{
  "best_metric": 0.402925044298172,
  "best_model_checkpoint": "mikhail-panzo/fil_b128_le5_s8000/checkpoint-4500",
  "epoch": 695.6521739130435,
  "eval_steps": 500,
  "global_step": 8000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.3478260869565215,
      "grad_norm": 4.460227966308594,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.814,
      "step": 50
    },
    {
      "epoch": 8.695652173913043,
      "grad_norm": 2.496962785720825,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.7805,
      "step": 100
    },
    {
      "epoch": 13.043478260869565,
      "grad_norm": 2.0162875652313232,
      "learning_rate": 7.5e-07,
      "loss": 0.7591,
      "step": 150
    },
    {
      "epoch": 17.391304347826086,
      "grad_norm": 1.1545039415359497,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.7243,
      "step": 200
    },
    {
      "epoch": 21.73913043478261,
      "grad_norm": 1.3521718978881836,
      "learning_rate": 1.25e-06,
      "loss": 0.7066,
      "step": 250
    },
    {
      "epoch": 26.08695652173913,
      "grad_norm": 2.4172327518463135,
      "learning_rate": 1.5e-06,
      "loss": 0.6976,
      "step": 300
    },
    {
      "epoch": 30.434782608695652,
      "grad_norm": 0.9816296696662903,
      "learning_rate": 1.75e-06,
      "loss": 0.6674,
      "step": 350
    },
    {
      "epoch": 34.78260869565217,
      "grad_norm": 2.370589017868042,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.6427,
      "step": 400
    },
    {
      "epoch": 39.130434782608695,
      "grad_norm": 1.0166665315628052,
      "learning_rate": 2.25e-06,
      "loss": 0.5999,
      "step": 450
    },
    {
      "epoch": 43.47826086956522,
      "grad_norm": 0.9980425238609314,
      "learning_rate": 2.5e-06,
      "loss": 0.5575,
      "step": 500
    },
    {
      "epoch": 43.47826086956522,
      "eval_loss": 0.4795418679714203,
      "eval_runtime": 12.6598,
      "eval_samples_per_second": 12.717,
      "eval_steps_per_second": 1.659,
      "step": 500
    },
    {
      "epoch": 47.82608695652174,
      "grad_norm": 1.1796629428863525,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.5467,
      "step": 550
    },
    {
      "epoch": 52.17391304347826,
      "grad_norm": 1.1312341690063477,
      "learning_rate": 3e-06,
      "loss": 0.5345,
      "step": 600
    },
    {
      "epoch": 56.52173913043478,
      "grad_norm": 0.855106770992279,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.5315,
      "step": 650
    },
    {
      "epoch": 60.869565217391305,
      "grad_norm": 1.4999840259552002,
      "learning_rate": 3.5e-06,
      "loss": 0.5258,
      "step": 700
    },
    {
      "epoch": 65.21739130434783,
      "grad_norm": 1.2242575883865356,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.5167,
      "step": 750
    },
    {
      "epoch": 69.56521739130434,
      "grad_norm": 0.9953985810279846,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.5032,
      "step": 800
    },
    {
      "epoch": 73.91304347826087,
      "grad_norm": 0.9109128713607788,
      "learning_rate": 4.25e-06,
      "loss": 0.5094,
      "step": 850
    },
    {
      "epoch": 78.26086956521739,
      "grad_norm": 1.020887851715088,
      "learning_rate": 4.5e-06,
      "loss": 0.4988,
      "step": 900
    },
    {
      "epoch": 82.6086956521739,
      "grad_norm": 1.0142054557800293,
      "learning_rate": 4.75e-06,
      "loss": 0.4911,
      "step": 950
    },
    {
      "epoch": 86.95652173913044,
      "grad_norm": 0.9454975128173828,
      "learning_rate": 5e-06,
      "loss": 0.4877,
      "step": 1000
    },
    {
      "epoch": 86.95652173913044,
      "eval_loss": 0.4371128976345062,
      "eval_runtime": 12.7935,
      "eval_samples_per_second": 12.585,
      "eval_steps_per_second": 1.641,
      "step": 1000
    },
    {
      "epoch": 91.30434782608695,
      "grad_norm": 1.0903185606002808,
      "learning_rate": 5.2500000000000006e-06,
      "loss": 0.4903,
      "step": 1050
    },
    {
      "epoch": 95.65217391304348,
      "grad_norm": 1.2410277128219604,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.4831,
      "step": 1100
    },
    {
      "epoch": 100.0,
      "grad_norm": 1.5670720338821411,
      "learning_rate": 5.75e-06,
      "loss": 0.4772,
      "step": 1150
    },
    {
      "epoch": 104.34782608695652,
      "grad_norm": 0.8733631372451782,
      "learning_rate": 6e-06,
      "loss": 0.4797,
      "step": 1200
    },
    {
      "epoch": 108.69565217391305,
      "grad_norm": 1.095543384552002,
      "learning_rate": 6.25e-06,
      "loss": 0.4785,
      "step": 1250
    },
    {
      "epoch": 113.04347826086956,
      "grad_norm": 1.7245070934295654,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 0.4734,
      "step": 1300
    },
    {
      "epoch": 117.3913043478261,
      "grad_norm": 1.1396911144256592,
      "learning_rate": 6.750000000000001e-06,
      "loss": 0.4742,
      "step": 1350
    },
    {
      "epoch": 121.73913043478261,
      "grad_norm": 1.195086121559143,
      "learning_rate": 7e-06,
      "loss": 0.4711,
      "step": 1400
    },
    {
      "epoch": 126.08695652173913,
      "grad_norm": 0.9940504431724548,
      "learning_rate": 7.25e-06,
      "loss": 0.4613,
      "step": 1450
    },
    {
      "epoch": 130.43478260869566,
      "grad_norm": 0.9867850542068481,
      "learning_rate": 7.500000000000001e-06,
      "loss": 0.4582,
      "step": 1500
    },
    {
      "epoch": 130.43478260869566,
      "eval_loss": 0.4196487069129944,
      "eval_runtime": 12.7443,
      "eval_samples_per_second": 12.633,
      "eval_steps_per_second": 1.648,
      "step": 1500
    },
    {
      "epoch": 134.7826086956522,
      "grad_norm": 0.9487758874893188,
      "learning_rate": 7.75e-06,
      "loss": 0.4613,
      "step": 1550
    },
    {
      "epoch": 139.1304347826087,
      "grad_norm": 1.4387928247451782,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.4602,
      "step": 1600
    },
    {
      "epoch": 143.47826086956522,
      "grad_norm": 1.1203278303146362,
      "learning_rate": 8.25e-06,
      "loss": 0.4583,
      "step": 1650
    },
    {
      "epoch": 147.82608695652175,
      "grad_norm": 0.8609073162078857,
      "learning_rate": 8.5e-06,
      "loss": 0.4535,
      "step": 1700
    },
    {
      "epoch": 152.17391304347825,
      "grad_norm": 0.8195172548294067,
      "learning_rate": 8.750000000000001e-06,
      "loss": 0.4546,
      "step": 1750
    },
    {
      "epoch": 156.52173913043478,
      "grad_norm": 0.9077767729759216,
      "learning_rate": 9e-06,
      "loss": 0.4466,
      "step": 1800
    },
    {
      "epoch": 160.8695652173913,
      "grad_norm": 0.9574931263923645,
      "learning_rate": 9.250000000000001e-06,
      "loss": 0.4537,
      "step": 1850
    },
    {
      "epoch": 165.2173913043478,
      "grad_norm": 0.8706139326095581,
      "learning_rate": 9.5e-06,
      "loss": 0.449,
      "step": 1900
    },
    {
      "epoch": 169.56521739130434,
      "grad_norm": 1.2113181352615356,
      "learning_rate": 9.75e-06,
      "loss": 0.4513,
      "step": 1950
    },
    {
      "epoch": 173.91304347826087,
      "grad_norm": 1.0548679828643799,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "step": 2000
    },
    {
      "epoch": 173.91304347826087,
      "eval_loss": 0.41451236605644226,
      "eval_runtime": 13.2202,
      "eval_samples_per_second": 12.178,
      "eval_steps_per_second": 1.588,
      "step": 2000
    },
    {
      "epoch": 178.2608695652174,
      "grad_norm": 1.1025744676589966,
      "learning_rate": 9.916666666666668e-06,
      "loss": 0.4463,
      "step": 2050
    },
    {
      "epoch": 182.6086956521739,
      "grad_norm": 1.0485080480575562,
      "learning_rate": 9.833333333333333e-06,
      "loss": 0.4452,
      "step": 2100
    },
    {
      "epoch": 186.95652173913044,
      "grad_norm": 0.9438555240631104,
      "learning_rate": 9.751666666666667e-06,
      "loss": 0.4401,
      "step": 2150
    },
    {
      "epoch": 191.30434782608697,
      "grad_norm": 0.8600581884384155,
      "learning_rate": 9.668333333333334e-06,
      "loss": 0.4406,
      "step": 2200
    },
    {
      "epoch": 195.65217391304347,
      "grad_norm": 0.915958046913147,
      "learning_rate": 9.585e-06,
      "loss": 0.4402,
      "step": 2250
    },
    {
      "epoch": 200.0,
      "grad_norm": 1.931151032447815,
      "learning_rate": 9.501666666666667e-06,
      "loss": 0.439,
      "step": 2300
    },
    {
      "epoch": 204.34782608695653,
      "grad_norm": 1.006366491317749,
      "learning_rate": 9.418333333333334e-06,
      "loss": 0.4386,
      "step": 2350
    },
    {
      "epoch": 208.69565217391303,
      "grad_norm": 1.2179653644561768,
      "learning_rate": 9.335000000000001e-06,
      "loss": 0.4381,
      "step": 2400
    },
    {
      "epoch": 213.04347826086956,
      "grad_norm": 1.779303789138794,
      "learning_rate": 9.251666666666668e-06,
      "loss": 0.4342,
      "step": 2450
    },
    {
      "epoch": 217.3913043478261,
      "grad_norm": 0.9080890417098999,
      "learning_rate": 9.168333333333333e-06,
      "loss": 0.4353,
      "step": 2500
    },
    {
      "epoch": 217.3913043478261,
      "eval_loss": 0.40957868099212646,
      "eval_runtime": 13.4977,
      "eval_samples_per_second": 11.928,
      "eval_steps_per_second": 1.556,
      "step": 2500
    },
    {
      "epoch": 221.7391304347826,
      "grad_norm": 0.7047795057296753,
      "learning_rate": 9.085e-06,
      "loss": 0.4285,
      "step": 2550
    },
    {
      "epoch": 226.08695652173913,
      "grad_norm": 1.3410165309906006,
      "learning_rate": 9.001666666666667e-06,
      "loss": 0.4294,
      "step": 2600
    },
    {
      "epoch": 230.43478260869566,
      "grad_norm": 1.68880033493042,
      "learning_rate": 8.918333333333334e-06,
      "loss": 0.437,
      "step": 2650
    },
    {
      "epoch": 234.7826086956522,
      "grad_norm": 0.8228254914283752,
      "learning_rate": 8.835000000000001e-06,
      "loss": 0.4305,
      "step": 2700
    },
    {
      "epoch": 239.1304347826087,
      "grad_norm": 1.1968687772750854,
      "learning_rate": 8.751666666666668e-06,
      "loss": 0.432,
      "step": 2750
    },
    {
      "epoch": 243.47826086956522,
      "grad_norm": 0.6908459663391113,
      "learning_rate": 8.668333333333335e-06,
      "loss": 0.4282,
      "step": 2800
    },
    {
      "epoch": 247.82608695652175,
      "grad_norm": 0.7675824165344238,
      "learning_rate": 8.585000000000002e-06,
      "loss": 0.427,
      "step": 2850
    },
    {
      "epoch": 252.17391304347825,
      "grad_norm": 0.893234133720398,
      "learning_rate": 8.501666666666667e-06,
      "loss": 0.4278,
      "step": 2900
    },
    {
      "epoch": 256.5217391304348,
      "grad_norm": 0.8852680921554565,
      "learning_rate": 8.418333333333334e-06,
      "loss": 0.4275,
      "step": 2950
    },
    {
      "epoch": 260.8695652173913,
      "grad_norm": 0.8615354299545288,
      "learning_rate": 8.335e-06,
      "loss": 0.4261,
      "step": 3000
    },
    {
      "epoch": 260.8695652173913,
      "eval_loss": 0.40570691227912903,
      "eval_runtime": 13.5692,
      "eval_samples_per_second": 11.865,
      "eval_steps_per_second": 1.548,
      "step": 3000
    },
    {
      "epoch": 265.2173913043478,
      "grad_norm": 1.5136363506317139,
      "learning_rate": 8.251666666666668e-06,
      "loss": 0.426,
      "step": 3050
    },
    {
      "epoch": 269.5652173913044,
      "grad_norm": 1.0311447381973267,
      "learning_rate": 8.168333333333334e-06,
      "loss": 0.4225,
      "step": 3100
    },
    {
      "epoch": 273.9130434782609,
      "grad_norm": 2.1306591033935547,
      "learning_rate": 8.085000000000001e-06,
      "loss": 0.4225,
      "step": 3150
    },
    {
      "epoch": 278.2608695652174,
      "grad_norm": 0.9035950303077698,
      "learning_rate": 8.001666666666668e-06,
      "loss": 0.4227,
      "step": 3200
    },
    {
      "epoch": 282.60869565217394,
      "grad_norm": 1.382899284362793,
      "learning_rate": 7.918333333333333e-06,
      "loss": 0.4222,
      "step": 3250
    },
    {
      "epoch": 286.95652173913044,
      "grad_norm": 0.9895361065864563,
      "learning_rate": 7.835e-06,
      "loss": 0.4215,
      "step": 3300
    },
    {
      "epoch": 291.30434782608694,
      "grad_norm": 0.7423685789108276,
      "learning_rate": 7.751666666666667e-06,
      "loss": 0.4218,
      "step": 3350
    },
    {
      "epoch": 295.6521739130435,
      "grad_norm": 1.230416178703308,
      "learning_rate": 7.668333333333334e-06,
      "loss": 0.4224,
      "step": 3400
    },
    {
      "epoch": 300.0,
      "grad_norm": 1.3802359104156494,
      "learning_rate": 7.585e-06,
      "loss": 0.4205,
      "step": 3450
    },
    {
      "epoch": 304.3478260869565,
      "grad_norm": 0.758198618888855,
      "learning_rate": 7.501666666666667e-06,
      "loss": 0.4206,
      "step": 3500
    },
    {
      "epoch": 304.3478260869565,
      "eval_loss": 0.4048587381839752,
      "eval_runtime": 14.3547,
      "eval_samples_per_second": 11.216,
      "eval_steps_per_second": 1.463,
      "step": 3500
    },
    {
      "epoch": 308.69565217391306,
      "grad_norm": 0.9321383833885193,
      "learning_rate": 7.418333333333334e-06,
      "loss": 0.4189,
      "step": 3550
    },
    {
      "epoch": 313.04347826086956,
      "grad_norm": 1.2097365856170654,
      "learning_rate": 7.335000000000001e-06,
      "loss": 0.4183,
      "step": 3600
    },
    {
      "epoch": 317.39130434782606,
      "grad_norm": 0.8793150782585144,
      "learning_rate": 7.251666666666667e-06,
      "loss": 0.4192,
      "step": 3650
    },
    {
      "epoch": 321.7391304347826,
      "grad_norm": 1.274627923965454,
      "learning_rate": 7.168333333333334e-06,
      "loss": 0.4131,
      "step": 3700
    },
    {
      "epoch": 326.0869565217391,
      "grad_norm": 0.7330735921859741,
      "learning_rate": 7.085000000000001e-06,
      "loss": 0.4173,
      "step": 3750
    },
    {
      "epoch": 330.4347826086956,
      "grad_norm": 0.9770796895027161,
      "learning_rate": 7.001666666666668e-06,
      "loss": 0.4136,
      "step": 3800
    },
    {
      "epoch": 334.7826086956522,
      "grad_norm": 0.8670146465301514,
      "learning_rate": 6.918333333333334e-06,
      "loss": 0.4124,
      "step": 3850
    },
    {
      "epoch": 339.1304347826087,
      "grad_norm": 1.0471419095993042,
      "learning_rate": 6.835000000000001e-06,
      "loss": 0.4152,
      "step": 3900
    },
    {
      "epoch": 343.4782608695652,
      "grad_norm": 0.6883171796798706,
      "learning_rate": 6.7516666666666675e-06,
      "loss": 0.4149,
      "step": 3950
    },
    {
      "epoch": 347.82608695652175,
      "grad_norm": 0.8158336877822876,
      "learning_rate": 6.668333333333334e-06,
      "loss": 0.4117,
      "step": 4000
    },
    {
      "epoch": 347.82608695652175,
      "eval_loss": 0.40652915835380554,
      "eval_runtime": 13.3468,
      "eval_samples_per_second": 12.063,
      "eval_steps_per_second": 1.573,
      "step": 4000
    },
    {
      "epoch": 352.17391304347825,
      "grad_norm": 0.8498961925506592,
      "learning_rate": 6.5850000000000005e-06,
      "loss": 0.4104,
      "step": 4050
    },
    {
      "epoch": 356.5217391304348,
      "grad_norm": 0.8166142702102661,
      "learning_rate": 6.501666666666667e-06,
      "loss": 0.4124,
      "step": 4100
    },
    {
      "epoch": 360.8695652173913,
      "grad_norm": 1.0088080167770386,
      "learning_rate": 6.418333333333334e-06,
      "loss": 0.4093,
      "step": 4150
    },
    {
      "epoch": 365.2173913043478,
      "grad_norm": 0.7711963653564453,
      "learning_rate": 6.335e-06,
      "loss": 0.4123,
      "step": 4200
    },
    {
      "epoch": 369.5652173913044,
      "grad_norm": 0.6871639490127563,
      "learning_rate": 6.253333333333333e-06,
      "loss": 0.4173,
      "step": 4250
    },
    {
      "epoch": 373.9130434782609,
      "grad_norm": 0.8275335431098938,
      "learning_rate": 6.17e-06,
      "loss": 0.4151,
      "step": 4300
    },
    {
      "epoch": 378.2608695652174,
      "grad_norm": 0.7653161883354187,
      "learning_rate": 6.086666666666667e-06,
      "loss": 0.4129,
      "step": 4350
    },
    {
      "epoch": 382.60869565217394,
      "grad_norm": 0.8653395771980286,
      "learning_rate": 6.003333333333334e-06,
      "loss": 0.4109,
      "step": 4400
    },
    {
      "epoch": 386.95652173913044,
      "grad_norm": 0.8151092529296875,
      "learning_rate": 5.92e-06,
      "loss": 0.4141,
      "step": 4450
    },
    {
      "epoch": 391.30434782608694,
      "grad_norm": 0.7493378520011902,
      "learning_rate": 5.836666666666667e-06,
      "loss": 0.4123,
      "step": 4500
    },
    {
      "epoch": 391.30434782608694,
      "eval_loss": 0.402925044298172,
      "eval_runtime": 13.3404,
      "eval_samples_per_second": 12.069,
      "eval_steps_per_second": 1.574,
      "step": 4500
    },
    {
      "epoch": 395.6521739130435,
      "grad_norm": 0.9856454133987427,
      "learning_rate": 5.753333333333334e-06,
      "loss": 0.4176,
      "step": 4550
    },
    {
      "epoch": 400.0,
      "grad_norm": 1.347724437713623,
      "learning_rate": 5.67e-06,
      "loss": 0.4077,
      "step": 4600
    },
    {
      "epoch": 404.3478260869565,
      "grad_norm": 0.6690952777862549,
      "learning_rate": 5.586666666666667e-06,
      "loss": 0.4099,
      "step": 4650
    },
    {
      "epoch": 408.69565217391306,
      "grad_norm": 0.7244478464126587,
      "learning_rate": 5.503333333333334e-06,
      "loss": 0.411,
      "step": 4700
    },
    {
      "epoch": 413.04347826086956,
      "grad_norm": 2.4231059551239014,
      "learning_rate": 5.420000000000001e-06,
      "loss": 0.4133,
      "step": 4750
    },
    {
      "epoch": 417.39130434782606,
      "grad_norm": 0.7757388353347778,
      "learning_rate": 5.336666666666667e-06,
      "loss": 0.4121,
      "step": 4800
    },
    {
      "epoch": 421.7391304347826,
      "grad_norm": 0.9179409146308899,
      "learning_rate": 5.2533333333333336e-06,
      "loss": 0.4089,
      "step": 4850
    },
    {
      "epoch": 426.0869565217391,
      "grad_norm": 1.120680809020996,
      "learning_rate": 5.1700000000000005e-06,
      "loss": 0.4141,
      "step": 4900
    },
    {
      "epoch": 430.4347826086956,
      "grad_norm": 0.9229013323783875,
      "learning_rate": 5.086666666666667e-06,
      "loss": 0.4086,
      "step": 4950
    },
    {
      "epoch": 434.7826086956522,
      "grad_norm": 0.7975584864616394,
      "learning_rate": 5.0033333333333334e-06,
      "loss": 0.4117,
      "step": 5000
    },
    {
      "epoch": 434.7826086956522,
      "eval_loss": 0.40458473563194275,
      "eval_runtime": 13.8621,
      "eval_samples_per_second": 11.614,
      "eval_steps_per_second": 1.515,
      "step": 5000
    },
    {
      "epoch": 439.1304347826087,
      "grad_norm": 0.6378179788589478,
      "learning_rate": 4.92e-06,
      "loss": 0.4081,
      "step": 5050
    },
    {
      "epoch": 443.4782608695652,
      "grad_norm": 1.1477278470993042,
      "learning_rate": 4.836666666666667e-06,
      "loss": 0.4077,
      "step": 5100
    },
    {
      "epoch": 447.82608695652175,
      "grad_norm": 0.8010798692703247,
      "learning_rate": 4.753333333333333e-06,
      "loss": 0.41,
      "step": 5150
    },
    {
      "epoch": 452.17391304347825,
      "grad_norm": 0.8638696074485779,
      "learning_rate": 4.670000000000001e-06,
      "loss": 0.41,
      "step": 5200
    },
    {
      "epoch": 456.5217391304348,
      "grad_norm": 0.6933819055557251,
      "learning_rate": 4.586666666666667e-06,
      "loss": 0.4063,
      "step": 5250
    },
    {
      "epoch": 460.8695652173913,
      "grad_norm": 0.8501341938972473,
      "learning_rate": 4.503333333333333e-06,
      "loss": 0.4077,
      "step": 5300
    },
    {
      "epoch": 465.2173913043478,
      "grad_norm": 0.6721881031990051,
      "learning_rate": 4.42e-06,
      "loss": 0.4069,
      "step": 5350
    },
    {
      "epoch": 469.5652173913044,
      "grad_norm": 1.053196668624878,
      "learning_rate": 4.336666666666667e-06,
      "loss": 0.4045,
      "step": 5400
    },
    {
      "epoch": 473.9130434782609,
      "grad_norm": 1.0177547931671143,
      "learning_rate": 4.253333333333334e-06,
      "loss": 0.4038,
      "step": 5450
    },
    {
      "epoch": 478.2608695652174,
      "grad_norm": 0.8582258820533752,
      "learning_rate": 4.17e-06,
      "loss": 0.4053,
      "step": 5500
    },
    {
      "epoch": 478.2608695652174,
      "eval_loss": 0.4044737219810486,
      "eval_runtime": 12.7138,
      "eval_samples_per_second": 12.663,
      "eval_steps_per_second": 1.652,
      "step": 5500
    },
    {
      "epoch": 482.60869565217394,
      "grad_norm": 0.7471197843551636,
      "learning_rate": 4.086666666666667e-06,
      "loss": 0.4008,
      "step": 5550
    },
    {
      "epoch": 486.95652173913044,
      "grad_norm": 0.7691963315010071,
      "learning_rate": 4.003333333333334e-06,
      "loss": 0.4076,
      "step": 5600
    },
    {
      "epoch": 491.30434782608694,
      "grad_norm": 0.7329992651939392,
      "learning_rate": 3.920000000000001e-06,
      "loss": 0.4039,
      "step": 5650
    },
    {
      "epoch": 495.6521739130435,
      "grad_norm": 1.2654510736465454,
      "learning_rate": 3.836666666666667e-06,
      "loss": 0.4057,
      "step": 5700
    },
    {
      "epoch": 500.0,
      "grad_norm": 2.19535756111145,
      "learning_rate": 3.753333333333334e-06,
      "loss": 0.4082,
      "step": 5750
    },
    {
      "epoch": 504.3478260869565,
      "grad_norm": 0.962538480758667,
      "learning_rate": 3.6700000000000004e-06,
      "loss": 0.4043,
      "step": 5800
    },
    {
      "epoch": 508.69565217391306,
      "grad_norm": 0.9239832162857056,
      "learning_rate": 3.5866666666666673e-06,
      "loss": 0.4069,
      "step": 5850
    },
    {
      "epoch": 513.0434782608696,
      "grad_norm": 2.786410331726074,
      "learning_rate": 3.5033333333333334e-06,
      "loss": 0.4081,
      "step": 5900
    },
    {
      "epoch": 517.3913043478261,
      "grad_norm": 0.7412209510803223,
      "learning_rate": 3.4200000000000007e-06,
      "loss": 0.4028,
      "step": 5950
    },
    {
      "epoch": 521.7391304347826,
      "grad_norm": 0.8102017045021057,
      "learning_rate": 3.3366666666666668e-06,
      "loss": 0.4064,
      "step": 6000
    },
    {
      "epoch": 521.7391304347826,
      "eval_loss": 0.4046972692012787,
      "eval_runtime": 12.7264,
      "eval_samples_per_second": 12.651,
      "eval_steps_per_second": 1.65,
      "step": 6000
    },
    {
      "epoch": 526.0869565217391,
      "grad_norm": 1.0029244422912598,
      "learning_rate": 3.2533333333333332e-06,
      "loss": 0.4032,
      "step": 6050
    },
    {
      "epoch": 530.4347826086956,
      "grad_norm": 0.7244696617126465,
      "learning_rate": 3.17e-06,
      "loss": 0.4082,
      "step": 6100
    },
    {
      "epoch": 534.7826086956521,
      "grad_norm": 0.7756295204162598,
      "learning_rate": 3.0866666666666666e-06,
      "loss": 0.4063,
      "step": 6150
    },
    {
      "epoch": 539.1304347826087,
      "grad_norm": 0.6352535486221313,
      "learning_rate": 3.0033333333333335e-06,
      "loss": 0.408,
      "step": 6200
    },
    {
      "epoch": 543.4782608695652,
      "grad_norm": 0.8275020122528076,
      "learning_rate": 2.92e-06,
      "loss": 0.4047,
      "step": 6250
    },
    {
      "epoch": 547.8260869565217,
      "grad_norm": 0.6523252129554749,
      "learning_rate": 2.836666666666667e-06,
      "loss": 0.4054,
      "step": 6300
    },
    {
      "epoch": 552.1739130434783,
      "grad_norm": 0.864088237285614,
      "learning_rate": 2.7533333333333334e-06,
      "loss": 0.4043,
      "step": 6350
    },
    {
      "epoch": 556.5217391304348,
      "grad_norm": 0.6934670805931091,
      "learning_rate": 2.6700000000000003e-06,
      "loss": 0.4014,
      "step": 6400
    },
    {
      "epoch": 560.8695652173913,
      "grad_norm": 0.6325249671936035,
      "learning_rate": 2.5866666666666667e-06,
      "loss": 0.4002,
      "step": 6450
    },
    {
      "epoch": 565.2173913043479,
      "grad_norm": 0.6945725083351135,
      "learning_rate": 2.505e-06,
      "loss": 0.4058,
      "step": 6500
    },
    {
      "epoch": 565.2173913043479,
      "eval_loss": 0.4041679799556732,
      "eval_runtime": 12.8417,
      "eval_samples_per_second": 12.537,
      "eval_steps_per_second": 1.635,
      "step": 6500
    },
    {
      "epoch": 569.5652173913044,
      "grad_norm": 0.7380744218826294,
      "learning_rate": 2.421666666666667e-06,
      "loss": 0.407,
      "step": 6550
    },
    {
      "epoch": 573.9130434782609,
      "grad_norm": 2.16170597076416,
      "learning_rate": 2.3383333333333335e-06,
      "loss": 0.3985,
      "step": 6600
    },
    {
      "epoch": 578.2608695652174,
      "grad_norm": 0.753318727016449,
      "learning_rate": 2.2550000000000004e-06,
      "loss": 0.4015,
      "step": 6650
    },
    {
      "epoch": 582.6086956521739,
      "grad_norm": 0.8654139637947083,
      "learning_rate": 2.171666666666667e-06,
      "loss": 0.4024,
      "step": 6700
    },
    {
      "epoch": 586.9565217391304,
      "grad_norm": 0.6627745032310486,
      "learning_rate": 2.088333333333334e-06,
      "loss": 0.4013,
      "step": 6750
    },
    {
      "epoch": 591.304347826087,
      "grad_norm": 0.7147239446640015,
      "learning_rate": 2.0050000000000003e-06,
      "loss": 0.3997,
      "step": 6800
    },
    {
      "epoch": 595.6521739130435,
      "grad_norm": 0.6943681836128235,
      "learning_rate": 1.9216666666666668e-06,
      "loss": 0.4043,
      "step": 6850
    },
    {
      "epoch": 600.0,
      "grad_norm": 0.94109046459198,
      "learning_rate": 1.8383333333333334e-06,
      "loss": 0.4031,
      "step": 6900
    },
    {
      "epoch": 604.3478260869565,
      "grad_norm": 0.6653702855110168,
      "learning_rate": 1.7550000000000001e-06,
      "loss": 0.3996,
      "step": 6950
    },
    {
      "epoch": 608.695652173913,
      "grad_norm": 0.8962767720222473,
      "learning_rate": 1.6716666666666666e-06,
      "loss": 0.3995,
      "step": 7000
    },
    {
      "epoch": 608.695652173913,
      "eval_loss": 0.4058188796043396,
      "eval_runtime": 12.7074,
      "eval_samples_per_second": 12.67,
      "eval_steps_per_second": 1.653,
      "step": 7000
    },
    {
      "epoch": 613.0434782608696,
      "grad_norm": 1.6197575330734253,
      "learning_rate": 1.5883333333333333e-06,
      "loss": 0.4003,
      "step": 7050
    },
    {
      "epoch": 617.3913043478261,
      "grad_norm": 0.6977828741073608,
      "learning_rate": 1.505e-06,
      "loss": 0.3998,
      "step": 7100
    },
    {
      "epoch": 621.7391304347826,
      "grad_norm": 0.665290355682373,
      "learning_rate": 1.4216666666666667e-06,
      "loss": 0.4003,
      "step": 7150
    },
    {
      "epoch": 626.0869565217391,
      "grad_norm": 0.7739102840423584,
      "learning_rate": 1.3383333333333334e-06,
      "loss": 0.4033,
      "step": 7200
    },
    {
      "epoch": 630.4347826086956,
      "grad_norm": 0.6389193534851074,
      "learning_rate": 1.255e-06,
      "loss": 0.4039,
      "step": 7250
    },
    {
      "epoch": 634.7826086956521,
      "grad_norm": 0.6950168609619141,
      "learning_rate": 1.1716666666666667e-06,
      "loss": 0.4017,
      "step": 7300
    },
    {
      "epoch": 639.1304347826087,
      "grad_norm": 0.6477563381195068,
      "learning_rate": 1.0883333333333334e-06,
      "loss": 0.3986,
      "step": 7350
    },
    {
      "epoch": 643.4782608695652,
      "grad_norm": 0.5801939964294434,
      "learning_rate": 1.0050000000000001e-06,
      "loss": 0.3996,
      "step": 7400
    },
    {
      "epoch": 647.8260869565217,
      "grad_norm": 0.6561626195907593,
      "learning_rate": 9.216666666666667e-07,
      "loss": 0.4035,
      "step": 7450
    },
    {
      "epoch": 652.1739130434783,
      "grad_norm": 0.6258853077888489,
      "learning_rate": 8.383333333333334e-07,
      "loss": 0.4076,
      "step": 7500
    },
    {
      "epoch": 652.1739130434783,
      "eval_loss": 0.40540868043899536,
      "eval_runtime": 13.1806,
      "eval_samples_per_second": 12.215,
      "eval_steps_per_second": 1.593,
      "step": 7500
    },
    {
      "epoch": 656.5217391304348,
      "grad_norm": 0.6540374159812927,
      "learning_rate": 7.550000000000001e-07,
      "loss": 0.4029,
      "step": 7550
    },
    {
      "epoch": 660.8695652173913,
      "grad_norm": 0.618146538734436,
      "learning_rate": 6.716666666666668e-07,
      "loss": 0.4018,
      "step": 7600
    },
    {
      "epoch": 665.2173913043479,
      "grad_norm": 0.7216867804527283,
      "learning_rate": 5.883333333333333e-07,
      "loss": 0.4012,
      "step": 7650
    },
    {
      "epoch": 669.5652173913044,
      "grad_norm": 0.6385495662689209,
      "learning_rate": 5.05e-07,
      "loss": 0.4028,
      "step": 7700
    },
    {
      "epoch": 673.9130434782609,
      "grad_norm": 0.7671997547149658,
      "learning_rate": 4.2166666666666667e-07,
      "loss": 0.4028,
      "step": 7750
    },
    {
      "epoch": 678.2608695652174,
      "grad_norm": 0.6697407960891724,
      "learning_rate": 3.3833333333333336e-07,
      "loss": 0.4024,
      "step": 7800
    },
    {
      "epoch": 682.6086956521739,
      "grad_norm": 0.6977794170379639,
      "learning_rate": 2.55e-07,
      "loss": 0.3988,
      "step": 7850
    },
    {
      "epoch": 686.9565217391304,
      "grad_norm": 0.6256734132766724,
      "learning_rate": 1.7166666666666668e-07,
      "loss": 0.3989,
      "step": 7900
    },
    {
      "epoch": 691.304347826087,
      "grad_norm": 0.7186506390571594,
      "learning_rate": 8.833333333333334e-08,
      "loss": 0.4027,
      "step": 7950
    },
    {
      "epoch": 695.6521739130435,
      "grad_norm": 0.7653046250343323,
      "learning_rate": 5e-09,
      "loss": 0.4042,
      "step": 8000
    },
    {
      "epoch": 695.6521739130435,
      "eval_loss": 0.4052514135837555,
      "eval_runtime": 13.2988,
      "eval_samples_per_second": 12.106,
      "eval_steps_per_second": 1.579,
      "step": 8000
    }
  ],
  "logging_steps": 50,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 728,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.9390598333064925e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}