{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 3933,
  "global_step": 3934,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_loss": 8.882436752319336,
      "eval_ppl": 7204.32452,
      "eval_runtime": 251.4297,
      "eval_samples_per_second": 1161.705,
      "eval_steps_per_second": 24.206,
      "memory/device_reserved (GiB)": 40.85,
      "memory/max_active (GiB)": 40.75,
      "memory/max_allocated (GiB)": 40.75,
      "step": 0
    },
    {
      "epoch": 0.012710921859607868,
      "grad_norm": 5.21875,
      "learning_rate": 0.00019999459941110254,
      "loss": 6.907210083007812,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 999.45495,
      "step": 50,
      "tokens/total": 78626816,
      "tokens/train_per_sec_per_gpu": 1148.16,
      "tokens/trainable": 78551592
    },
    {
      "epoch": 0.025421843719215735,
      "grad_norm": 2.15625,
      "learning_rate": 0.00019997187610487207,
      "loss": 5.735242919921875,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 309.58817,
      "step": 100,
      "tokens/total": 157257728,
      "tokens/train_per_sec_per_gpu": 1149.32,
      "tokens/trainable": 157108256
    },
    {
      "epoch": 0.038132765578823606,
      "grad_norm": 1.453125,
      "learning_rate": 0.00019993140447924126,
      "loss": 5.411905517578125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 224.05813,
      "step": 150,
      "tokens/total": 235888640,
      "tokens/train_per_sec_per_gpu": 1154.11,
      "tokens/trainable": 235656864
    },
    {
      "epoch": 0.05084368743843147,
      "grad_norm": 1.25,
      "learning_rate": 0.00019987319171926423,
      "loss": 5.301961669921875,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 200.73019,
      "step": 200,
      "tokens/total": 314519552,
      "tokens/train_per_sec_per_gpu": 1152.76,
      "tokens/trainable": 314208672
    },
    {
      "epoch": 0.06355460929803934,
      "grad_norm": 1.5234375,
      "learning_rate": 0.00019979724815963408,
      "loss": 5.258994140625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 192.28798,
      "step": 250,
      "tokens/total": 393146368,
      "tokens/train_per_sec_per_gpu": 738.98,
      "tokens/trainable": 392753440
    },
    {
      "epoch": 0.07626553115764721,
      "grad_norm": 1.078125,
      "learning_rate": 0.0001997035872828481,
      "loss": 5.239044189453125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 188.48986,
      "step": 300,
      "tokens/total": 471785472,
      "tokens/train_per_sec_per_gpu": 1147.68,
      "tokens/trainable": 471316160
    },
    {
      "epoch": 0.08897645301725508,
      "grad_norm": 1.015625,
      "learning_rate": 0.00019959222571681436,
      "loss": 5.224154663085938,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 185.70412,
      "step": 350,
      "tokens/total": 550420480,
      "tokens/train_per_sec_per_gpu": 1149.3,
      "tokens/trainable": 549873472
    },
    {
      "epoch": 0.10168737487686294,
      "grad_norm": 1.359375,
      "learning_rate": 0.00019946318323189936,
      "loss": 5.208404541015625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 182.80217,
      "step": 400,
      "tokens/total": 629022720,
      "tokens/train_per_sec_per_gpu": 1151.54,
      "tokens/trainable": 628395264
    },
    {
      "epoch": 0.11439829673647081,
      "grad_norm": 1.546875,
      "learning_rate": 0.00019931648273741856,
      "loss": 5.199439697265625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 181.1707,
      "step": 450,
      "tokens/total": 707637248,
      "tokens/train_per_sec_per_gpu": 1012.12,
      "tokens/trainable": 706928384
    },
    {
      "epoch": 0.12710921859607868,
      "grad_norm": 1.8359375,
      "learning_rate": 0.00019915215027756892,
      "loss": 5.187381591796875,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 178.99925,
      "step": 500,
      "tokens/total": 786255872,
      "tokens/train_per_sec_per_gpu": 1148.12,
      "tokens/trainable": 785466688
    },
    {
      "epoch": 0.13982014045568655,
      "grad_norm": 1.75,
      "learning_rate": 0.00019897021502680525,
      "loss": 5.182908935546875,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 178.20043,
      "step": 550,
      "tokens/total": 864882688,
      "tokens/train_per_sec_per_gpu": 1150.21,
      "tokens/trainable": 864011904
    },
    {
      "epoch": 0.15253106231529442,
      "grad_norm": 5.6875,
      "learning_rate": 0.00019877070928466087,
      "loss": 5.1762158203125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 177.0117,
      "step": 600,
      "tokens/total": 943525888,
      "tokens/train_per_sec_per_gpu": 1149.34,
      "tokens/trainable": 942579392
    },
    {
      "epoch": 0.1652419841749023,
      "grad_norm": 1.390625,
      "learning_rate": 0.00019855366847001328,
      "loss": 5.169808349609375,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 175.88113,
      "step": 650,
      "tokens/total": 1022140416,
      "tokens/train_per_sec_per_gpu": 1151.84,
      "tokens/trainable": 1021113408
    },
    {
      "epoch": 0.17795290603451017,
      "grad_norm": 1.2734375,
      "learning_rate": 0.00019831913111479618,
      "loss": 5.165040283203125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 175.04451,
      "step": 700,
      "tokens/total": 1100767232,
      "tokens/train_per_sec_per_gpu": 1152.16,
      "tokens/trainable": 1099662208
    },
    {
      "epoch": 0.190663827894118,
      "grad_norm": 1.75,
      "learning_rate": 0.00019806713885715875,
      "loss": 5.161105346679688,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 174.35707,
      "step": 750,
      "tokens/total": 1179402240,
      "tokens/train_per_sec_per_gpu": 1143.29,
      "tokens/trainable": 1178215296
    },
    {
      "epoch": 0.20337474975372588,
      "grad_norm": 1.3046875,
      "learning_rate": 0.00019779773643407352,
      "loss": 5.1522760009765625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 172.82439,
      "step": 800,
      "tokens/total": 1258024960,
      "tokens/train_per_sec_per_gpu": 1140.81,
      "tokens/trainable": 1256756096
    },
    {
      "epoch": 0.21608567161333375,
      "grad_norm": 0.796875,
      "learning_rate": 0.00019751097167339406,
      "loss": 5.149384765625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 172.32544,
      "step": 850,
      "tokens/total": 1336647680,
      "tokens/train_per_sec_per_gpu": 1101.91,
      "tokens/trainable": 1335300352
    },
    {
      "epoch": 0.22879659347294162,
      "grad_norm": 1.2421875,
      "learning_rate": 0.00019720689548536398,
      "loss": 5.145052490234375,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 171.58049,
      "step": 900,
      "tokens/total": 1415266304,
      "tokens/train_per_sec_per_gpu": 1149.04,
      "tokens/trainable": 1413838848
    },
    {
      "epoch": 0.2415075153325495,
      "grad_norm": 0.73828125,
      "learning_rate": 0.00019688556185357863,
      "loss": 5.139707641601563,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 170.66587,
      "step": 950,
      "tokens/total": 1493872640,
      "tokens/train_per_sec_per_gpu": 891.45,
      "tokens/trainable": 1492370688
    },
    {
      "epoch": 0.25421843719215736,
      "grad_norm": 0.8359375,
      "learning_rate": 0.00019654702782540126,
      "loss": 5.133409423828125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 169.59435,
      "step": 1000,
      "tokens/total": 1572495360,
      "tokens/train_per_sec_per_gpu": 1148.59,
      "tokens/trainable": 1570911360
    },
    {
      "epoch": 0.26692935905176524,
      "grad_norm": 0.9609375,
      "learning_rate": 0.00019619135350183525,
      "loss": 5.134027099609375,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 169.69914,
      "step": 1050,
      "tokens/total": 1651118080,
      "tokens/train_per_sec_per_gpu": 1148.45,
      "tokens/trainable": 1649451776
    },
    {
      "epoch": 0.2796402809113731,
      "grad_norm": 0.87109375,
      "learning_rate": 0.00019581860202685402,
      "loss": 5.128356323242188,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 168.73954,
      "step": 1100,
      "tokens/total": 1729744896,
      "tokens/train_per_sec_per_gpu": 690.35,
      "tokens/trainable": 1727996928
    },
    {
      "epoch": 0.292351202770981,
      "grad_norm": 0.63671875,
      "learning_rate": 0.00019542883957619122,
      "loss": 5.123819580078125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 167.97574,
      "step": 1150,
      "tokens/total": 1808388096,
      "tokens/train_per_sec_per_gpu": 881.69,
      "tokens/trainable": 1806564992
    },
    {
      "epoch": 0.30506212463058885,
      "grad_norm": 1.6328125,
      "learning_rate": 0.00019502213534559198,
      "loss": 5.119311828613281,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 167.22025,
      "step": 1200,
      "tokens/total": 1887014912,
      "tokens/train_per_sec_per_gpu": 1140.25,
      "tokens/trainable": 1885116160
    },
    {
      "epoch": 0.3177730464901967,
      "grad_norm": 0.9921875,
      "learning_rate": 0.00019459856153852863,
      "loss": 5.118416137695313,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 167.07054,
      "step": 1250,
      "tokens/total": 1965645824,
      "tokens/train_per_sec_per_gpu": 1141.31,
      "tokens/trainable": 1963666176
    },
    {
      "epoch": 0.3304839683498046,
      "grad_norm": 1.7109375,
      "learning_rate": 0.000194158193353382,
      "loss": 5.116240844726563,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 166.70751,
      "step": 1300,
      "tokens/total": 2044252160,
      "tokens/train_per_sec_per_gpu": 1148.85,
      "tokens/trainable": 2042193664
    },
    {
      "epoch": 0.34319489020941246,
      "grad_norm": 1.1875,
      "learning_rate": 0.0001937011089700914,
      "loss": 5.1140374755859375,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 166.3406,
      "step": 1350,
      "tokens/total": 2122862592,
      "tokens/train_per_sec_per_gpu": 1143.78,
      "tokens/trainable": 2120723840
    },
    {
      "epoch": 0.35590581206902033,
      "grad_norm": 0.95703125,
      "learning_rate": 0.000193227389536275,
      "loss": 5.111378479003906,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 165.89889,
      "step": 1400,
      "tokens/total": 2201497600,
      "tokens/train_per_sec_per_gpu": 900.16,
      "tokens/trainable": 2199277824
    },
    {
      "epoch": 0.36861673392862815,
      "grad_norm": 1.0,
      "learning_rate": 0.00019273711915282339,
      "loss": 5.110181274414063,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 165.70039,
      "step": 1450,
      "tokens/total": 2280120320,
      "tokens/train_per_sec_per_gpu": 1153.87,
      "tokens/trainable": 2277814272
    },
    {
      "epoch": 0.381327655788236,
      "grad_norm": 1.0234375,
      "learning_rate": 0.00019223038485896896,
      "loss": 5.105591430664062,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 164.94159,
      "step": 1500,
      "tokens/total": 2358730752,
      "tokens/train_per_sec_per_gpu": 1137.7,
      "tokens/trainable": 2356336384
    },
    {
      "epoch": 0.3940385776478439,
      "grad_norm": 0.890625,
      "learning_rate": 0.00019170727661683358,
      "loss": 5.106539916992188,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 165.09811,
      "step": 1550,
      "tokens/total": 2437357568,
      "tokens/train_per_sec_per_gpu": 1150.48,
      "tokens/trainable": 2434874624
    },
    {
      "epoch": 0.40674949950745176,
      "grad_norm": 1.6796875,
      "learning_rate": 0.00019116788729545728,
      "loss": 5.101802978515625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 164.3179,
      "step": 1600,
      "tokens/total": 2515980288,
      "tokens/train_per_sec_per_gpu": 1008.97,
      "tokens/trainable": 2513408512
    },
    {
      "epoch": 0.41946042136705963,
      "grad_norm": 0.76953125,
      "learning_rate": 0.00019061231265431086,
      "loss": 5.1004248046875,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 164.0916,
      "step": 1650,
      "tokens/total": 2594619392,
      "tokens/train_per_sec_per_gpu": 996.46,
      "tokens/trainable": 2591954176
    },
    {
      "epoch": 0.4321713432266675,
      "grad_norm": 1.015625,
      "learning_rate": 0.0001900406513262956,
      "loss": 5.095927734375,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 163.35532,
      "step": 1700,
      "tokens/total": 2673242112,
      "tokens/train_per_sec_per_gpu": 1143.08,
      "tokens/trainable": 2670494464
    },
    {
      "epoch": 0.4448822650862754,
      "grad_norm": 1.0625,
      "learning_rate": 0.0001894530048002325,
      "loss": 5.0997857666015625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 163.98677,
      "step": 1750,
      "tokens/total": 2751873024,
      "tokens/train_per_sec_per_gpu": 1147.88,
      "tokens/trainable": 2749035264
    },
    {
      "epoch": 0.45759318694588325,
      "grad_norm": 1.171875,
      "learning_rate": 0.00018884947740284465,
      "loss": 5.0969171142578125,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 163.51703,
      "step": 1800,
      "tokens/total": 2830499840,
      "tokens/train_per_sec_per_gpu": 1003.19,
      "tokens/trainable": 2827570944
    },
    {
      "epoch": 0.4703041088054911,
      "grad_norm": 0.875,
      "learning_rate": 0.00018823017628023588,
      "loss": 5.09571044921875,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 163.31983,
      "step": 1850,
      "tokens/total": 2909130752,
      "tokens/train_per_sec_per_gpu": 899.92,
      "tokens/trainable": 2906111744
    },
    {
      "epoch": 0.483015030665099,
      "grad_norm": 0.7890625,
      "learning_rate": 0.00018759521137886874,
      "loss": 5.092830200195312,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 162.85011,
      "step": 1900,
      "tokens/total": 2987749376,
      "tokens/train_per_sec_per_gpu": 1144.23,
      "tokens/trainable": 2984636672
    },
    {
      "epoch": 0.49572595252470686,
      "grad_norm": 1.1171875,
      "learning_rate": 0.00018694469542604526,
      "loss": 5.093046264648438,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 162.8853,
      "step": 1950,
      "tokens/total": 3066388480,
      "tokens/train_per_sec_per_gpu": 1139.03,
      "tokens/trainable": 3063186944
    },
    {
      "epoch": 0.5084368743843147,
      "grad_norm": 0.7578125,
      "learning_rate": 0.00018627874390989425,
      "loss": 5.092771301269531,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 162.84052,
      "step": 2000,
      "tokens/total": 3145011200,
      "tokens/train_per_sec_per_gpu": 1148.33,
      "tokens/trainable": 3141722368
    },
    {
      "epoch": 0.5211477962439226,
      "grad_norm": 0.734375,
      "learning_rate": 0.0001855974750588684,
      "loss": 5.088687438964843,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 162.17686,
      "step": 2050,
      "tokens/total": 3223629824,
      "tokens/train_per_sec_per_gpu": 895.72,
      "tokens/trainable": 3220251904
    },
    {
      "epoch": 0.5338587181035305,
      "grad_norm": 0.94140625,
      "learning_rate": 0.00018490100982075446,
      "loss": 5.087530822753906,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.98939,
      "step": 2100,
      "tokens/total": 3302248448,
      "tokens/train_per_sec_per_gpu": 887.24,
      "tokens/trainable": 3298778880
    },
    {
      "epoch": 0.5465696399631383,
      "grad_norm": 0.83203125,
      "learning_rate": 0.00018418947184120147,
      "loss": 5.090943908691406,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 162.54322,
      "step": 2150,
      "tokens/total": 3380875264,
      "tokens/train_per_sec_per_gpu": 1144.27,
      "tokens/trainable": 3377318656
    },
    {
      "epoch": 0.5592805618227462,
      "grad_norm": 0.98828125,
      "learning_rate": 0.0001834629874417692,
      "loss": 5.08627197265625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.7856,
      "step": 2200,
      "tokens/total": 3459510272,
      "tokens/train_per_sec_per_gpu": 1141.64,
      "tokens/trainable": 3455864576
    },
    {
      "epoch": 0.5719914836823541,
      "grad_norm": 1.640625,
      "learning_rate": 0.00018272168559750206,
      "loss": 5.087280883789062,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.9489,
      "step": 2250,
      "tokens/total": 3538141184,
      "tokens/train_per_sec_per_gpu": 1137.59,
      "tokens/trainable": 3534408704
    },
    {
      "epoch": 0.584702405541962,
      "grad_norm": 0.74609375,
      "learning_rate": 0.00018196569791403174,
      "loss": 5.0849853515625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.57757,
      "step": 2300,
      "tokens/total": 3616763904,
      "tokens/train_per_sec_per_gpu": 889.48,
      "tokens/trainable": 3612942848
    },
    {
      "epoch": 0.5974133274015698,
      "grad_norm": 0.96484375,
      "learning_rate": 0.0001811951586042128,
      "loss": 5.08361572265625,
      "memory/device_reserved (GiB)": 60.46,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.35642,
      "step": 2350,
      "tokens/total": 3695378432,
      "tokens/train_per_sec_per_gpu": 1140.81,
      "tokens/trainable": 3691470080
    },
    {
      "epoch": 0.6101242492611777,
      "grad_norm": 0.87890625,
      "learning_rate": 0.00018041020446429547,
      "loss": 5.084944152832032,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.57092,
      "step": 2400,
      "tokens/total": 3774013440,
      "tokens/train_per_sec_per_gpu": 1147.78,
      "tokens/trainable": 3770012672
    },
    {
      "epoch": 0.6228351711207856,
      "grad_norm": 0.984375,
      "learning_rate": 0.0001796109748496398,
      "loss": 5.079226684570313,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.64977,
      "step": 2450,
      "tokens/total": 3852640256,
      "tokens/train_per_sec_per_gpu": 1148.28,
      "tokens/trainable": 3848548096
    },
    {
      "epoch": 0.6355460929803934,
      "grad_norm": 1.1171875,
      "learning_rate": 0.00017879761164997553,
      "loss": 5.082112426757813,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.11404,
      "step": 2500,
      "tokens/total": 3931279360,
      "tokens/train_per_sec_per_gpu": 1145.45,
      "tokens/trainable": 3927099136
    },
    {
      "epoch": 0.6482570148400013,
      "grad_norm": 1.2109375,
      "learning_rate": 0.00017797025926421172,
      "loss": 5.081408081054687,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 161.0006,
      "step": 2550,
      "tokens/total": 4009910272,
      "tokens/train_per_sec_per_gpu": 995.23,
      "tokens/trainable": 4005641472
    },
    {
      "epoch": 0.6609679366996092,
      "grad_norm": 0.80859375,
      "learning_rate": 0.0001771290645748015,
      "loss": 5.0802920532226565,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.82102,
      "step": 2600,
      "tokens/total": 4088532992,
      "tokens/train_per_sec_per_gpu": 1141.69,
      "tokens/trainable": 4084177152
    },
    {
      "epoch": 0.673678858559217,
      "grad_norm": 1.03125,
      "learning_rate": 0.00017627417692166528,
      "loss": 5.079031066894531,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.61835,
      "step": 2650,
      "tokens/total": 4167155712,
      "tokens/train_per_sec_per_gpu": 1138.96,
      "tokens/trainable": 4162711296
    },
    {
      "epoch": 0.6863897804188249,
      "grad_norm": 1.265625,
      "learning_rate": 0.00017540574807567818,
      "loss": 5.079642639160157,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.71661,
      "step": 2700,
      "tokens/total": 4245798912,
      "tokens/train_per_sec_per_gpu": 1144.53,
      "tokens/trainable": 4241262592
    },
    {
      "epoch": 0.6991007022784328,
      "grad_norm": 0.69140625,
      "learning_rate": 0.0001745239322117255,
      "loss": 5.078349304199219,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.50889,
      "step": 2750,
      "tokens/total": 4324409344,
      "tokens/train_per_sec_per_gpu": 892.26,
      "tokens/trainable": 4319796224
    },
    {
      "epoch": 0.7118116241380407,
      "grad_norm": 0.90234375,
      "learning_rate": 0.0001736288858813317,
      "loss": 5.078107299804688,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.47005,
      "step": 2800,
      "tokens/total": 4403052544,
      "tokens/train_per_sec_per_gpu": 984.67,
      "tokens/trainable": 4398406144
    },
    {
      "epoch": 0.7245225459976484,
      "grad_norm": 0.8359375,
      "learning_rate": 0.00017272076798486726,
      "loss": 5.076226806640625,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.16857,
      "step": 2850,
      "tokens/total": 4481658880,
      "tokens/train_per_sec_per_gpu": 1142.23,
      "tokens/trainable": 4476985344
    },
    {
      "epoch": 0.7372334678572563,
      "grad_norm": 0.70703125,
      "learning_rate": 0.00017179973974333857,
      "loss": 5.076329650878907,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.18504,
      "step": 2900,
      "tokens/total": 4560277504,
      "tokens/train_per_sec_per_gpu": 1143.25,
      "tokens/trainable": 4555567616
    },
    {
      "epoch": 0.7499443897168642,
      "grad_norm": 0.890625,
      "learning_rate": 0.00017086596466976594,
      "loss": 5.074002990722656,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.81278,
      "step": 2950,
      "tokens/total": 4638900224,
      "tokens/train_per_sec_per_gpu": 1007.19,
      "tokens/trainable": 4634149888
    },
    {
      "epoch": 0.762655311576472,
      "grad_norm": 0.859375,
      "learning_rate": 0.00016991960854015464,
      "loss": 5.075715942382812,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 160.08676,
      "step": 3000,
      "tokens/total": 4717506560,
      "tokens/train_per_sec_per_gpu": 1142.63,
      "tokens/trainable": 4712714752
    },
    {
      "epoch": 0.7753662334360799,
      "grad_norm": 0.984375,
      "learning_rate": 0.00016896083936406394,
      "loss": 5.0732049560546875,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.68529,
      "step": 3050,
      "tokens/total": 4796145664,
      "tokens/train_per_sec_per_gpu": 1140.38,
      "tokens/trainable": 4791320576
    },
    {
      "epoch": 0.7880771552956878,
      "grad_norm": 0.83203125,
      "learning_rate": 0.00016798982735478023,
      "loss": 5.072262573242187,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.53488,
      "step": 3100,
      "tokens/total": 4874768384,
      "tokens/train_per_sec_per_gpu": 1010.94,
      "tokens/trainable": 4869914112
    },
    {
      "epoch": 0.8007880771552957,
      "grad_norm": 1.1015625,
      "learning_rate": 0.00016700674489909817,
      "loss": 5.071590881347657,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.42776,
      "step": 3150,
      "tokens/total": 4953391104,
      "tokens/train_per_sec_per_gpu": 1134.74,
      "tokens/trainable": 4948505600
    },
    {
      "epoch": 0.8134989990149035,
      "grad_norm": 1.3828125,
      "learning_rate": 0.00016601176652671654,
      "loss": 5.071939392089844,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.48333,
      "step": 3200,
      "tokens/total": 5032034304,
      "tokens/train_per_sec_per_gpu": 998.75,
      "tokens/trainable": 5027120640
    },
    {
      "epoch": 0.8262099208745114,
      "grad_norm": 0.77734375,
      "learning_rate": 0.00016500506887925332,
      "loss": 5.070206298828125,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.20717,
      "step": 3250,
      "tokens/total": 5110657024,
      "tokens/train_per_sec_per_gpu": 987.57,
      "tokens/trainable": 5105710592
    },
    {
      "epoch": 0.8389208427341193,
      "grad_norm": 0.6484375,
      "learning_rate": 0.000163986830678886,
      "loss": 5.072908325195312,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.63793,
      "step": 3300,
      "tokens/total": 5189275648,
      "tokens/train_per_sec_per_gpu": 1137.51,
      "tokens/trainable": 5184292864
    },
    {
      "epoch": 0.8516317645937271,
      "grad_norm": 0.94140625,
      "learning_rate": 0.00016295723269662247,
      "loss": 5.069717712402344,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.1294,
      "step": 3350,
      "tokens/total": 5267918848,
      "tokens/train_per_sec_per_gpu": 1141.74,
      "tokens/trainable": 5262910976
    },
    {
      "epoch": 0.864342686453335,
      "grad_norm": 0.98828125,
      "learning_rate": 0.00016191645772020827,
      "loss": 5.071309204101563,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.38286,
      "step": 3400,
      "tokens/total": 5346545664,
      "tokens/train_per_sec_per_gpu": 998.73,
      "tokens/trainable": 5341506560
    },
    {
      "epoch": 0.8770536083129429,
      "grad_norm": 1.21875,
      "learning_rate": 0.00016086469052167547,
      "loss": 5.069463500976562,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.08895,
      "step": 3450,
      "tokens/total": 5425172480,
      "tokens/train_per_sec_per_gpu": 990.05,
      "tokens/trainable": 5420105216
    },
    {
      "epoch": 0.8897645301725507,
      "grad_norm": 0.67578125,
      "learning_rate": 0.00015980211782453977,
      "loss": 5.070612182617188,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.2718,
      "step": 3500,
      "tokens/total": 5503782912,
      "tokens/train_per_sec_per_gpu": 1139.16,
      "tokens/trainable": 5498675200
    },
    {
      "epoch": 0.9024754520321586,
      "grad_norm": 0.8984375,
      "learning_rate": 0.0001587289282706508,
      "loss": 5.069385375976562,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.07653,
      "step": 3550,
      "tokens/total": 5582409728,
      "tokens/train_per_sec_per_gpu": 991.04,
      "tokens/trainable": 5577267712
    },
    {
      "epoch": 0.9151863738917665,
      "grad_norm": 1.0078125,
      "learning_rate": 0.00015764531238670175,
      "loss": 5.068446044921875,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 158.92717,
      "step": 3600,
      "tokens/total": 5661048832,
      "tokens/train_per_sec_per_gpu": 1146.26,
      "tokens/trainable": 5655876608
    },
    {
      "epoch": 0.9278972957513744,
      "grad_norm": 1.09375,
      "learning_rate": 0.00015655146255040501,
      "loss": 5.070053405761719,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.18283,
      "step": 3650,
      "tokens/total": 5739687936,
      "tokens/train_per_sec_per_gpu": 1143.44,
      "tokens/trainable": 5734480384
    },
    {
      "epoch": 0.9406082176109822,
      "grad_norm": 0.71484375,
      "learning_rate": 0.00015544757295633837,
      "loss": 5.067489013671875,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 158.77514,
      "step": 3700,
      "tokens/total": 5818314752,
      "tokens/train_per_sec_per_gpu": 1016.18,
      "tokens/trainable": 5813072896
    },
    {
      "epoch": 0.9533191394705901,
      "grad_norm": 0.73046875,
      "learning_rate": 0.00015433383958146906,
      "loss": 5.07120849609375,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 159.36681,
      "step": 3750,
      "tokens/total": 5896941568,
      "tokens/train_per_sec_per_gpu": 1158.51,
      "tokens/trainable": 5891662848
    },
    {
      "epoch": 0.966030061330198,
      "grad_norm": 0.8359375,
      "learning_rate": 0.00015321046015036167,
      "loss": 5.0687188720703125,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 158.97054,
      "step": 3800,
      "tokens/total": 5975556096,
      "tokens/train_per_sec_per_gpu": 1155.48,
      "tokens/trainable": 5970243072
    },
    {
      "epoch": 0.9787409831898058,
      "grad_norm": 1.2421875,
      "learning_rate": 0.0001520776341000751,
      "loss": 5.067420959472656,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 158.76434,
      "step": 3850,
      "tokens/total": 6054170624,
      "tokens/train_per_sec_per_gpu": 1157.96,
      "tokens/trainable": 6048813568
    },
    {
      "epoch": 0.9914519050494137,
      "grad_norm": 0.78125,
      "learning_rate": 0.00015093556254475618,
      "loss": 5.067011108398438,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "ppl": 158.69928,
      "step": 3900,
      "tokens/total": 6132797440,
      "tokens/train_per_sec_per_gpu": 1153.86,
      "tokens/trainable": 6127400960
    },
    {
      "epoch": 0.9998411134767549,
      "eval_loss": 5.067268371582031,
      "eval_ppl": 158.74012,
      "eval_runtime": 255.6347,
      "eval_samples_per_second": 1142.595,
      "eval_steps_per_second": 23.807,
      "memory/device_reserved (GiB)": 60.62,
      "memory/max_active (GiB)": 52.53,
      "memory/max_allocated (GiB)": 52.53,
      "step": 3933
    }
  ],
  "logging_steps": 50,
  "max_steps": 11799,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.3283000173283246e+19,
  "train_batch_size": 6,
  "trial_name": null,
  "trial_params": null
}