986266e-05,
      "loss": 0.1047,
      "step": 2050
    },
    {
      "epoch": 5.300581771170006,
      "eval_loss": 0.3803362548351288,
      "eval_runtime": 13.3828,
      "eval_samples_per_second": 14.496,
      "eval_steps_per_second": 1.868,
      "step": 2050
    },
    {
      "epoch": 5.365223012281835,
      "grad_norm": 0.31524232029914856,
      "learning_rate": 9.251101321585903e-05,
      "loss": 0.101,
      "step": 2075
    },
    {
      "epoch": 5.365223012281835,
      "eval_loss": 0.3836386799812317,
      "eval_runtime": 13.3933,
      "eval_samples_per_second": 14.485,
      "eval_steps_per_second": 1.867,
      "step": 2075
    },
    {
      "epoch": 5.429864253393665,
      "grad_norm": 0.4523369073867798,
      "learning_rate": 9.12153407618554e-05,
      "loss": 0.1018,
      "step": 2100
    },
    {
      "epoch": 5.429864253393665,
      "eval_loss": 0.38159799575805664,
      "eval_runtime": 13.3858,
      "eval_samples_per_second": 14.493,
      "eval_steps_per_second": 1.868,
      "step": 2100
    },
    {
      "epoch": 5.4945054945054945,
      "grad_norm": 0.2764829099178314,
      "learning_rate": 8.991966830785178e-05,
      "loss": 0.1034,
      "step": 2125
    },
    {
      "epoch": 5.4945054945054945,
      "eval_loss": 0.3741620182991028,
      "eval_runtime": 13.3746,
      "eval_samples_per_second": 14.505,
      "eval_steps_per_second": 1.869,
      "step": 2125
    },
    {
      "epoch": 5.559146735617324,
      "grad_norm": 0.3252582848072052,
      "learning_rate": 8.862399585384815e-05,
      "loss": 0.1047,
      "step": 2150
    },
    {
      "epoch": 5.559146735617324,
      "eval_loss": 0.3792659342288971,
      "eval_runtime": 13.3818,
      "eval_samples_per_second": 14.497,
      "eval_steps_per_second": 1.868,
      "step": 2150
    },
    {
      "epoch": 5.623787976729153,
      "grad_norm": 0.31311649084091187,
      "learning_rate": 8.732832339984453e-05,
      "loss": 0.1061,
      "step": 2175
    },
    {
      "epoch": 5.623787976729153,
      "eval_loss": 0.3811538815498352,
      "eval_runtime": 13.3767,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 2175
    },
    {
      "epoch": 5.688429217840983,
      "grad_norm": 0.6086262464523315,
      "learning_rate": 8.60326509458409e-05,
      "loss": 0.1041,
      "step": 2200
    },
    {
      "epoch": 5.688429217840983,
      "eval_loss": 0.36947333812713623,
      "eval_runtime": 13.3773,
      "eval_samples_per_second": 14.502,
      "eval_steps_per_second": 1.869,
      "step": 2200
    },
    {
      "epoch": 5.753070458952812,
      "grad_norm": 0.3273337781429291,
      "learning_rate": 8.473697849183728e-05,
      "loss": 0.1092,
      "step": 2225
    },
    {
      "epoch": 5.753070458952812,
      "eval_loss": 0.3786868751049042,
      "eval_runtime": 13.371,
      "eval_samples_per_second": 14.509,
      "eval_steps_per_second": 1.87,
      "step": 2225
    },
    {
      "epoch": 5.817711700064641,
      "grad_norm": 0.5605947375297546,
      "learning_rate": 8.344130603783364e-05,
      "loss": 0.1055,
      "step": 2250
    },
    {
      "epoch": 5.817711700064641,
      "eval_loss": 0.37696948647499084,
      "eval_runtime": 13.3873,
      "eval_samples_per_second": 14.491,
      "eval_steps_per_second": 1.867,
      "step": 2250
    },
    {
      "epoch": 5.882352941176471,
      "grad_norm": 0.3829572796821594,
      "learning_rate": 8.214563358383001e-05,
      "loss": 0.1047,
      "step": 2275
    },
    {
      "epoch": 5.882352941176471,
      "eval_loss": 0.3816302716732025,
      "eval_runtime": 13.375,
      "eval_samples_per_second": 14.505,
      "eval_steps_per_second": 1.869,
      "step": 2275
    },
    {
      "epoch": 5.9469941822883,
      "grad_norm": 0.36125028133392334,
      "learning_rate": 8.084996112982637e-05,
      "loss": 0.1005,
      "step": 2300
    },
    {
      "epoch": 5.9469941822883,
      "eval_loss": 0.37172621488571167,
      "eval_runtime": 13.3728,
      "eval_samples_per_second": 14.507,
      "eval_steps_per_second": 1.869,
      "step": 2300
    },
    {
      "epoch": 6.011635423400129,
      "grad_norm": 0.37611111998558044,
      "learning_rate": 7.955428867582275e-05,
      "loss": 0.1007,
      "step": 2325
    },
    {
      "epoch": 6.011635423400129,
      "eval_loss": 0.3778633773326874,
      "eval_runtime": 13.3736,
      "eval_samples_per_second": 14.506,
      "eval_steps_per_second": 1.869,
      "step": 2325
    },
    {
      "epoch": 6.076276664511958,
      "grad_norm": 0.23287895321846008,
      "learning_rate": 7.825861622181913e-05,
      "loss": 0.0853,
      "step": 2350
    },
    {
      "epoch": 6.076276664511958,
      "eval_loss": 0.3875974416732788,
      "eval_runtime": 13.3815,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 2350
    },
    {
      "epoch": 6.140917905623788,
      "grad_norm": 0.9243698120117188,
      "learning_rate": 7.69629437678155e-05,
      "loss": 0.0941,
      "step": 2375
    },
    {
      "epoch": 6.140917905623788,
      "eval_loss": 0.3978659212589264,
      "eval_runtime": 13.3745,
      "eval_samples_per_second": 14.505,
      "eval_steps_per_second": 1.869,
      "step": 2375
    },
    {
      "epoch": 6.2055591467356175,
      "grad_norm": 0.2637142241001129,
      "learning_rate": 7.566727131381188e-05,
      "loss": 0.087,
      "step": 2400
    },
    {
      "epoch": 6.2055591467356175,
      "eval_loss": 0.3923758268356323,
      "eval_runtime": 13.3843,
      "eval_samples_per_second": 14.495,
      "eval_steps_per_second": 1.868,
      "step": 2400
    },
    {
      "epoch": 6.270200387847447,
      "grad_norm": 0.48433196544647217,
      "learning_rate": 7.437159885980824e-05,
      "loss": 0.0969,
      "step": 2425
    },
    {
      "epoch": 6.270200387847447,
      "eval_loss": 0.38719630241394043,
      "eval_runtime": 13.3714,
      "eval_samples_per_second": 14.509,
      "eval_steps_per_second": 1.87,
      "step": 2425
    },
    {
      "epoch": 6.334841628959276,
      "grad_norm": 0.3420272767543793,
      "learning_rate": 7.307592640580462e-05,
      "loss": 0.091,
      "step": 2450
    },
    {
      "epoch": 6.334841628959276,
      "eval_loss": 0.39057913422584534,
      "eval_runtime": 13.3771,
      "eval_samples_per_second": 14.502,
      "eval_steps_per_second": 1.869,
      "step": 2450
    },
    {
      "epoch": 6.399482870071106,
      "grad_norm": 0.48248687386512756,
      "learning_rate": 7.178025395180099e-05,
      "loss": 0.0983,
      "step": 2475
    },
    {
      "epoch": 6.399482870071106,
      "eval_loss": 0.38908788561820984,
      "eval_runtime": 13.3905,
      "eval_samples_per_second": 14.488,
      "eval_steps_per_second": 1.867,
      "step": 2475
    },
    {
      "epoch": 6.464124111182935,
      "grad_norm": 0.2842746078968048,
      "learning_rate": 7.048458149779737e-05,
      "loss": 0.0908,
      "step": 2500
    },
    {
      "epoch": 6.464124111182935,
      "eval_loss": 0.387993723154068,
      "eval_runtime": 13.4025,
      "eval_samples_per_second": 14.475,
      "eval_steps_per_second": 1.865,
      "step": 2500
    },
    {
      "epoch": 6.528765352294764,
      "grad_norm": 0.31269609928131104,
      "learning_rate": 6.918890904379373e-05,
      "loss": 0.0942,
      "step": 2525
    },
    {
      "epoch": 6.528765352294764,
      "eval_loss": 0.38906431198120117,
      "eval_runtime": 13.3831,
      "eval_samples_per_second": 14.496,
      "eval_steps_per_second": 1.868,
      "step": 2525
    },
    {
      "epoch": 6.593406593406593,
      "grad_norm": 0.2563638687133789,
      "learning_rate": 6.78932365897901e-05,
      "loss": 0.0906,
      "step": 2550
    },
    {
      "epoch": 6.593406593406593,
      "eval_loss": 0.39094778895378113,
      "eval_runtime": 13.3827,
      "eval_samples_per_second": 14.496,
      "eval_steps_per_second": 1.868,
      "step": 2550
    },
    {
      "epoch": 6.658047834518423,
      "grad_norm": 0.228476420044899,
      "learning_rate": 6.659756413578648e-05,
      "loss": 0.0969,
      "step": 2575
    },
    {
      "epoch": 6.658047834518423,
      "eval_loss": 0.3924230933189392,
      "eval_runtime": 13.3727,
      "eval_samples_per_second": 14.507,
      "eval_steps_per_second": 1.869,
      "step": 2575
    },
    {
      "epoch": 6.722689075630252,
      "grad_norm": 0.19723258912563324,
      "learning_rate": 6.530189168178284e-05,
      "loss": 0.0922,
      "step": 2600
    },
    {
      "epoch": 6.722689075630252,
      "eval_loss": 0.3915347158908844,
      "eval_runtime": 13.3804,
      "eval_samples_per_second": 14.499,
      "eval_steps_per_second": 1.868,
      "step": 2600
    },
    {
      "epoch": 6.787330316742081,
      "grad_norm": 0.3364139199256897,
      "learning_rate": 6.400621922777922e-05,
      "loss": 0.0989,
      "step": 2625
    },
    {
      "epoch": 6.787330316742081,
      "eval_loss": 0.3891298770904541,
      "eval_runtime": 13.3771,
      "eval_samples_per_second": 14.502,
      "eval_steps_per_second": 1.869,
      "step": 2625
    },
    {
      "epoch": 6.85197155785391,
      "grad_norm": 0.31284722685813904,
      "learning_rate": 6.271054677377559e-05,
      "loss": 0.0947,
      "step": 2650
    },
    {
      "epoch": 6.85197155785391,
      "eval_loss": 0.3864246904850006,
      "eval_runtime": 13.3787,
      "eval_samples_per_second": 14.501,
      "eval_steps_per_second": 1.869,
      "step": 2650
    },
    {
      "epoch": 6.91661279896574,
      "grad_norm": 0.4304758906364441,
      "learning_rate": 6.141487431977197e-05,
      "loss": 0.0955,
      "step": 2675
    },
    {
      "epoch": 6.91661279896574,
      "eval_loss": 0.38957709074020386,
      "eval_runtime": 13.3927,
      "eval_samples_per_second": 14.485,
      "eval_steps_per_second": 1.867,
      "step": 2675
    },
    {
      "epoch": 6.98125404007757,
      "grad_norm": 0.21785129606723785,
      "learning_rate": 6.0119201865768335e-05,
      "loss": 0.0957,
      "step": 2700
    },
    {
      "epoch": 6.98125404007757,
      "eval_loss": 0.3870868384838104,
      "eval_runtime": 13.3762,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 2700
    },
    {
      "epoch": 7.045895281189399,
      "grad_norm": 0.31381458044052124,
      "learning_rate": 5.882352941176471e-05,
      "loss": 0.0844,
      "step": 2725
    },
    {
      "epoch": 7.045895281189399,
      "eval_loss": 0.3945547640323639,
      "eval_runtime": 13.3799,
      "eval_samples_per_second": 14.499,
      "eval_steps_per_second": 1.868,
      "step": 2725
    },
    {
      "epoch": 7.110536522301228,
      "grad_norm": 0.33102476596832275,
      "learning_rate": 5.752785695776109e-05,
      "loss": 0.0883,
      "step": 2750
    },
    {
      "epoch": 7.110536522301228,
      "eval_loss": 0.4064219295978546,
      "eval_runtime": 13.3763,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 2750
    },
    {
      "epoch": 7.175177763413058,
      "grad_norm": 0.2840661406517029,
      "learning_rate": 5.623218450375746e-05,
      "loss": 0.0812,
      "step": 2775
    },
    {
      "epoch": 7.175177763413058,
      "eval_loss": 0.40618330240249634,
      "eval_runtime": 13.391,
      "eval_samples_per_second": 14.487,
      "eval_steps_per_second": 1.867,
      "step": 2775
    },
    {
      "epoch": 7.239819004524887,
      "grad_norm": 0.3565811216831207,
      "learning_rate": 5.493651204975383e-05,
      "loss": 0.0894,
      "step": 2800
    },
    {
      "epoch": 7.239819004524887,
      "eval_loss": 0.4039536416530609,
      "eval_runtime": 13.3689,
      "eval_samples_per_second": 14.511,
      "eval_steps_per_second": 1.87,
      "step": 2800
    },
    {
      "epoch": 7.304460245636716,
      "grad_norm": 0.19193735718727112,
      "learning_rate": 5.364083959575019e-05,
      "loss": 0.0827,
      "step": 2825
    },
    {
      "epoch": 7.304460245636716,
      "eval_loss": 0.40351200103759766,
      "eval_runtime": 13.3814,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 2825
    },
    {
      "epoch": 7.369101486748546,
      "grad_norm": 0.28010833263397217,
      "learning_rate": 5.2345167141746564e-05,
      "loss": 0.0937,
      "step": 2850
    },
    {
      "epoch": 7.369101486748546,
      "eval_loss": 0.40530896186828613,
      "eval_runtime": 13.371,
      "eval_samples_per_second": 14.509,
      "eval_steps_per_second": 1.87,
      "step": 2850
    },
    {
      "epoch": 7.433742727860375,
      "grad_norm": 0.1792915165424347,
      "learning_rate": 5.1049494687742936e-05,
      "loss": 0.0839,
      "step": 2875
    },
    {
      "epoch": 7.433742727860375,
      "eval_loss": 0.40639927983283997,
      "eval_runtime": 13.3697,
      "eval_samples_per_second": 14.51,
      "eval_steps_per_second": 1.87,
      "step": 2875
    },
    {
      "epoch": 7.498383968972204,
      "grad_norm": 0.23624753952026367,
      "learning_rate": 4.975382223373931e-05,
      "loss": 0.0888,
      "step": 2900
    },
    {
      "epoch": 7.498383968972204,
      "eval_loss": 0.4040166437625885,
      "eval_runtime": 13.3871,
      "eval_samples_per_second": 14.492,
      "eval_steps_per_second": 1.867,
      "step": 2900
    },
    {
      "epoch": 7.563025210084033,
      "grad_norm": 0.3217020630836487,
      "learning_rate": 4.845814977973568e-05,
      "loss": 0.0839,
      "step": 2925
    },
    {
      "epoch": 7.563025210084033,
      "eval_loss": 0.40058034658432007,
      "eval_runtime": 13.3755,
      "eval_samples_per_second": 14.504,
      "eval_steps_per_second": 1.869,
      "step": 2925
    },
    {
      "epoch": 7.6276664511958625,
      "grad_norm": 0.19924822449684143,
      "learning_rate": 4.7162477325732054e-05,
      "loss": 0.0931,
      "step": 2950
    },
    {
      "epoch": 7.6276664511958625,
      "eval_loss": 0.3974081575870514,
      "eval_runtime": 13.3797,
      "eval_samples_per_second": 14.5,
      "eval_steps_per_second": 1.869,
      "step": 2950
    },
    {
      "epoch": 7.6923076923076925,
      "grad_norm": 0.22853316366672516,
      "learning_rate": 4.5866804871728434e-05,
      "loss": 0.0857,
      "step": 2975
    },
    {
      "epoch": 7.6923076923076925,
      "eval_loss": 0.3978106379508972,
      "eval_runtime": 13.3649,
      "eval_samples_per_second": 14.516,
      "eval_steps_per_second": 1.871,
      "step": 2975
    },
    {
      "epoch": 7.756948933419522,
      "grad_norm": 0.3232315480709076,
      "learning_rate": 4.4571132417724806e-05,
      "loss": 0.0944,
      "step": 3000
    },
    {
      "epoch": 7.756948933419522,
      "eval_loss": 0.39671608805656433,
      "eval_runtime": 13.3837,
      "eval_samples_per_second": 14.495,
      "eval_steps_per_second": 1.868,
      "step": 3000
    },
    {
      "epoch": 7.821590174531351,
      "grad_norm": 0.19052205979824066,
      "learning_rate": 4.327545996372117e-05,
      "loss": 0.0825,
      "step": 3025
    },
    {
      "epoch": 7.821590174531351,
      "eval_loss": 0.39850884675979614,
      "eval_runtime": 13.3766,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 3025
    },
    {
      "epoch": 7.886231415643181,
      "grad_norm": 0.24069742858409882,
      "learning_rate": 4.1979787509717545e-05,
      "loss": 0.0926,
      "step": 3050
    },
    {
      "epoch": 7.886231415643181,
      "eval_loss": 0.40030214190483093,
      "eval_runtime": 13.3678,
      "eval_samples_per_second": 14.512,
      "eval_steps_per_second": 1.87,
      "step": 3050
    },
    {
      "epoch": 7.95087265675501,
      "grad_norm": 0.26551541686058044,
      "learning_rate": 4.068411505571392e-05,
      "loss": 0.0858,
      "step": 3075
    },
    {
      "epoch": 7.95087265675501,
      "eval_loss": 0.39577826857566833,
      "eval_runtime": 13.3848,
      "eval_samples_per_second": 14.494,
      "eval_steps_per_second": 1.868,
      "step": 3075
    },
    {
      "epoch": 8.015513897866839,
      "grad_norm": 0.23020051419734955,
      "learning_rate": 3.938844260171029e-05,
      "loss": 0.0841,
      "step": 3100
    },
    {
      "epoch": 8.015513897866839,
      "eval_loss": 0.4009804129600525,
      "eval_runtime": 13.3828,
      "eval_samples_per_second": 14.496,
      "eval_steps_per_second": 1.868,
      "step": 3100
    },
    {
      "epoch": 8.080155138978668,
      "grad_norm": 0.23779013752937317,
      "learning_rate": 3.809277014770666e-05,
      "loss": 0.0795,
      "step": 3125
    },
    {
      "epoch": 8.080155138978668,
      "eval_loss": 0.40682119131088257,
      "eval_runtime": 13.3724,
      "eval_samples_per_second": 14.507,
      "eval_steps_per_second": 1.87,
      "step": 3125
    },
    {
      "epoch": 8.144796380090497,
      "grad_norm": 0.30244529247283936,
      "learning_rate": 3.679709769370303e-05,
      "loss": 0.083,
      "step": 3150
    },
    {
      "epoch": 8.144796380090497,
      "eval_loss": 0.40890321135520935,
      "eval_runtime": 13.3736,
      "eval_samples_per_second": 14.506,
      "eval_steps_per_second": 1.869,
      "step": 3150
    },
    {
      "epoch": 8.209437621202326,
      "grad_norm": 0.2585161328315735,
      "learning_rate": 3.55014252396994e-05,
      "loss": 0.078,
      "step": 3175
    },
    {
      "epoch": 8.209437621202326,
      "eval_loss": 0.4109211564064026,
      "eval_runtime": 13.3935,
      "eval_samples_per_second": 14.485,
      "eval_steps_per_second": 1.867,
      "step": 3175
    },
    {
      "epoch": 8.274078862314157,
      "grad_norm": 0.28934046626091003,
      "learning_rate": 3.420575278569578e-05,
      "loss": 0.0841,
      "step": 3200
    },
    {
      "epoch": 8.274078862314157,
      "eval_loss": 0.41459396481513977,
      "eval_runtime": 13.3855,
      "eval_samples_per_second": 14.493,
      "eval_steps_per_second": 1.868,
      "step": 3200
    },
    {
      "epoch": 8.338720103425986,
      "grad_norm": 0.23310010135173798,
      "learning_rate": 3.291008033169215e-05,
      "loss": 0.0775,
      "step": 3225
    },
    {
      "epoch": 8.338720103425986,
      "eval_loss": 0.41512471437454224,
      "eval_runtime": 13.3818,
      "eval_samples_per_second": 14.497,
      "eval_steps_per_second": 1.868,
      "step": 3225
    },
    {
      "epoch": 8.403361344537815,
      "grad_norm": 0.22439134120941162,
      "learning_rate": 3.1614407877688526e-05,
      "loss": 0.0837,
      "step": 3250
    },
    {
      "epoch": 8.403361344537815,
      "eval_loss": 0.41561707854270935,
      "eval_runtime": 13.3826,
      "eval_samples_per_second": 14.496,
      "eval_steps_per_second": 1.868,
      "step": 3250
    },
    {
      "epoch": 8.468002585649645,
      "grad_norm": 0.2346993237733841,
      "learning_rate": 3.0318735423684895e-05,
      "loss": 0.0799,
      "step": 3275
    },
    {
      "epoch": 8.468002585649645,
      "eval_loss": 0.4051866829395294,
      "eval_runtime": 13.3812,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 3275
    },
    {
      "epoch": 8.532643826761474,
      "grad_norm": 0.22098496556282043,
      "learning_rate": 2.9023062969681264e-05,
      "loss": 0.0849,
      "step": 3300
    },
    {
      "epoch": 8.532643826761474,
      "eval_loss": 0.40784117579460144,
      "eval_runtime": 13.3808,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 3300
    },
    {
      "epoch": 8.597285067873303,
      "grad_norm": 0.2729458808898926,
      "learning_rate": 2.7727390515677637e-05,
      "loss": 0.0803,
      "step": 3325
    },
    {
      "epoch": 8.597285067873303,
      "eval_loss": 0.4111315608024597,
      "eval_runtime": 13.3646,
      "eval_samples_per_second": 14.516,
      "eval_steps_per_second": 1.871,
      "step": 3325
    },
    {
      "epoch": 8.661926308985132,
      "grad_norm": 0.19786176085472107,
      "learning_rate": 2.643171806167401e-05,
      "loss": 0.0849,
      "step": 3350
    },
    {
      "epoch": 8.661926308985132,
      "eval_loss": 0.415315181016922,
      "eval_runtime": 13.3797,
      "eval_samples_per_second": 14.5,
      "eval_steps_per_second": 1.868,
      "step": 3350
    },
    {
      "epoch": 8.726567550096961,
      "grad_norm": 0.26568812131881714,
      "learning_rate": 2.5136045607670382e-05,
      "loss": 0.0806,
      "step": 3375
    },
    {
      "epoch": 8.726567550096961,
      "eval_loss": 0.4111620783805847,
      "eval_runtime": 13.3769,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 3375
    },
    {
      "epoch": 8.791208791208792,
      "grad_norm": 0.2054780125617981,
      "learning_rate": 2.3840373153666755e-05,
      "loss": 0.0881,
      "step": 3400
    },
    {
      "epoch": 8.791208791208792,
      "eval_loss": 0.4074566960334778,
      "eval_runtime": 13.384,
      "eval_samples_per_second": 14.495,
      "eval_steps_per_second": 1.868,
      "step": 3400
    },
    {
      "epoch": 8.855850032320621,
      "grad_norm": 0.29870903491973877,
      "learning_rate": 2.2544700699663127e-05,
      "loss": 0.0806,
      "step": 3425
    },
    {
      "epoch": 8.855850032320621,
      "eval_loss": 0.40841713547706604,
      "eval_runtime": 13.3768,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 3425
    },
    {
      "epoch": 8.92049127343245,
      "grad_norm": 0.21767568588256836,
      "learning_rate": 2.12490282456595e-05,
      "loss": 0.085,
      "step": 3450
    },
    {
      "epoch": 8.92049127343245,
      "eval_loss": 0.4094007909297943,
      "eval_runtime": 13.3757,
      "eval_samples_per_second": 14.504,
      "eval_steps_per_second": 1.869,
      "step": 3450
    },
    {
      "epoch": 8.98513251454428,
      "grad_norm": 0.20575569570064545,
      "learning_rate": 1.995335579165587e-05,
      "loss": 0.0873,
      "step": 3475
    },
    {
      "epoch": 8.98513251454428,
      "eval_loss": 0.40901434421539307,
      "eval_runtime": 13.3813,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 3475
    },
    {
      "epoch": 9.049773755656108,
      "grad_norm": 0.23535218834877014,
      "learning_rate": 1.8657683337652242e-05,
      "loss": 0.0744,
      "step": 3500
    },
    {
      "epoch": 9.049773755656108,
      "eval_loss": 0.41001009941101074,
      "eval_runtime": 13.3767,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 3500
    },
    {
      "epoch": 9.114414996767938,
      "grad_norm": 0.22550158202648163,
      "learning_rate": 1.7362010883648614e-05,
      "loss": 0.0815,
      "step": 3525
    },
    {
      "epoch": 9.114414996767938,
      "eval_loss": 0.4145725667476654,
      "eval_runtime": 13.372,
      "eval_samples_per_second": 14.508,
      "eval_steps_per_second": 1.87,
      "step": 3525
    },
    {
      "epoch": 9.179056237879767,
      "grad_norm": 0.21262843906879425,
      "learning_rate": 1.6066338429644987e-05,
      "loss": 0.0732,
      "step": 3550
    },
    {
      "epoch": 9.179056237879767,
      "eval_loss": 0.4180243909358978,
      "eval_runtime": 13.3877,
      "eval_samples_per_second": 14.491,
      "eval_steps_per_second": 1.867,
      "step": 3550
    },
    {
      "epoch": 9.243697478991596,
      "grad_norm": 0.22902533411979675,
      "learning_rate": 1.477066597564136e-05,
      "loss": 0.0821,
      "step": 3575
    },
    {
      "epoch": 9.243697478991596,
      "eval_loss": 0.4216705858707428,
      "eval_runtime": 13.3663,
      "eval_samples_per_second": 14.514,
      "eval_steps_per_second": 1.87,
      "step": 3575
    },
    {
      "epoch": 9.308338720103427,
      "grad_norm": 0.2216034233570099,
      "learning_rate": 1.347499352163773e-05,
      "loss": 0.0718,
      "step": 3600
    },
    {
      "epoch": 9.308338720103427,
      "eval_loss": 0.4226977825164795,
      "eval_runtime": 13.3768,
      "eval_samples_per_second": 14.503,
      "eval_steps_per_second": 1.869,
      "step": 3600
    },
    {
      "epoch": 9.372979961215256,
      "grad_norm": 0.27297767996788025,
      "learning_rate": 1.2179321067634103e-05,
      "loss": 0.0823,
      "step": 3625
    },
    {
      "epoch": 9.372979961215256,
      "eval_loss": 0.42199140787124634,
      "eval_runtime": 13.3655,
      "eval_samples_per_second": 14.515,
      "eval_steps_per_second": 1.87,
      "step": 3625
    },
    {
      "epoch": 9.437621202327085,
      "grad_norm": 0.2388414740562439,
      "learning_rate": 1.0883648613630474e-05,
      "loss": 0.073,
      "step": 3650
    },
    {
      "epoch": 9.437621202327085,
      "eval_loss": 0.42137593030929565,
      "eval_runtime": 13.3793,
      "eval_samples_per_second": 14.5,
      "eval_steps_per_second": 1.869,
      "step": 3650
    },
    {
      "epoch": 9.502262443438914,
      "grad_norm": 0.25000351667404175,
      "learning_rate": 9.587976159626847e-06,
      "loss": 0.0843,
      "step": 3675
    },
    {
      "epoch": 9.502262443438914,
      "eval_loss": 0.42024168372154236,
      "eval_runtime": 13.3731,
      "eval_samples_per_second": 14.507,
      "eval_steps_per_second": 1.869,
      "step": 3675
    },
    {
      "epoch": 9.566903684550743,
      "grad_norm": 0.24509260058403015,
      "learning_rate": 8.29230370562322e-06,
      "loss": 0.0753,
      "step": 3700
    },
    {
      "epoch": 9.566903684550743,
      "eval_loss": 0.41970470547676086,
      "eval_runtime": 13.378,
      "eval_samples_per_second": 14.501,
      "eval_steps_per_second": 1.869,
      "step": 3700
    },
    {
      "epoch": 9.631544925662572,
      "grad_norm": 0.28298047184944153,
      "learning_rate": 6.996631251619591e-06,
      "loss": 0.0835,
      "step": 3725
    },
    {
      "epoch": 9.631544925662572,
      "eval_loss": 0.4190382957458496,
      "eval_runtime": 13.3719,
      "eval_samples_per_second": 14.508,
      "eval_steps_per_second": 1.87,
      "step": 3725
    },
    {
      "epoch": 9.696186166774401,
      "grad_norm": 0.28533321619033813,
      "learning_rate": 5.700958797615963e-06,
      "loss": 0.0735,
      "step": 3750
    },
    {
      "epoch": 9.696186166774401,
      "eval_loss": 0.4197899401187897,
      "eval_runtime": 13.3702,
      "eval_samples_per_second": 14.51,
      "eval_steps_per_second": 1.87,
      "step": 3750
    },
    {
      "epoch": 9.760827407886232,
      "grad_norm": 0.2747386693954468,
      "learning_rate": 4.4052863436123355e-06,
      "loss": 0.0846,
      "step": 3775
    },
    {
      "epoch": 9.760827407886232,
      "eval_loss": 0.420250803232193,
      "eval_runtime": 13.3816,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 1.868,
      "step": 3775
    }
  ],
  "logging_steps": 25,
  "max_steps": 3860,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 25,
  "total_flos": 1.232289499803648e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}