{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.1328462304882099,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.001328462304882099,
      "grad_norm": 3.849614381790161,
      "learning_rate": 0.0,
      "loss": 1.2204899787902832,
      "memory/device_reserved (GiB)": 74.02,
      "memory/max_active (GiB)": 73.31,
      "memory/max_allocated (GiB)": 73.31,
      "ppl": 3.38885,
      "step": 1,
      "tokens/total": 241280,
      "tokens/trainable": 1844
    },
    {
      "epoch": 0.002656924609764198,
      "grad_norm": 3.235618829727173,
      "learning_rate": 2.2222222222222225e-06,
      "loss": 1.208996057510376,
      "memory/device_reserved (GiB)": 74.03,
      "memory/max_active (GiB)": 71.94,
      "memory/max_allocated (GiB)": 71.94,
      "ppl": 3.35012,
      "step": 2,
      "tokens/total": 481728,
      "tokens/train_per_sec_per_gpu": 12.87,
      "tokens/trainable": 3740
    },
    {
      "epoch": 0.003985386914646297,
      "grad_norm": 3.656043767929077,
      "learning_rate": 4.444444444444445e-06,
      "loss": 1.4389150142669678,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 75.91,
      "memory/max_allocated (GiB)": 75.91,
      "ppl": 4.21612,
      "step": 3,
      "tokens/total": 699776,
      "tokens/train_per_sec_per_gpu": 13.14,
      "tokens/trainable": 5674
    },
    {
      "epoch": 0.005313849219528396,
      "grad_norm": 2.997349739074707,
      "learning_rate": 6.666666666666667e-06,
      "loss": 1.1131364107131958,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 75.18,
      "memory/max_allocated (GiB)": 75.18,
      "ppl": 3.04389,
      "step": 4,
      "tokens/total": 961536,
      "tokens/train_per_sec_per_gpu": 10.46,
      "tokens/trainable": 7547
    },
    {
      "epoch": 0.006642311524410495,
      "grad_norm": 3.100785493850708,
      "learning_rate": 8.88888888888889e-06,
      "loss": 1.2307677268981934,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 75.5,
      "memory/max_allocated (GiB)": 75.5,
      "ppl": 3.42386,
      "step": 5,
      "tokens/total": 1220800,
      "tokens/train_per_sec_per_gpu": 11.03,
      "tokens/trainable": 9461
    },
    {
      "epoch": 0.007970773829292593,
      "grad_norm": 2.7882754802703857,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 1.1808981895446777,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 75.28,
      "memory/max_allocated (GiB)": 75.28,
      "ppl": 3.2573,
      "step": 6,
      "tokens/total": 1486464,
      "tokens/train_per_sec_per_gpu": 14.7,
      "tokens/trainable": 12044
    },
    {
      "epoch": 0.009299236134174693,
      "grad_norm": 3.3219873905181885,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 1.3643646240234375,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 74.3,
      "memory/max_allocated (GiB)": 74.3,
      "ppl": 3.91324,
      "step": 7,
      "tokens/total": 1721408,
      "tokens/train_per_sec_per_gpu": 12.8,
      "tokens/trainable": 14012
    },
    {
      "epoch": 0.010627698439056792,
      "grad_norm": 3.5518031120300293,
      "learning_rate": 1.5555555555555555e-05,
      "loss": 1.4114501476287842,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 74.33,
      "memory/max_allocated (GiB)": 74.33,
      "ppl": 4.1019,
      "step": 8,
      "tokens/total": 1929152,
      "tokens/train_per_sec_per_gpu": 14.39,
      "tokens/trainable": 15860
    },
    {
      "epoch": 0.01195616074393889,
      "grad_norm": 3.1157267093658447,
      "learning_rate": 1.777777777777778e-05,
      "loss": 1.2192461490631104,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 72.51,
      "memory/max_allocated (GiB)": 72.51,
      "ppl": 3.38464,
      "step": 9,
      "tokens/total": 2134464,
      "tokens/train_per_sec_per_gpu": 13.52,
      "tokens/trainable": 17682
    },
    {
      "epoch": 0.01328462304882099,
      "grad_norm": 3.2054903507232666,
      "learning_rate": 2e-05,
      "loss": 1.0837433338165283,
      "memory/device_reserved (GiB)": 76.73,
      "memory/max_active (GiB)": 72.07,
      "memory/max_allocated (GiB)": 72.07,
      "ppl": 2.95572,
      "step": 10,
      "tokens/total": 2366720,
      "tokens/train_per_sec_per_gpu": 13.52,
      "tokens/trainable": 19525
    },
    {
      "epoch": 0.014613085353703089,
      "grad_norm": 2.874584913253784,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 1.0766642093658447,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 76.02,
      "memory/max_allocated (GiB)": 76.02,
      "ppl": 2.93487,
      "step": 11,
      "tokens/total": 2609600,
      "tokens/train_per_sec_per_gpu": 12.69,
      "tokens/trainable": 21534
    },
    {
      "epoch": 0.015941547658585187,
      "grad_norm": 2.1507201194763184,
      "learning_rate": 2.4444444444444445e-05,
      "loss": 0.8726168274879456,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.23,
      "memory/max_allocated (GiB)": 75.23,
      "ppl": 2.39317,
      "step": 12,
      "tokens/total": 2881216,
      "tokens/train_per_sec_per_gpu": 11.29,
      "tokens/trainable": 23688
    },
    {
      "epoch": 0.017270009963467288,
      "grad_norm": 2.2943410873413086,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 0.8717596530914307,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.53,
      "memory/max_allocated (GiB)": 75.53,
      "ppl": 2.39111,
      "step": 13,
      "tokens/total": 3154112,
      "tokens/train_per_sec_per_gpu": 10.45,
      "tokens/trainable": 25514
    },
    {
      "epoch": 0.018598472268349386,
      "grad_norm": 1.9567185640335083,
      "learning_rate": 2.8888888888888888e-05,
      "loss": 0.7917066812515259,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 74.19,
      "memory/max_allocated (GiB)": 74.19,
      "ppl": 2.20716,
      "step": 14,
      "tokens/total": 3410304,
      "tokens/train_per_sec_per_gpu": 12.17,
      "tokens/trainable": 27547
    },
    {
      "epoch": 0.019926934573231483,
      "grad_norm": 1.7974358797073364,
      "learning_rate": 3.111111111111111e-05,
      "loss": 0.8320116400718689,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.77,
      "memory/max_allocated (GiB)": 75.77,
      "ppl": 2.29794,
      "step": 15,
      "tokens/total": 3654464,
      "tokens/train_per_sec_per_gpu": 13.46,
      "tokens/trainable": 29471
    },
    {
      "epoch": 0.021255396878113585,
      "grad_norm": 1.3337464332580566,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.6287137866020203,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.2,
      "memory/max_allocated (GiB)": 75.2,
      "ppl": 1.8752,
      "step": 16,
      "tokens/total": 3909248,
      "tokens/train_per_sec_per_gpu": 12.03,
      "tokens/trainable": 31464
    },
    {
      "epoch": 0.022583859182995682,
      "grad_norm": 1.8309210538864136,
      "learning_rate": 3.555555555555556e-05,
      "loss": 0.7073560953140259,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 74.85,
      "memory/max_allocated (GiB)": 74.85,
      "ppl": 2.02862,
      "step": 17,
      "tokens/total": 4131008,
      "tokens/train_per_sec_per_gpu": 11.95,
      "tokens/trainable": 33074
    },
    {
      "epoch": 0.02391232148787778,
      "grad_norm": 0.980492115020752,
      "learning_rate": 3.777777777777778e-05,
      "loss": 0.589785635471344,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 72.75,
      "memory/max_allocated (GiB)": 72.75,
      "ppl": 1.8036,
      "step": 18,
      "tokens/total": 4342272,
      "tokens/train_per_sec_per_gpu": 15.5,
      "tokens/trainable": 35185
    },
    {
      "epoch": 0.02524078379275988,
      "grad_norm": 0.8374645709991455,
      "learning_rate": 4e-05,
      "loss": 0.49727869033813477,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.23,
      "memory/max_allocated (GiB)": 75.23,
      "ppl": 1.64424,
      "step": 19,
      "tokens/total": 4594880,
      "tokens/train_per_sec_per_gpu": 12.79,
      "tokens/trainable": 37280
    },
    {
      "epoch": 0.02656924609764198,
      "grad_norm": 0.586373507976532,
      "learning_rate": 4.222222222222222e-05,
      "loss": 0.37330132722854614,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.34,
      "memory/max_allocated (GiB)": 75.34,
      "ppl": 1.45252,
      "step": 20,
      "tokens/total": 4858752,
      "tokens/train_per_sec_per_gpu": 10.82,
      "tokens/trainable": 39261
    },
    {
      "epoch": 0.027897708402524077,
      "grad_norm": 7.131537914276123,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.4620007276535034,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 74.66,
      "memory/max_allocated (GiB)": 74.66,
      "ppl": 1.58725,
      "step": 21,
      "tokens/total": 5107648,
      "tokens/train_per_sec_per_gpu": 12.27,
      "tokens/trainable": 41174
    },
    {
      "epoch": 0.029226170707406178,
      "grad_norm": 0.5388134717941284,
      "learning_rate": 4.666666666666667e-05,
      "loss": 0.4094924330711365,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 72.94,
      "memory/max_allocated (GiB)": 72.94,
      "ppl": 1.50605,
      "step": 22,
      "tokens/total": 5310976,
      "tokens/train_per_sec_per_gpu": 15.93,
      "tokens/trainable": 43051
    },
    {
      "epoch": 0.030554633012288276,
      "grad_norm": 0.46234583854675293,
      "learning_rate": 4.888888888888889e-05,
      "loss": 0.4509355425834656,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 76.1,
      "memory/max_allocated (GiB)": 76.1,
      "ppl": 1.56978,
      "step": 23,
      "tokens/total": 5557504,
      "tokens/train_per_sec_per_gpu": 14.78,
      "tokens/trainable": 45232
    },
    {
      "epoch": 0.031883095317170373,
      "grad_norm": 0.4812721610069275,
      "learning_rate": 5.111111111111111e-05,
      "loss": 0.32048141956329346,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.94,
      "memory/max_allocated (GiB)": 75.94,
      "ppl": 1.37779,
      "step": 24,
      "tokens/total": 5795904,
      "tokens/train_per_sec_per_gpu": 10.11,
      "tokens/trainable": 46943
    },
    {
      "epoch": 0.033211557622052475,
      "grad_norm": 0.3919018805027008,
      "learning_rate": 5.333333333333333e-05,
      "loss": 0.4536985158920288,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 74.17,
      "memory/max_allocated (GiB)": 74.17,
      "ppl": 1.57412,
      "step": 25,
      "tokens/total": 6040896,
      "tokens/train_per_sec_per_gpu": 13.85,
      "tokens/trainable": 48930
    },
    {
      "epoch": 0.034540019926934576,
      "grad_norm": 0.37939736247062683,
      "learning_rate": 5.555555555555556e-05,
      "loss": 0.32177501916885376,
      "memory/device_reserved (GiB)": 77.14,
      "memory/max_active (GiB)": 75.69,
      "memory/max_allocated (GiB)": 75.69,
      "ppl": 1.37957,
      "step": 26,
      "tokens/total": 6295680,
      "tokens/train_per_sec_per_gpu": 8.93,
      "tokens/trainable": 51085
    },
    {
      "epoch": 0.03586848223181667,
      "grad_norm": 2.0314126014709473,
      "learning_rate": 5.7777777777777776e-05,
      "loss": 0.46417510509490967,
      "memory/device_reserved (GiB)": 77.47,
      "memory/max_active (GiB)": 76.65,
      "memory/max_allocated (GiB)": 76.65,
      "ppl": 1.5907,
      "step": 27,
      "tokens/total": 6534144,
      "tokens/train_per_sec_per_gpu": 13.58,
      "tokens/trainable": 53256
    },
    {
      "epoch": 0.03719694453669877,
      "grad_norm": 0.43113988637924194,
      "learning_rate": 6e-05,
      "loss": 0.4589465260505676,
      "memory/device_reserved (GiB)": 77.47,
      "memory/max_active (GiB)": 73.51,
      "memory/max_allocated (GiB)": 73.51,
      "ppl": 1.58241,
      "step": 28,
      "tokens/total": 6745152,
      "tokens/train_per_sec_per_gpu": 15.98,
      "tokens/trainable": 55353
    },
    {
      "epoch": 0.03852540684158087,
      "grad_norm": 0.3234078288078308,
      "learning_rate": 6.222222222222222e-05,
      "loss": 0.3227987289428711,
      "memory/device_reserved (GiB)": 77.47,
      "memory/max_active (GiB)": 75.75,
      "memory/max_allocated (GiB)": 75.75,
      "ppl": 1.38099,
      "step": 29,
      "tokens/total": 6994688,
      "tokens/train_per_sec_per_gpu": 12.33,
      "tokens/trainable": 57312
    },
    {
      "epoch": 0.03985386914646297,
      "grad_norm": 0.3677234649658203,
      "learning_rate": 6.444444444444446e-05,
      "loss": 0.43566569685935974,
      "memory/device_reserved (GiB)": 77.47,
      "memory/max_active (GiB)": 76.4,
      "memory/max_allocated (GiB)": 76.4,
      "ppl": 1.54599,
      "step": 30,
      "tokens/total": 7220288,
      "tokens/train_per_sec_per_gpu": 13.43,
      "tokens/trainable": 59213
    },
    {
      "epoch": 0.04118233145134507,
      "grad_norm": 0.29641425609588623,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.40512919425964355,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 76.43,
      "memory/max_allocated (GiB)": 76.43,
      "ppl": 1.4995,
      "step": 31,
      "tokens/total": 7466432,
      "tokens/train_per_sec_per_gpu": 13.24,
      "tokens/trainable": 61213
    },
    {
      "epoch": 0.04251079375622717,
      "grad_norm": 0.3371325433254242,
      "learning_rate": 6.88888888888889e-05,
      "loss": 0.4463217258453369,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 72.26,
      "memory/max_allocated (GiB)": 72.26,
      "ppl": 1.56255,
      "step": 32,
      "tokens/total": 7682816,
      "tokens/train_per_sec_per_gpu": 14.94,
      "tokens/trainable": 63202
    },
    {
      "epoch": 0.043839256061109264,
      "grad_norm": 0.3139035999774933,
      "learning_rate": 7.111111111111112e-05,
      "loss": 0.3831879496574402,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 71.72,
      "memory/max_allocated (GiB)": 71.72,
      "ppl": 1.46695,
      "step": 33,
      "tokens/total": 7905344,
      "tokens/train_per_sec_per_gpu": 12.46,
      "tokens/trainable": 65086
    },
    {
      "epoch": 0.045167718365991365,
      "grad_norm": 0.2688130736351013,
      "learning_rate": 7.333333333333333e-05,
      "loss": 0.2790850102901459,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 76.48,
      "memory/max_allocated (GiB)": 76.48,
      "ppl": 1.32192,
      "step": 34,
      "tokens/total": 8152960,
      "tokens/train_per_sec_per_gpu": 10.56,
      "tokens/trainable": 66938
    },
    {
      "epoch": 0.046496180670873466,
      "grad_norm": 0.2653343677520752,
      "learning_rate": 7.555555555555556e-05,
      "loss": 0.3251267075538635,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 72.7,
      "memory/max_allocated (GiB)": 72.7,
      "ppl": 1.38421,
      "step": 35,
      "tokens/total": 8406912,
      "tokens/train_per_sec_per_gpu": 11.68,
      "tokens/trainable": 68947
    },
    {
      "epoch": 0.04782464297575556,
      "grad_norm": 0.2554786801338196,
      "learning_rate": 7.777777777777778e-05,
      "loss": 0.35603782534599304,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 74.66,
      "memory/max_allocated (GiB)": 74.66,
      "ppl": 1.42766,
      "step": 36,
      "tokens/total": 8657536,
      "tokens/train_per_sec_per_gpu": 12.21,
      "tokens/trainable": 70906
    },
    {
      "epoch": 0.04915310528063766,
      "grad_norm": 0.329858660697937,
      "learning_rate": 8e-05,
      "loss": 0.3854241371154785,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 73.35,
      "memory/max_allocated (GiB)": 73.35,
      "ppl": 1.47024,
      "step": 37,
      "tokens/total": 8891968,
      "tokens/train_per_sec_per_gpu": 14.54,
      "tokens/trainable": 72932
    },
    {
      "epoch": 0.05048156758551976,
      "grad_norm": 0.3251776397228241,
      "learning_rate": 8.222222222222222e-05,
      "loss": 0.35971540212631226,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 73.19,
      "memory/max_allocated (GiB)": 73.19,
      "ppl": 1.43292,
      "step": 38,
      "tokens/total": 9137728,
      "tokens/train_per_sec_per_gpu": 12.29,
      "tokens/trainable": 74902
    },
    {
      "epoch": 0.05181002989040186,
      "grad_norm": 0.27980777621269226,
      "learning_rate": 8.444444444444444e-05,
      "loss": 0.247895747423172,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 75.69,
      "memory/max_allocated (GiB)": 75.69,
      "ppl": 1.28133,
      "step": 39,
      "tokens/total": 9373760,
      "tokens/train_per_sec_per_gpu": 12.65,
      "tokens/trainable": 76811
    },
    {
      "epoch": 0.05313849219528396,
      "grad_norm": 0.26269567012786865,
      "learning_rate": 8.666666666666667e-05,
      "loss": 0.3823346495628357,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 73.81,
      "memory/max_allocated (GiB)": 73.81,
      "ppl": 1.4657,
      "step": 40,
      "tokens/total": 9608640,
      "tokens/train_per_sec_per_gpu": 15.26,
      "tokens/trainable": 79037
    },
    {
      "epoch": 0.05446695450016606,
      "grad_norm": 0.3128836750984192,
      "learning_rate": 8.888888888888889e-05,
      "loss": 0.3774491548538208,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 73.95,
      "memory/max_allocated (GiB)": 73.95,
      "ppl": 1.45856,
      "step": 41,
      "tokens/total": 9850688,
      "tokens/train_per_sec_per_gpu": 13.66,
      "tokens/trainable": 81194
    },
    {
      "epoch": 0.055795416805048154,
      "grad_norm": 0.2611078917980194,
      "learning_rate": 9.111111111111112e-05,
      "loss": 0.37145930528640747,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 75.53,
      "memory/max_allocated (GiB)": 75.53,
      "ppl": 1.44985,
      "step": 42,
      "tokens/total": 10095872,
      "tokens/train_per_sec_per_gpu": 13.83,
      "tokens/trainable": 83205
    },
    {
      "epoch": 0.057123879109930255,
      "grad_norm": 0.2708030343055725,
      "learning_rate": 9.333333333333334e-05,
      "loss": 0.3306182622909546,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 75.53,
      "memory/max_allocated (GiB)": 75.53,
      "ppl": 1.39183,
      "step": 43,
      "tokens/total": 10312640,
      "tokens/train_per_sec_per_gpu": 14.12,
      "tokens/trainable": 85157
    },
    {
      "epoch": 0.058452341414812356,
      "grad_norm": 0.3672192692756653,
      "learning_rate": 9.555555555555557e-05,
      "loss": 0.38120734691619873,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 71.88,
      "memory/max_allocated (GiB)": 71.88,
      "ppl": 1.46405,
      "step": 44,
      "tokens/total": 10522944,
      "tokens/train_per_sec_per_gpu": 14.59,
      "tokens/trainable": 87062
    },
    {
      "epoch": 0.05978080371969445,
      "grad_norm": 0.8446953892707825,
      "learning_rate": 9.777777777777778e-05,
      "loss": 0.281764417886734,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 73.76,
      "memory/max_allocated (GiB)": 73.76,
      "ppl": 1.32547,
      "step": 45,
      "tokens/total": 10763456,
      "tokens/train_per_sec_per_gpu": 11.81,
      "tokens/trainable": 88876
    },
    {
      "epoch": 0.06110926602457655,
      "grad_norm": 0.2893065810203552,
      "learning_rate": 0.0001,
      "loss": 0.30848920345306396,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 76.21,
      "memory/max_allocated (GiB)": 76.21,
      "ppl": 1.36137,
      "step": 46,
      "tokens/total": 11004864,
      "tokens/train_per_sec_per_gpu": 10.95,
      "tokens/trainable": 90781
    },
    {
      "epoch": 0.06243772832945865,
      "grad_norm": 0.2893041670322418,
      "learning_rate": 9.999988440497192e-05,
      "loss": 0.3450867533683777,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 76.16,
      "memory/max_allocated (GiB)": 76.16,
      "ppl": 1.41211,
      "step": 47,
      "tokens/total": 11245440,
      "tokens/train_per_sec_per_gpu": 12.74,
      "tokens/trainable": 92871
    },
    {
      "epoch": 0.06376619063434075,
      "grad_norm": 0.3253316879272461,
      "learning_rate": 9.999953762042213e-05,
      "loss": 0.4527391791343689,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 72.78,
      "memory/max_allocated (GiB)": 72.78,
      "ppl": 1.57261,
      "step": 48,
      "tokens/total": 11478272,
      "tokens/train_per_sec_per_gpu": 14.65,
      "tokens/trainable": 95022
    },
    {
      "epoch": 0.06509465293922286,
      "grad_norm": 0.4156225025653839,
      "learning_rate": 9.999895964795412e-05,
      "loss": 0.3814932703971863,
      "memory/device_reserved (GiB)": 77.49,
      "memory/max_active (GiB)": 73.32,
      "memory/max_allocated (GiB)": 73.32,
      "ppl": 1.46447,
      "step": 49,
      "tokens/total": 11701120,
      "tokens/train_per_sec_per_gpu": 15.7,
      "tokens/trainable": 97209
    },
    {
      "epoch": 0.06642311524410495,
      "grad_norm": 0.34428590536117554,
      "learning_rate": 9.99981504902403e-05,
      "loss": 0.4007723927497864,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.62,
      "memory/max_allocated (GiB)": 76.62,
      "ppl": 1.49298,
      "step": 50,
      "tokens/total": 11962496,
      "tokens/train_per_sec_per_gpu": 12.07,
      "tokens/trainable": 99480
    },
    {
      "epoch": 0.06775157754898704,
      "grad_norm": 0.28597790002822876,
      "learning_rate": 9.999711015102209e-05,
      "loss": 0.34098297357559204,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 71.69,
      "memory/max_allocated (GiB)": 71.69,
      "ppl": 1.40633,
      "step": 51,
      "tokens/total": 12175616,
      "tokens/train_per_sec_per_gpu": 11.89,
      "tokens/trainable": 101480
    },
    {
      "epoch": 0.06908003985386915,
      "grad_norm": 0.27698570489883423,
      "learning_rate": 9.999583863510977e-05,
      "loss": 0.333741694688797,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 76.83,
      "memory/max_allocated (GiB)": 76.83,
      "ppl": 1.39618,
      "step": 52,
      "tokens/total": 12425856,
      "tokens/train_per_sec_per_gpu": 10.53,
      "tokens/trainable": 103474
    },
    {
      "epoch": 0.07040850215875125,
      "grad_norm": 0.27645811438560486,
      "learning_rate": 9.999433594838259e-05,
      "loss": 0.2773073613643646,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.28,
      "memory/max_allocated (GiB)": 74.28,
      "ppl": 1.31957,
      "step": 53,
      "tokens/total": 12673472,
      "tokens/train_per_sec_per_gpu": 10.96,
      "tokens/trainable": 105399
    },
    {
      "epoch": 0.07173696446363334,
      "grad_norm": 0.23398520052433014,
      "learning_rate": 9.999260209778867e-05,
      "loss": 0.31913238763809204,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.58,
      "memory/max_allocated (GiB)": 75.58,
      "ppl": 1.37593,
      "step": 54,
      "tokens/total": 12919936,
      "tokens/train_per_sec_per_gpu": 12.06,
      "tokens/trainable": 107320
    },
    {
      "epoch": 0.07306542676851545,
      "grad_norm": 0.38037580251693726,
      "learning_rate": 9.999063709134502e-05,
      "loss": 0.28329750895500183,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 71.94,
      "memory/max_allocated (GiB)": 71.94,
      "ppl": 1.3275,
      "step": 55,
      "tokens/total": 13149568,
      "tokens/train_per_sec_per_gpu": 12.49,
      "tokens/trainable": 109265
    },
    {
      "epoch": 0.07439388907339754,
      "grad_norm": 0.266801118850708,
      "learning_rate": 9.998844093813741e-05,
      "loss": 0.3162427842617035,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.94,
      "memory/max_allocated (GiB)": 75.94,
      "ppl": 1.37196,
      "step": 56,
      "tokens/total": 13412032,
      "tokens/train_per_sec_per_gpu": 11.08,
      "tokens/trainable": 111171
    },
    {
      "epoch": 0.07572235137827964,
      "grad_norm": 0.3075020909309387,
      "learning_rate": 9.998601364832042e-05,
      "loss": 0.35588574409484863,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.44,
      "memory/max_allocated (GiB)": 74.44,
      "ppl": 1.42744,
      "step": 57,
      "tokens/total": 13670784,
      "tokens/train_per_sec_per_gpu": 13.97,
      "tokens/trainable": 113207
    },
    {
      "epoch": 0.07705081368316175,
      "grad_norm": 0.296512246131897,
      "learning_rate": 9.998335523311734e-05,
      "loss": 0.3532083034515381,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.6,
      "memory/max_allocated (GiB)": 74.6,
      "ppl": 1.42363,
      "step": 58,
      "tokens/total": 13901888,
      "tokens/train_per_sec_per_gpu": 13.23,
      "tokens/trainable": 115221
    },
    {
      "epoch": 0.07837927598804384,
      "grad_norm": 0.34292277693748474,
      "learning_rate": 9.998046570482021e-05,
      "loss": 0.2728872299194336,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 76.02,
      "memory/max_allocated (GiB)": 76.02,
      "ppl": 1.31375,
      "step": 59,
      "tokens/total": 14140224,
      "tokens/train_per_sec_per_gpu": 11.28,
      "tokens/trainable": 117086
    },
    {
      "epoch": 0.07970773829292593,
      "grad_norm": 0.30472293496131897,
      "learning_rate": 9.997734507678958e-05,
      "loss": 0.34186768531799316,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.6,
      "memory/max_allocated (GiB)": 74.6,
      "ppl": 1.40757,
      "step": 60,
      "tokens/total": 14384256,
      "tokens/train_per_sec_per_gpu": 14.36,
      "tokens/trainable": 119322
    },
    {
      "epoch": 0.08103620059780804,
      "grad_norm": 0.2648812532424927,
      "learning_rate": 9.99739933634546e-05,
      "loss": 0.3356971740722656,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.72,
      "memory/max_allocated (GiB)": 75.72,
      "ppl": 1.39892,
      "step": 61,
      "tokens/total": 14647168,
      "tokens/train_per_sec_per_gpu": 11.08,
      "tokens/trainable": 121273
    },
    {
      "epoch": 0.08236466290269014,
      "grad_norm": 0.2972364127635956,
      "learning_rate": 9.9970410580313e-05,
      "loss": 0.29585444927215576,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.26,
      "memory/max_allocated (GiB)": 75.26,
      "ppl": 1.34427,
      "step": 62,
      "tokens/total": 14905792,
      "tokens/train_per_sec_per_gpu": 13.09,
      "tokens/trainable": 123323
    },
    {
      "epoch": 0.08369312520757223,
      "grad_norm": 0.29903343319892883,
      "learning_rate": 9.996659674393078e-05,
      "loss": 0.3641955852508545,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.28,
      "memory/max_allocated (GiB)": 75.28,
      "ppl": 1.43936,
      "step": 63,
      "tokens/total": 15153664,
      "tokens/train_per_sec_per_gpu": 10.88,
      "tokens/trainable": 125382
    },
    {
      "epoch": 0.08502158751245434,
      "grad_norm": 0.33454394340515137,
      "learning_rate": 9.996255187194241e-05,
      "loss": 0.35934415459632874,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 72.29,
      "memory/max_allocated (GiB)": 72.29,
      "ppl": 1.43239,
      "step": 64,
      "tokens/total": 15367552,
      "tokens/train_per_sec_per_gpu": 15.22,
      "tokens/trainable": 127310
    },
    {
      "epoch": 0.08635004981733643,
      "grad_norm": 0.30956289172172546,
      "learning_rate": 9.995827598305055e-05,
      "loss": 0.4041798710823059,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 72.56,
      "memory/max_allocated (GiB)": 72.56,
      "ppl": 1.49807,
      "step": 65,
      "tokens/total": 15590848,
      "tokens/train_per_sec_per_gpu": 14.12,
      "tokens/trainable": 129279
    },
    {
      "epoch": 0.08767851212221853,
      "grad_norm": 0.26368483901023865,
      "learning_rate": 9.995376909702607e-05,
      "loss": 0.3347601890563965,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 73.44,
      "memory/max_allocated (GiB)": 73.44,
      "ppl": 1.39761,
      "step": 66,
      "tokens/total": 15858432,
      "tokens/train_per_sec_per_gpu": 11.04,
      "tokens/trainable": 131138
    },
    {
      "epoch": 0.08900697442710064,
      "grad_norm": 0.7567566633224487,
      "learning_rate": 9.99490312347079e-05,
      "loss": 0.3867908716201782,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.88,
      "memory/max_allocated (GiB)": 74.88,
      "ppl": 1.47225,
      "step": 67,
      "tokens/total": 16085504,
      "tokens/train_per_sec_per_gpu": 17.45,
      "tokens/trainable": 133359
    },
    {
      "epoch": 0.09033543673198273,
      "grad_norm": 0.27797308564186096,
      "learning_rate": 9.994406241800298e-05,
      "loss": 0.2849214971065521,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.74,
      "memory/max_allocated (GiB)": 74.74,
      "ppl": 1.32966,
      "step": 68,
      "tokens/total": 16326272,
      "tokens/train_per_sec_per_gpu": 12.31,
      "tokens/trainable": 135250
    },
    {
      "epoch": 0.09166389903686482,
      "grad_norm": 0.27896398305892944,
      "learning_rate": 9.993886266988615e-05,
      "loss": 0.2861061096191406,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.96,
      "memory/max_allocated (GiB)": 75.96,
      "ppl": 1.33123,
      "step": 69,
      "tokens/total": 16568384,
      "tokens/train_per_sec_per_gpu": 12.93,
      "tokens/trainable": 137383
    },
    {
      "epoch": 0.09299236134174693,
      "grad_norm": 0.2257593721151352,
      "learning_rate": 9.99334320144e-05,
      "loss": 0.2536161541938782,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 75.47,
      "memory/max_allocated (GiB)": 75.47,
      "ppl": 1.28868,
      "step": 70,
      "tokens/total": 16793536,
      "tokens/train_per_sec_per_gpu": 12.98,
      "tokens/trainable": 139172
    },
    {
      "epoch": 0.09432082364662903,
      "grad_norm": 0.30535292625427246,
      "learning_rate": 9.992777047665477e-05,
      "loss": 0.30237969756126404,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.82,
      "memory/max_allocated (GiB)": 74.82,
      "ppl": 1.35307,
      "step": 71,
      "tokens/total": 17018816,
      "tokens/train_per_sec_per_gpu": 13.41,
      "tokens/trainable": 141287
    },
    {
      "epoch": 0.09564928595151112,
      "grad_norm": 0.23031526803970337,
      "learning_rate": 9.992187808282834e-05,
      "loss": 0.30215615034103394,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.39,
      "memory/max_allocated (GiB)": 74.39,
      "ppl": 1.35277,
      "step": 72,
      "tokens/total": 17269248,
      "tokens/train_per_sec_per_gpu": 11.75,
      "tokens/trainable": 143236
    },
    {
      "epoch": 0.09697774825639323,
      "grad_norm": 0.3049887716770172,
      "learning_rate": 9.991575486016592e-05,
      "loss": 0.34131675958633423,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 76.24,
      "memory/max_allocated (GiB)": 76.24,
      "ppl": 1.4068,
      "step": 73,
      "tokens/total": 17508992,
      "tokens/train_per_sec_per_gpu": 14.61,
      "tokens/trainable": 145571
    },
    {
      "epoch": 0.09830621056127532,
      "grad_norm": 0.2670894265174866,
      "learning_rate": 9.99094008369801e-05,
      "loss": 0.32349398732185364,
      "memory/device_reserved (GiB)": 77.55,
      "memory/max_active (GiB)": 74.9,
      "memory/max_allocated (GiB)": 74.9,
      "ppl": 1.38195,
      "step": 74,
      "tokens/total": 17736448,
      "tokens/train_per_sec_per_gpu": 13.21,
      "tokens/trainable": 147614
    },
    {
      "epoch": 0.09963467286615742,
      "grad_norm": 0.1849091500043869,
      "learning_rate": 9.99028160426506e-05,
      "loss": 0.21922682225704193,
      "memory/device_reserved (GiB)": 77.79,
      "memory/max_active (GiB)": 76.86,
      "memory/max_allocated (GiB)": 76.86,
      "ppl": 1.24511,
      "step": 75,
      "tokens/total": 18018752,
      "tokens/train_per_sec_per_gpu": 11.37,
      "tokens/trainable": 149818
    },
    {
      "epoch": 0.10096313517103953,
      "grad_norm": 0.29884207248687744,
      "learning_rate": 9.989600050762422e-05,
      "loss": 0.35047096014022827,
      "memory/device_reserved (GiB)": 77.79,
      "memory/max_active (GiB)": 75.58,
      "memory/max_allocated (GiB)": 75.58,
      "ppl": 1.41974,
      "step": 76,
      "tokens/total": 18233856,
      "tokens/train_per_sec_per_gpu": 9.97,
      "tokens/trainable": 151707
    },
    {
      "epoch": 0.10229159747592162,
      "grad_norm": 0.3146807849407196,
      "learning_rate": 9.988895426341463e-05,
      "loss": 0.28694236278533936,
      "memory/device_reserved (GiB)": 77.25,
      "memory/max_active (GiB)": 76.32,
      "memory/max_allocated (GiB)": 76.32,
      "ppl": 1.33235,
      "step": 77,
      "tokens/total": 18452672,
      "tokens/train_per_sec_per_gpu": 11.96,
      "tokens/trainable": 153598
    },
    {
      "epoch": 0.10362005978080371,
      "grad_norm": 0.5249087810516357,
      "learning_rate": 9.988167734260226e-05,
      "loss": 0.282268762588501,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 75.34,
      "memory/max_allocated (GiB)": 75.34,
      "ppl": 1.32614,
      "step": 78,
      "tokens/total": 18695552,
      "tokens/train_per_sec_per_gpu": 12.26,
      "tokens/trainable": 155561
    },
    {
      "epoch": 0.10494852208568582,
      "grad_norm": 0.3543905019760132,
      "learning_rate": 9.987416977883415e-05,
      "loss": 0.3822178244590759,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 74.52,
      "memory/max_allocated (GiB)": 74.52,
      "ppl": 1.46553,
      "step": 79,
      "tokens/total": 18907712,
      "tokens/train_per_sec_per_gpu": 15.2,
      "tokens/trainable": 157505
    },
    {
      "epoch": 0.10627698439056792,
      "grad_norm": 0.26824578642845154,
      "learning_rate": 9.986643160682376e-05,
      "loss": 0.30315613746643066,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 76.4,
      "memory/max_allocated (GiB)": 76.4,
      "ppl": 1.35413,
      "step": 80,
      "tokens/total": 19134144,
      "tokens/train_per_sec_per_gpu": 12.57,
      "tokens/trainable": 159379
    },
    {
      "epoch": 0.10760544669545001,
      "grad_norm": 0.2979929447174072,
      "learning_rate": 9.98584628623509e-05,
      "loss": 0.3564770519733429,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 72.18,
      "memory/max_allocated (GiB)": 72.18,
      "ppl": 1.42829,
      "step": 81,
      "tokens/total": 19376000,
      "tokens/train_per_sec_per_gpu": 12.8,
      "tokens/trainable": 161333
    },
    {
      "epoch": 0.10893390900033212,
      "grad_norm": 0.25432249903678894,
      "learning_rate": 9.985026358226142e-05,
      "loss": 0.3223087191581726,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 74.09,
      "memory/max_allocated (GiB)": 74.09,
      "ppl": 1.38031,
      "step": 82,
      "tokens/total": 19613056,
      "tokens/train_per_sec_per_gpu": 16.53,
      "tokens/trainable": 163537
    },
    {
      "epoch": 0.11026237130521421,
      "grad_norm": 0.2386687844991684,
      "learning_rate": 9.984183380446718e-05,
      "loss": 0.2596937119960785,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 75.37,
      "memory/max_allocated (GiB)": 75.37,
      "ppl": 1.29653,
      "step": 83,
      "tokens/total": 19840960,
      "tokens/train_per_sec_per_gpu": 10.24,
      "tokens/trainable": 165415
    },
    {
      "epoch": 0.11159083361009631,
      "grad_norm": 0.2683466672897339,
      "learning_rate": 9.983317356794579e-05,
      "loss": 0.2844763994216919,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 74.69,
      "memory/max_allocated (GiB)": 74.69,
      "ppl": 1.32907,
      "step": 84,
      "tokens/total": 20084608,
      "tokens/train_per_sec_per_gpu": 10.79,
      "tokens/trainable": 167156
    },
    {
      "epoch": 0.11291929591497842,
      "grad_norm": 0.22222431004047394,
      "learning_rate": 9.982428291274046e-05,
      "loss": 0.2596544921398163,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 73.24,
      "memory/max_allocated (GiB)": 73.24,
      "ppl": 1.29648,
      "step": 85,
      "tokens/total": 20300224,
      "tokens/train_per_sec_per_gpu": 13.84,
      "tokens/trainable": 169157
    },
    {
      "epoch": 0.11424775821986051,
      "grad_norm": 0.24243441224098206,
      "learning_rate": 9.981516187995983e-05,
      "loss": 0.3357563614845276,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 73.6,
      "memory/max_allocated (GiB)": 73.6,
      "ppl": 1.399,
      "step": 86,
      "tokens/total": 20541056,
      "tokens/train_per_sec_per_gpu": 13.71,
      "tokens/trainable": 171164
    },
    {
      "epoch": 0.1155762205247426,
      "grad_norm": 0.27081701159477234,
      "learning_rate": 9.980581051177771e-05,
      "loss": 0.3430708050727844,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 76.13,
      "memory/max_allocated (GiB)": 76.13,
      "ppl": 1.40927,
      "step": 87,
      "tokens/total": 20785216,
      "tokens/train_per_sec_per_gpu": 13.84,
      "tokens/trainable": 173239
    },
    {
      "epoch": 0.11690468282962471,
      "grad_norm": 0.23188258707523346,
      "learning_rate": 9.979622885143301e-05,
      "loss": 0.27370986342430115,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 72.67,
      "memory/max_allocated (GiB)": 72.67,
      "ppl": 1.31483,
      "step": 88,
      "tokens/total": 21016768,
      "tokens/train_per_sec_per_gpu": 12.38,
      "tokens/trainable": 175096
    },
    {
      "epoch": 0.1182331451345068,
      "grad_norm": 0.26299989223480225,
      "learning_rate": 9.978641694322937e-05,
      "loss": 0.30719634890556335,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 74.71,
      "memory/max_allocated (GiB)": 74.71,
      "ppl": 1.35961,
      "step": 89,
      "tokens/total": 21246976,
      "tokens/train_per_sec_per_gpu": 13.23,
      "tokens/trainable": 177111
    },
    {
      "epoch": 0.1195616074393889,
      "grad_norm": 0.23563504219055176,
      "learning_rate": 9.977637483253513e-05,
      "loss": 0.3038645386695862,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 74.09,
      "memory/max_allocated (GiB)": 74.09,
      "ppl": 1.35509,
      "step": 90,
      "tokens/total": 21491072,
      "tokens/train_per_sec_per_gpu": 12.0,
      "tokens/trainable": 179077
    },
    {
      "epoch": 0.12089006974427101,
      "grad_norm": 0.24628512561321259,
      "learning_rate": 9.976610256578301e-05,
      "loss": 0.3164127469062805,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 75.86,
      "memory/max_allocated (GiB)": 75.86,
      "ppl": 1.3722,
      "step": 91,
      "tokens/total": 21734208,
      "tokens/train_per_sec_per_gpu": 13.08,
      "tokens/trainable": 181123
    },
    {
      "epoch": 0.1222185320491531,
      "grad_norm": 1.0121843814849854,
      "learning_rate": 9.975560019046994e-05,
      "loss": 0.30315762758255005,
      "memory/device_reserved (GiB)": 77.26,
      "memory/max_active (GiB)": 75.53,
      "memory/max_allocated (GiB)": 75.53,
      "ppl": 1.35413,
      "step": 92,
      "tokens/total": 21955776,
      "tokens/train_per_sec_per_gpu": 15.95,
      "tokens/trainable": 183236
    },
    {
      "epoch": 0.12354699435403521,
      "grad_norm": 0.2596462666988373,
      "learning_rate": 9.974486775515679e-05,
      "loss": 0.38250499963760376,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.84,
      "memory/max_allocated (GiB)": 76.84,
      "ppl": 1.46595,
      "step": 93,
      "tokens/total": 22220800,
      "tokens/train_per_sec_per_gpu": 13.69,
      "tokens/trainable": 185470
    },
    {
      "epoch": 0.1248754566589173,
      "grad_norm": 0.2512047588825226,
      "learning_rate": 9.973390530946824e-05,
      "loss": 0.2767234444618225,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.16,
      "memory/max_allocated (GiB)": 76.16,
      "ppl": 1.3188,
      "step": 94,
      "tokens/total": 22472000,
      "tokens/train_per_sec_per_gpu": 12.01,
      "tokens/trainable": 187632
    },
    {
      "epoch": 0.1262039189637994,
      "grad_norm": 0.2514182925224304,
      "learning_rate": 9.972271290409241e-05,
      "loss": 0.3130686581134796,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 72.1,
      "memory/max_allocated (GiB)": 72.1,
      "ppl": 1.36762,
      "step": 95,
      "tokens/total": 22724928,
      "tokens/train_per_sec_per_gpu": 10.76,
      "tokens/trainable": 189511
    },
    {
      "epoch": 0.1275323812686815,
      "grad_norm": 0.2720886170864105,
      "learning_rate": 9.971129059078082e-05,
      "loss": 0.31837907433509827,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 73.43,
      "memory/max_allocated (GiB)": 73.43,
      "ppl": 1.3749,
      "step": 96,
      "tokens/total": 22930688,
      "tokens/train_per_sec_per_gpu": 17.6,
      "tokens/trainable": 191527
    },
    {
      "epoch": 0.1288608435735636,
      "grad_norm": 0.2725374102592468,
      "learning_rate": 9.96996384223479e-05,
      "loss": 0.30179908871650696,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 72.92,
      "memory/max_allocated (GiB)": 72.92,
      "ppl": 1.35229,
      "step": 97,
      "tokens/total": 23159808,
      "tokens/train_per_sec_per_gpu": 13.92,
      "tokens/trainable": 193452
    },
    {
      "epoch": 0.1301893058784457,
      "grad_norm": 0.26988229155540466,
      "learning_rate": 9.968775645267103e-05,
      "loss": 0.27448657155036926,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.51,
      "memory/max_allocated (GiB)": 76.51,
      "ppl": 1.31585,
      "step": 98,
      "tokens/total": 23417536,
      "tokens/train_per_sec_per_gpu": 11.31,
      "tokens/trainable": 195481
    },
    {
      "epoch": 0.1315177681833278,
      "grad_norm": 0.2829805910587311,
      "learning_rate": 9.967564473669005e-05,
      "loss": 0.34205272793769836,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 73.65,
      "memory/max_allocated (GiB)": 73.65,
      "ppl": 1.40783,
      "step": 99,
      "tokens/total": 23646592,
      "tokens/train_per_sec_per_gpu": 15.37,
      "tokens/trainable": 197676
    },
    {
      "epoch": 0.1328462304882099,
      "grad_norm": 0.2858264446258545,
      "learning_rate": 9.96633033304071e-05,
      "loss": 0.3625262379646301,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.02,
      "memory/max_allocated (GiB)": 76.02,
      "ppl": 1.43695,
      "step": 100,
      "tokens/total": 23892672,
      "tokens/train_per_sec_per_gpu": 11.79,
      "tokens/trainable": 199656
    }
  ],
  "logging_steps": 1,
  "max_steps": 1506,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.4244649666108457e+18,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}