{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.20408163265306123,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0020408163265306124,
      "grad_norm": 4.2167510986328125,
      "learning_rate": 0.0,
      "loss": 1.313438892364502,
      "memory/device_reserved (GiB)": 75.76,
      "memory/max_active (GiB)": 74.61,
      "memory/max_allocated (GiB)": 74.61,
      "ppl": 3.71894,
      "step": 1,
      "tokens/total": 249728,
      "tokens/trainable": 1992
    },
    {
      "epoch": 0.004081632653061225,
      "grad_norm": 3.8650901317596436,
      "learning_rate": 3.448275862068966e-06,
      "loss": 1.2385082244873047,
      "memory/device_reserved (GiB)": 75.77,
      "memory/max_active (GiB)": 74.28,
      "memory/max_allocated (GiB)": 74.28,
      "ppl": 3.45046,
      "step": 2,
      "tokens/total": 483392,
      "tokens/train_per_sec_per_gpu": 12.03,
      "tokens/trainable": 3913
    },
    {
      "epoch": 0.006122448979591836,
      "grad_norm": 3.7181951999664307,
      "learning_rate": 6.896551724137932e-06,
      "loss": 1.2442840337753296,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 75.5,
      "memory/max_allocated (GiB)": 75.5,
      "ppl": 3.47045,
      "step": 3,
      "tokens/total": 726144,
      "tokens/train_per_sec_per_gpu": 12.05,
      "tokens/trainable": 5903
    },
    {
      "epoch": 0.00816326530612245,
      "grad_norm": 44.8486328125,
      "learning_rate": 1.0344827586206897e-05,
      "loss": 1.3412562608718872,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 72.18,
      "memory/max_allocated (GiB)": 72.18,
      "ppl": 3.82384,
      "step": 4,
      "tokens/total": 943168,
      "tokens/train_per_sec_per_gpu": 15.39,
      "tokens/trainable": 7867
    },
    {
      "epoch": 0.01020408163265306,
      "grad_norm": 3.929412841796875,
      "learning_rate": 1.3793103448275863e-05,
      "loss": 1.3407814502716064,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 75.42,
      "memory/max_allocated (GiB)": 75.42,
      "ppl": 3.82203,
      "step": 5,
      "tokens/total": 1169664,
      "tokens/train_per_sec_per_gpu": 12.41,
      "tokens/trainable": 9745
    },
    {
      "epoch": 0.012244897959183673,
      "grad_norm": 3.892376661300659,
      "learning_rate": 1.7241379310344828e-05,
      "loss": 1.2880868911743164,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 73.05,
      "memory/max_allocated (GiB)": 73.05,
      "ppl": 3.62584,
      "step": 6,
      "tokens/total": 1406336,
      "tokens/train_per_sec_per_gpu": 13.98,
      "tokens/trainable": 11650
    },
    {
      "epoch": 0.014285714285714285,
      "grad_norm": 4.0600128173828125,
      "learning_rate": 2.0689655172413793e-05,
      "loss": 1.3101682662963867,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 72.54,
      "memory/max_allocated (GiB)": 72.54,
      "ppl": 3.7068,
      "step": 7,
      "tokens/total": 1647104,
      "tokens/train_per_sec_per_gpu": 13.02,
      "tokens/trainable": 13571
    },
    {
      "epoch": 0.0163265306122449,
      "grad_norm": 3.508254289627075,
      "learning_rate": 2.413793103448276e-05,
      "loss": 1.1288416385650635,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 74.44,
      "memory/max_allocated (GiB)": 74.44,
      "ppl": 3.09207,
      "step": 8,
      "tokens/total": 1913472,
      "tokens/train_per_sec_per_gpu": 11.37,
      "tokens/trainable": 15743
    },
    {
      "epoch": 0.018367346938775512,
      "grad_norm": 4.13547945022583,
      "learning_rate": 2.7586206896551727e-05,
      "loss": 1.1055421829223633,
      "memory/device_reserved (GiB)": 76.4,
      "memory/max_active (GiB)": 75.26,
      "memory/max_allocated (GiB)": 75.26,
      "ppl": 3.02086,
      "step": 9,
      "tokens/total": 2159104,
      "tokens/train_per_sec_per_gpu": 13.39,
      "tokens/trainable": 17993
    },
    {
      "epoch": 0.02040816326530612,
      "grad_norm": 3.8564958572387695,
      "learning_rate": 3.103448275862069e-05,
      "loss": 1.1748199462890625,
      "memory/device_reserved (GiB)": 76.61,
      "memory/max_active (GiB)": 75.69,
      "memory/max_allocated (GiB)": 75.69,
      "ppl": 3.23756,
      "step": 10,
      "tokens/total": 2406336,
      "tokens/train_per_sec_per_gpu": 9.87,
      "tokens/trainable": 19603
    },
    {
      "epoch": 0.022448979591836733,
      "grad_norm": 2.662527322769165,
      "learning_rate": 3.4482758620689657e-05,
      "loss": 1.0491664409637451,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 76.81,
      "memory/max_allocated (GiB)": 76.81,
      "ppl": 2.85527,
      "step": 11,
      "tokens/total": 2637440,
      "tokens/train_per_sec_per_gpu": 13.08,
      "tokens/trainable": 21674
    },
    {
      "epoch": 0.024489795918367346,
      "grad_norm": 2.4675278663635254,
      "learning_rate": 3.793103448275862e-05,
      "loss": 0.7866945266723633,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 73.65,
      "memory/max_allocated (GiB)": 73.65,
      "ppl": 2.19613,
      "step": 12,
      "tokens/total": 2872896,
      "tokens/train_per_sec_per_gpu": 10.39,
      "tokens/trainable": 23481
    },
    {
      "epoch": 0.026530612244897958,
      "grad_norm": 2.2947044372558594,
      "learning_rate": 4.1379310344827587e-05,
      "loss": 0.7748140692710876,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 74.77,
      "memory/max_allocated (GiB)": 74.77,
      "ppl": 2.17019,
      "step": 13,
      "tokens/total": 3128768,
      "tokens/train_per_sec_per_gpu": 11.12,
      "tokens/trainable": 25355
    },
    {
      "epoch": 0.02857142857142857,
      "grad_norm": 1.5860671997070312,
      "learning_rate": 4.482758620689655e-05,
      "loss": 0.7104668021202087,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 75.83,
      "memory/max_allocated (GiB)": 75.83,
      "ppl": 2.03494,
      "step": 14,
      "tokens/total": 3354624,
      "tokens/train_per_sec_per_gpu": 13.4,
      "tokens/trainable": 27379
    },
    {
      "epoch": 0.030612244897959183,
      "grad_norm": 1.2193500995635986,
      "learning_rate": 4.827586206896552e-05,
      "loss": 0.6342951059341431,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 76.32,
      "memory/max_allocated (GiB)": 76.32,
      "ppl": 1.88569,
      "step": 15,
      "tokens/total": 3632576,
      "tokens/train_per_sec_per_gpu": 10.45,
      "tokens/trainable": 29348
    },
    {
      "epoch": 0.0326530612244898,
      "grad_norm": 0.8882229924201965,
      "learning_rate": 5.172413793103449e-05,
      "loss": 0.4914810061454773,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 76.86,
      "memory/max_allocated (GiB)": 76.86,
      "ppl": 1.63474,
      "step": 16,
      "tokens/total": 3893632,
      "tokens/train_per_sec_per_gpu": 11.24,
      "tokens/trainable": 31409
    },
    {
      "epoch": 0.03469387755102041,
      "grad_norm": 0.750299870967865,
      "learning_rate": 5.517241379310345e-05,
      "loss": 0.533316969871521,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 73.6,
      "memory/max_allocated (GiB)": 73.6,
      "ppl": 1.70458,
      "step": 17,
      "tokens/total": 4120000,
      "tokens/train_per_sec_per_gpu": 16.43,
      "tokens/trainable": 33895
    },
    {
      "epoch": 0.036734693877551024,
      "grad_norm": 0.9324926137924194,
      "learning_rate": 5.862068965517241e-05,
      "loss": 0.4928280711174011,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 73.19,
      "memory/max_allocated (GiB)": 73.19,
      "ppl": 1.63694,
      "step": 18,
      "tokens/total": 4334208,
      "tokens/train_per_sec_per_gpu": 13.05,
      "tokens/trainable": 35668
    },
    {
      "epoch": 0.03877551020408163,
      "grad_norm": 0.4430578649044037,
      "learning_rate": 6.206896551724138e-05,
      "loss": 0.4595395028591156,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 75.47,
      "memory/max_allocated (GiB)": 75.47,
      "ppl": 1.58334,
      "step": 19,
      "tokens/total": 4595520,
      "tokens/train_per_sec_per_gpu": 15.09,
      "tokens/trainable": 38006
    },
    {
      "epoch": 0.04081632653061224,
      "grad_norm": 0.6463696360588074,
      "learning_rate": 6.551724137931034e-05,
      "loss": 0.35258132219314575,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 73.57,
      "memory/max_allocated (GiB)": 73.57,
      "ppl": 1.42274,
      "step": 20,
      "tokens/total": 4823424,
      "tokens/train_per_sec_per_gpu": 15.34,
      "tokens/trainable": 40245
    },
    {
      "epoch": 0.04285714285714286,
      "grad_norm": 0.5994691252708435,
      "learning_rate": 6.896551724137931e-05,
      "loss": 0.4857128858566284,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 74.82,
      "memory/max_allocated (GiB)": 74.82,
      "ppl": 1.62533,
      "step": 21,
      "tokens/total": 5095296,
      "tokens/train_per_sec_per_gpu": 12.55,
      "tokens/trainable": 42288
    },
    {
      "epoch": 0.044897959183673466,
      "grad_norm": 0.37300923466682434,
      "learning_rate": 7.241379310344828e-05,
      "loss": 0.3811933100223541,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 76.37,
      "memory/max_allocated (GiB)": 76.37,
      "ppl": 1.46403,
      "step": 22,
      "tokens/total": 5361536,
      "tokens/train_per_sec_per_gpu": 12.0,
      "tokens/trainable": 44219
    },
    {
      "epoch": 0.04693877551020408,
      "grad_norm": 0.3530994951725006,
      "learning_rate": 7.586206896551724e-05,
      "loss": 0.47953876852989197,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 76.18,
      "memory/max_allocated (GiB)": 76.18,
      "ppl": 1.61533,
      "step": 23,
      "tokens/total": 5611776,
      "tokens/train_per_sec_per_gpu": 11.94,
      "tokens/trainable": 46265
    },
    {
      "epoch": 0.04897959183673469,
      "grad_norm": 0.4046308696269989,
      "learning_rate": 7.931034482758621e-05,
      "loss": 0.4454658031463623,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 75.09,
      "memory/max_allocated (GiB)": 75.09,
      "ppl": 1.56122,
      "step": 24,
      "tokens/total": 5842944,
      "tokens/train_per_sec_per_gpu": 13.15,
      "tokens/trainable": 48369
    },
    {
      "epoch": 0.05102040816326531,
      "grad_norm": 0.36315539479255676,
      "learning_rate": 8.275862068965517e-05,
      "loss": 0.31761664152145386,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 75.72,
      "memory/max_allocated (GiB)": 75.72,
      "ppl": 1.37385,
      "step": 25,
      "tokens/total": 6081152,
      "tokens/train_per_sec_per_gpu": 13.54,
      "tokens/trainable": 50389
    },
    {
      "epoch": 0.053061224489795916,
      "grad_norm": 0.5031322240829468,
      "learning_rate": 8.620689655172413e-05,
      "loss": 0.4392084777355194,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 73.22,
      "memory/max_allocated (GiB)": 73.22,
      "ppl": 1.55148,
      "step": 26,
      "tokens/total": 6343808,
      "tokens/train_per_sec_per_gpu": 7.89,
      "tokens/trainable": 52264
    },
    {
      "epoch": 0.05510204081632653,
      "grad_norm": 0.33433303236961365,
      "learning_rate": 8.96551724137931e-05,
      "loss": 0.3558434844017029,
      "memory/device_reserved (GiB)": 76.65,
      "memory/max_active (GiB)": 75.86,
      "memory/max_allocated (GiB)": 75.86,
      "ppl": 1.42738,
      "step": 27,
      "tokens/total": 6588608,
      "tokens/train_per_sec_per_gpu": 10.7,
      "tokens/trainable": 54059
    },
    {
      "epoch": 0.05714285714285714,
      "grad_norm": 0.39968281984329224,
      "learning_rate": 9.310344827586207e-05,
      "loss": 0.32551220059394836,
      "memory/device_reserved (GiB)": 76.65,
      "memory/max_active (GiB)": 73.19,
      "memory/max_allocated (GiB)": 73.19,
      "ppl": 1.38474,
      "step": 28,
      "tokens/total": 6826560,
      "tokens/train_per_sec_per_gpu": 10.04,
      "tokens/trainable": 55763
    },
    {
      "epoch": 0.05918367346938776,
      "grad_norm": 0.4150298833847046,
      "learning_rate": 9.655172413793105e-05,
      "loss": 0.40454256534576416,
      "memory/device_reserved (GiB)": 77.57,
      "memory/max_active (GiB)": 76.86,
      "memory/max_allocated (GiB)": 76.86,
      "ppl": 1.49862,
      "step": 29,
      "tokens/total": 7088000,
      "tokens/train_per_sec_per_gpu": 12.62,
      "tokens/trainable": 57812
    },
    {
      "epoch": 0.061224489795918366,
      "grad_norm": 0.6367387771606445,
      "learning_rate": 0.0001,
      "loss": 0.42207133769989014,
      "memory/device_reserved (GiB)": 77.57,
      "memory/max_active (GiB)": 75.91,
      "memory/max_allocated (GiB)": 75.91,
      "ppl": 1.52512,
      "step": 30,
      "tokens/total": 7351744,
      "tokens/train_per_sec_per_gpu": 11.4,
      "tokens/trainable": 59788
    },
    {
      "epoch": 0.06326530612244897,
      "grad_norm": 0.27459612488746643,
      "learning_rate": 9.999972717866783e-05,
      "loss": 0.26368996500968933,
      "memory/device_reserved (GiB)": 77.57,
      "memory/max_active (GiB)": 75.28,
      "memory/max_allocated (GiB)": 75.28,
      "ppl": 1.30172,
      "step": 31,
      "tokens/total": 7589888,
      "tokens/train_per_sec_per_gpu": 11.98,
      "tokens/trainable": 61698
    },
    {
      "epoch": 0.0653061224489796,
      "grad_norm": 0.3257054090499878,
      "learning_rate": 9.999890871764856e-05,
      "loss": 0.32186996936798096,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 76.75,
      "memory/max_allocated (GiB)": 76.75,
      "ppl": 1.37971,
      "step": 32,
      "tokens/total": 7812352,
      "tokens/train_per_sec_per_gpu": 11.42,
      "tokens/trainable": 63730
    },
    {
      "epoch": 0.0673469387755102,
      "grad_norm": 0.43009933829307556,
      "learning_rate": 9.999754462587395e-05,
      "loss": 0.3165552020072937,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 74.25,
      "memory/max_allocated (GiB)": 74.25,
      "ppl": 1.37239,
      "step": 33,
      "tokens/total": 8090112,
      "tokens/train_per_sec_per_gpu": 12.14,
      "tokens/trainable": 65782
    },
    {
      "epoch": 0.06938775510204082,
      "grad_norm": 0.29606345295906067,
      "learning_rate": 9.999563491823011e-05,
      "loss": 0.3038017749786377,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 76.51,
      "memory/max_allocated (GiB)": 76.51,
      "ppl": 1.355,
      "step": 34,
      "tokens/total": 8340224,
      "tokens/train_per_sec_per_gpu": 10.36,
      "tokens/trainable": 67558
    },
    {
      "epoch": 0.07142857142857142,
      "grad_norm": 0.26481109857559204,
      "learning_rate": 9.999317961555746e-05,
      "loss": 0.2744923532009125,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 74.41,
      "memory/max_allocated (GiB)": 74.41,
      "ppl": 1.31586,
      "step": 35,
      "tokens/total": 8599168,
      "tokens/train_per_sec_per_gpu": 9.93,
      "tokens/trainable": 69256
    },
    {
      "epoch": 0.07346938775510205,
      "grad_norm": 0.2818477153778076,
      "learning_rate": 9.999017874465026e-05,
      "loss": 0.32188448309898376,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 75.12,
      "memory/max_allocated (GiB)": 75.12,
      "ppl": 1.37973,
      "step": 36,
      "tokens/total": 8879872,
      "tokens/train_per_sec_per_gpu": 11.11,
      "tokens/trainable": 71213
    },
    {
      "epoch": 0.07551020408163266,
      "grad_norm": 0.31210315227508545,
      "learning_rate": 9.998663233825667e-05,
      "loss": 0.35000020265579224,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 72.67,
      "memory/max_allocated (GiB)": 72.67,
      "ppl": 1.41907,
      "step": 37,
      "tokens/total": 9092480,
      "tokens/train_per_sec_per_gpu": 15.52,
      "tokens/trainable": 73212
    },
    {
      "epoch": 0.07755102040816327,
      "grad_norm": 0.4802961051464081,
      "learning_rate": 9.998254043507804e-05,
      "loss": 0.3628653287887573,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 73.13,
      "memory/max_allocated (GiB)": 73.13,
      "ppl": 1.43744,
      "step": 38,
      "tokens/total": 9338944,
      "tokens/train_per_sec_per_gpu": 10.57,
      "tokens/trainable": 75027
    },
    {
      "epoch": 0.07959183673469387,
      "grad_norm": 0.3178337812423706,
      "learning_rate": 9.997790307976873e-05,
      "loss": 0.3139835596084595,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 72.86,
      "memory/max_allocated (GiB)": 72.86,
      "ppl": 1.36887,
      "step": 39,
      "tokens/total": 9552064,
      "tokens/train_per_sec_per_gpu": 13.09,
      "tokens/trainable": 76940
    },
    {
      "epoch": 0.08163265306122448,
      "grad_norm": 0.2966010868549347,
      "learning_rate": 9.997272032293553e-05,
      "loss": 0.34238967299461365,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 75.53,
      "memory/max_allocated (GiB)": 75.53,
      "ppl": 1.40831,
      "step": 40,
      "tokens/total": 9780480,
      "tokens/train_per_sec_per_gpu": 12.68,
      "tokens/trainable": 78924
    },
    {
      "epoch": 0.0836734693877551,
      "grad_norm": 0.3346101641654968,
      "learning_rate": 9.996699222113709e-05,
      "loss": 0.2840029001235962,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 74.11,
      "memory/max_allocated (GiB)": 74.11,
      "ppl": 1.32844,
      "step": 41,
      "tokens/total": 10032896,
      "tokens/train_per_sec_per_gpu": 10.74,
      "tokens/trainable": 80796
    },
    {
      "epoch": 0.08571428571428572,
      "grad_norm": 0.3669697940349579,
      "learning_rate": 9.996071883688334e-05,
      "loss": 0.38213422894477844,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 75.45,
      "memory/max_allocated (GiB)": 75.45,
      "ppl": 1.46541,
      "step": 42,
      "tokens/total": 10258880,
      "tokens/train_per_sec_per_gpu": 13.46,
      "tokens/trainable": 82842
    },
    {
      "epoch": 0.08775510204081632,
      "grad_norm": 0.3180081248283386,
      "learning_rate": 9.995390023863483e-05,
      "loss": 0.3140525817871094,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 74.41,
      "memory/max_allocated (GiB)": 74.41,
      "ppl": 1.36896,
      "step": 43,
      "tokens/total": 10509120,
      "tokens/train_per_sec_per_gpu": 13.79,
      "tokens/trainable": 84934
    },
    {
      "epoch": 0.08979591836734693,
      "grad_norm": 0.26853886246681213,
      "learning_rate": 9.994653650080188e-05,
      "loss": 0.3507641553878784,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 72.23,
      "memory/max_allocated (GiB)": 72.23,
      "ppl": 1.42015,
      "step": 44,
      "tokens/total": 10749568,
      "tokens/train_per_sec_per_gpu": 12.49,
      "tokens/trainable": 86784
    },
    {
      "epoch": 0.09183673469387756,
      "grad_norm": 0.26175931096076965,
      "learning_rate": 9.99386277037439e-05,
      "loss": 0.27149611711502075,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 74.66,
      "memory/max_allocated (GiB)": 74.66,
      "ppl": 1.31193,
      "step": 45,
      "tokens/total": 10991040,
      "tokens/train_per_sec_per_gpu": 12.04,
      "tokens/trainable": 88613
    },
    {
      "epoch": 0.09387755102040816,
      "grad_norm": 0.28962087631225586,
      "learning_rate": 9.993017393376844e-05,
      "loss": 0.34222665429115295,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 72.51,
      "memory/max_allocated (GiB)": 72.51,
      "ppl": 1.40808,
      "step": 46,
      "tokens/total": 11252224,
      "tokens/train_per_sec_per_gpu": 12.59,
      "tokens/trainable": 90773
    },
    {
      "epoch": 0.09591836734693877,
      "grad_norm": 0.41601240634918213,
      "learning_rate": 9.992117528313025e-05,
      "loss": 0.41904228925704956,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 74.3,
      "memory/max_allocated (GiB)": 74.3,
      "ppl": 1.5205,
      "step": 47,
      "tokens/total": 11496768,
      "tokens/train_per_sec_per_gpu": 13.52,
      "tokens/trainable": 93011
    },
    {
      "epoch": 0.09795918367346938,
      "grad_norm": 0.2995956838130951,
      "learning_rate": 9.991163185003029e-05,
      "loss": 0.32476919889450073,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 76.81,
      "memory/max_allocated (GiB)": 76.81,
      "ppl": 1.38371,
      "step": 48,
      "tokens/total": 11754816,
      "tokens/train_per_sec_per_gpu": 11.42,
      "tokens/trainable": 94980
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.35347771644592285,
      "learning_rate": 9.990154373861462e-05,
      "loss": 0.3826776444911957,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 72.81,
      "memory/max_allocated (GiB)": 72.81,
      "ppl": 1.46621,
      "step": 49,
      "tokens/total": 11997568,
      "tokens/train_per_sec_per_gpu": 14.25,
      "tokens/trainable": 97237
    },
    {
      "epoch": 0.10204081632653061,
      "grad_norm": 0.37251126766204834,
      "learning_rate": 9.989091105897335e-05,
      "loss": 0.2627924084663391,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 73.71,
      "memory/max_allocated (GiB)": 73.71,
      "ppl": 1.30056,
      "step": 50,
      "tokens/total": 12199616,
      "tokens/train_per_sec_per_gpu": 13.62,
      "tokens/trainable": 99077
    },
    {
      "epoch": 0.10408163265306122,
      "grad_norm": 0.4113054871559143,
      "learning_rate": 9.987973392713932e-05,
      "loss": 0.26638543605804443,
      "memory/device_reserved (GiB)": 77.69,
      "memory/max_active (GiB)": 75.09,
      "memory/max_allocated (GiB)": 75.09,
      "ppl": 1.30524,
      "step": 51,
      "tokens/total": 12458944,
      "tokens/train_per_sec_per_gpu": 8.67,
      "tokens/trainable": 100904
    },
    {
      "epoch": 0.10612244897959183,
      "grad_norm": 0.27514827251434326,
      "learning_rate": 9.986801246508695e-05,
      "loss": 0.3189134895801544,
      "memory/device_reserved (GiB)": 76.16,
      "memory/max_active (GiB)": 75.12,
      "memory/max_allocated (GiB)": 75.12,
      "ppl": 1.37563,
      "step": 52,
      "tokens/total": 12717056,
      "tokens/train_per_sec_per_gpu": 9.76,
      "tokens/trainable": 102655
    },
    {
      "epoch": 0.10816326530612246,
      "grad_norm": 0.2949848473072052,
      "learning_rate": 9.985574680073085e-05,
      "loss": 0.2989092767238617,
      "memory/device_reserved (GiB)": 76.16,
      "memory/max_active (GiB)": 74.2,
      "memory/max_allocated (GiB)": 74.2,
      "ppl": 1.34839,
      "step": 53,
      "tokens/total": 12979520,
      "tokens/train_per_sec_per_gpu": 10.41,
      "tokens/trainable": 104528
    },
    {
      "epoch": 0.11020408163265306,
      "grad_norm": 0.2654709815979004,
      "learning_rate": 9.984293706792438e-05,
      "loss": 0.3334200382232666,
      "memory/device_reserved (GiB)": 76.16,
      "memory/max_active (GiB)": 74.44,
      "memory/max_allocated (GiB)": 74.44,
      "ppl": 1.39573,
      "step": 54,
      "tokens/total": 13256256,
      "tokens/train_per_sec_per_gpu": 12.32,
      "tokens/trainable": 106680
    },
    {
      "epoch": 0.11224489795918367,
      "grad_norm": 0.36602842807769775,
      "learning_rate": 9.982958340645831e-05,
      "loss": 0.39756059646606445,
      "memory/device_reserved (GiB)": 76.16,
      "memory/max_active (GiB)": 72.13,
      "memory/max_allocated (GiB)": 72.13,
      "ppl": 1.48819,
      "step": 55,
      "tokens/total": 13483968,
      "tokens/train_per_sec_per_gpu": 13.09,
      "tokens/trainable": 108647
    },
    {
      "epoch": 0.11428571428571428,
      "grad_norm": 0.2748600244522095,
      "learning_rate": 9.981568596205917e-05,
      "loss": 0.3055681586265564,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 76.7,
      "memory/max_allocated (GiB)": 76.7,
      "ppl": 1.3574,
      "step": 56,
      "tokens/total": 13749568,
      "tokens/train_per_sec_per_gpu": 11.05,
      "tokens/trainable": 110620
    },
    {
      "epoch": 0.11632653061224489,
      "grad_norm": 0.32185810804367065,
      "learning_rate": 9.980124488638774e-05,
      "loss": 0.37074923515319824,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 72.1,
      "memory/max_allocated (GiB)": 72.1,
      "ppl": 1.44882,
      "step": 57,
      "tokens/total": 13985344,
      "tokens/train_per_sec_per_gpu": 16.08,
      "tokens/trainable": 112743
    },
    {
      "epoch": 0.11836734693877551,
      "grad_norm": 0.3911129832267761,
      "learning_rate": 9.978626033703732e-05,
      "loss": 0.3393978774547577,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 73.43,
      "memory/max_allocated (GiB)": 73.43,
      "ppl": 1.4041,
      "step": 58,
      "tokens/total": 14196480,
      "tokens/train_per_sec_per_gpu": 13.31,
      "tokens/trainable": 114572
    },
    {
      "epoch": 0.12040816326530612,
      "grad_norm": 0.3574847877025604,
      "learning_rate": 9.977073247753218e-05,
      "loss": 0.3169441223144531,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 74.69,
      "memory/max_allocated (GiB)": 74.69,
      "ppl": 1.37293,
      "step": 59,
      "tokens/total": 14435008,
      "tokens/train_per_sec_per_gpu": 13.0,
      "tokens/trainable": 116707
    },
    {
      "epoch": 0.12244897959183673,
      "grad_norm": 0.30045002698898315,
      "learning_rate": 9.97546614773255e-05,
      "loss": 0.2758142650127411,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 74.39,
      "memory/max_allocated (GiB)": 74.39,
      "ppl": 1.3176,
      "step": 60,
      "tokens/total": 14648576,
      "tokens/train_per_sec_per_gpu": 11.21,
      "tokens/trainable": 118447
    },
    {
      "epoch": 0.12448979591836734,
      "grad_norm": 0.25962135195732117,
      "learning_rate": 9.973804751179779e-05,
      "loss": 0.27344828844070435,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 75.67,
      "memory/max_allocated (GiB)": 75.67,
      "ppl": 1.31449,
      "step": 61,
      "tokens/total": 14898368,
      "tokens/train_per_sec_per_gpu": 12.14,
      "tokens/trainable": 120410
    },
    {
      "epoch": 0.12653061224489795,
      "grad_norm": 0.2542338967323303,
      "learning_rate": 9.972089076225477e-05,
      "loss": 0.2752627432346344,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 75.34,
      "memory/max_allocated (GiB)": 75.34,
      "ppl": 1.31688,
      "step": 62,
      "tokens/total": 15121792,
      "tokens/train_per_sec_per_gpu": 11.59,
      "tokens/trainable": 122325
    },
    {
      "epoch": 0.12857142857142856,
      "grad_norm": 0.3297300338745117,
      "learning_rate": 9.97031914159256e-05,
      "loss": 0.3394104242324829,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 74.49,
      "memory/max_allocated (GiB)": 74.49,
      "ppl": 1.40412,
      "step": 63,
      "tokens/total": 15389312,
      "tokens/train_per_sec_per_gpu": 12.41,
      "tokens/trainable": 124224
    },
    {
      "epoch": 0.1306122448979592,
      "grad_norm": 0.28276708722114563,
      "learning_rate": 9.968494966596058e-05,
      "loss": 0.31572309136390686,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 73.95,
      "memory/max_allocated (GiB)": 73.95,
      "ppl": 1.37125,
      "step": 64,
      "tokens/total": 15630144,
      "tokens/train_per_sec_per_gpu": 11.28,
      "tokens/trainable": 126004
    },
    {
      "epoch": 0.1326530612244898,
      "grad_norm": 0.31448298692703247,
      "learning_rate": 9.966616571142931e-05,
      "loss": 0.333346962928772,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 76.32,
      "memory/max_allocated (GiB)": 76.32,
      "ppl": 1.39563,
      "step": 65,
      "tokens/total": 15912896,
      "tokens/train_per_sec_per_gpu": 10.31,
      "tokens/trainable": 128003
    },
    {
      "epoch": 0.1346938775510204,
      "grad_norm": 0.31218236684799194,
      "learning_rate": 9.964683975731828e-05,
      "loss": 0.3107539415359497,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 75.94,
      "memory/max_allocated (GiB)": 75.94,
      "ppl": 1.36445,
      "step": 66,
      "tokens/total": 16185728,
      "tokens/train_per_sec_per_gpu": 12.49,
      "tokens/trainable": 130139
    },
    {
      "epoch": 0.13673469387755102,
      "grad_norm": 0.35470691323280334,
      "learning_rate": 9.962697201452881e-05,
      "loss": 0.3213728368282318,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 76.32,
      "memory/max_allocated (GiB)": 76.32,
      "ppl": 1.37902,
      "step": 67,
      "tokens/total": 16449728,
      "tokens/train_per_sec_per_gpu": 12.06,
      "tokens/trainable": 132201
    },
    {
      "epoch": 0.13877551020408163,
      "grad_norm": 0.34666892886161804,
      "learning_rate": 9.96065626998747e-05,
      "loss": 0.37185078859329224,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 76.48,
      "memory/max_allocated (GiB)": 76.48,
      "ppl": 1.45042,
      "step": 68,
      "tokens/total": 16719488,
      "tokens/train_per_sec_per_gpu": 9.9,
      "tokens/trainable": 134160
    },
    {
      "epoch": 0.14081632653061224,
      "grad_norm": 0.32578787207603455,
      "learning_rate": 9.958561203607975e-05,
      "loss": 0.3230299651622772,
      "memory/device_reserved (GiB)": 77.65,
      "memory/max_active (GiB)": 73.57,
      "memory/max_allocated (GiB)": 73.57,
      "ppl": 1.38131,
      "step": 69,
      "tokens/total": 16953152,
      "tokens/train_per_sec_per_gpu": 13.59,
      "tokens/trainable": 136020
    },
    {
      "epoch": 0.14285714285714285,
      "grad_norm": 0.5893422961235046,
      "learning_rate": 9.956412025177551e-05,
      "loss": 0.35762453079223633,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 76.73,
      "memory/max_allocated (GiB)": 76.73,
      "ppl": 1.42993,
      "step": 70,
      "tokens/total": 17219072,
      "tokens/train_per_sec_per_gpu": 11.73,
      "tokens/trainable": 137981
    },
    {
      "epoch": 0.14489795918367346,
      "grad_norm": 0.2695072293281555,
      "learning_rate": 9.954208758149867e-05,
      "loss": 0.23460917174816132,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 73.92,
      "memory/max_allocated (GiB)": 73.92,
      "ppl": 1.26441,
      "step": 71,
      "tokens/total": 17490944,
      "tokens/train_per_sec_per_gpu": 10.35,
      "tokens/trainable": 139847
    },
    {
      "epoch": 0.1469387755102041,
      "grad_norm": 0.2971181869506836,
      "learning_rate": 9.951951426568851e-05,
      "loss": 0.2664044499397278,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 72.34,
      "memory/max_allocated (GiB)": 72.34,
      "ppl": 1.30526,
      "step": 72,
      "tokens/total": 17732032,
      "tokens/train_per_sec_per_gpu": 13.7,
      "tokens/trainable": 141879
    },
    {
      "epoch": 0.1489795918367347,
      "grad_norm": 0.3867640495300293,
      "learning_rate": 9.949640055068432e-05,
      "loss": 0.4099368751049042,
      "memory/device_reserved (GiB)": 77.67,
      "memory/max_active (GiB)": 72.29,
      "memory/max_allocated (GiB)": 72.29,
      "ppl": 1.50672,
      "step": 73,
      "tokens/total": 17969088,
      "tokens/train_per_sec_per_gpu": 12.42,
      "tokens/trainable": 143793
    },
    {
      "epoch": 0.1510204081632653,
      "grad_norm": 0.28894367814064026,
      "learning_rate": 9.94727466887227e-05,
      "loss": 0.34861448407173157,
      "memory/device_reserved (GiB)": 77.77,
      "memory/max_active (GiB)": 76.84,
      "memory/max_allocated (GiB)": 76.84,
      "ppl": 1.4171,
      "step": 74,
      "tokens/total": 18238784,
      "tokens/train_per_sec_per_gpu": 11.22,
      "tokens/trainable": 145834
    },
    {
      "epoch": 0.15306122448979592,
      "grad_norm": 0.2793218493461609,
      "learning_rate": 9.944855293793475e-05,
      "loss": 0.2703927457332611,
      "memory/device_reserved (GiB)": 77.77,
      "memory/max_active (GiB)": 76.51,
      "memory/max_allocated (GiB)": 76.51,
      "ppl": 1.31048,
      "step": 75,
      "tokens/total": 18495168,
      "tokens/train_per_sec_per_gpu": 11.85,
      "tokens/trainable": 147888
    },
    {
      "epoch": 0.15510204081632653,
      "grad_norm": 0.2610492408275604,
      "learning_rate": 9.942381956234335e-05,
      "loss": 0.2886396050453186,
      "memory/device_reserved (GiB)": 77.77,
      "memory/max_active (GiB)": 76.4,
      "memory/max_allocated (GiB)": 76.4,
      "ppl": 1.33461,
      "step": 76,
      "tokens/total": 18734016,
      "tokens/train_per_sec_per_gpu": 9.23,
      "tokens/trainable": 149764
    },
    {
      "epoch": 0.15714285714285714,
      "grad_norm": 0.31321877241134644,
      "learning_rate": 9.939854683186016e-05,
      "loss": 0.32045096158981323,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 73.62,
      "memory/max_allocated (GiB)": 73.62,
      "ppl": 1.37775,
      "step": 77,
      "tokens/total": 18972672,
      "tokens/train_per_sec_per_gpu": 12.79,
      "tokens/trainable": 151857
    },
    {
      "epoch": 0.15918367346938775,
      "grad_norm": 0.34637442231178284,
      "learning_rate": 9.937273502228281e-05,
      "loss": 0.36426636576652527,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 75.8,
      "memory/max_allocated (GiB)": 75.8,
      "ppl": 1.43946,
      "step": 78,
      "tokens/total": 19195392,
      "tokens/train_per_sec_per_gpu": 13.15,
      "tokens/trainable": 153856
    },
    {
      "epoch": 0.16122448979591836,
      "grad_norm": 0.34072041511535645,
      "learning_rate": 9.934638441529179e-05,
      "loss": 0.34764987230300903,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 74.06,
      "memory/max_allocated (GiB)": 74.06,
      "ppl": 1.41574,
      "step": 79,
      "tokens/total": 19472064,
      "tokens/train_per_sec_per_gpu": 10.01,
      "tokens/trainable": 155739
    },
    {
      "epoch": 0.16326530612244897,
      "grad_norm": 0.4057715833187103,
      "learning_rate": 9.93194952984474e-05,
      "loss": 0.38709670305252075,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 74.06,
      "memory/max_allocated (GiB)": 74.06,
      "ppl": 1.4727,
      "step": 80,
      "tokens/total": 19724544,
      "tokens/train_per_sec_per_gpu": 13.56,
      "tokens/trainable": 158060
    },
    {
      "epoch": 0.1653061224489796,
      "grad_norm": 0.3533332347869873,
      "learning_rate": 9.929206796518662e-05,
      "loss": 0.3303971588611603,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 71.64,
      "memory/max_allocated (GiB)": 71.64,
      "ppl": 1.39152,
      "step": 81,
      "tokens/total": 19941440,
      "tokens/train_per_sec_per_gpu": 14.34,
      "tokens/trainable": 159968
    },
    {
      "epoch": 0.1673469387755102,
      "grad_norm": 0.576310932636261,
      "learning_rate": 9.926410271481994e-05,
      "loss": 0.29558199644088745,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 74.52,
      "memory/max_allocated (GiB)": 74.52,
      "ppl": 1.34391,
      "step": 82,
      "tokens/total": 20194560,
      "tokens/train_per_sec_per_gpu": 11.0,
      "tokens/trainable": 161784
    },
    {
      "epoch": 0.16938775510204082,
      "grad_norm": 0.2885003685951233,
      "learning_rate": 9.9235599852528e-05,
      "loss": 0.2939549684524536,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 75.09,
      "memory/max_allocated (GiB)": 75.09,
      "ppl": 1.34172,
      "step": 83,
      "tokens/total": 20438272,
      "tokens/train_per_sec_per_gpu": 12.77,
      "tokens/trainable": 163745
    },
    {
      "epoch": 0.17142857142857143,
      "grad_norm": 0.27577269077301025,
      "learning_rate": 9.920655968935838e-05,
      "loss": 0.3108862042427063,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 73.62,
      "memory/max_allocated (GiB)": 73.62,
      "ppl": 1.36463,
      "step": 84,
      "tokens/total": 20647488,
      "tokens/train_per_sec_per_gpu": 12.17,
      "tokens/trainable": 165491
    },
    {
      "epoch": 0.17346938775510204,
      "grad_norm": 0.2597094178199768,
      "learning_rate": 9.91769825422221e-05,
      "loss": 0.29625654220581055,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 73.54,
      "memory/max_allocated (GiB)": 73.54,
      "ppl": 1.34482,
      "step": 85,
      "tokens/total": 20908352,
      "tokens/train_per_sec_per_gpu": 12.03,
      "tokens/trainable": 167486
    },
    {
      "epoch": 0.17551020408163265,
      "grad_norm": 0.31459540128707886,
      "learning_rate": 9.914686873389025e-05,
      "loss": 0.3925963342189789,
      "memory/device_reserved (GiB)": 77.37,
      "memory/max_active (GiB)": 73.95,
      "memory/max_allocated (GiB)": 73.95,
      "ppl": 1.48082,
      "step": 86,
      "tokens/total": 21166656,
      "tokens/train_per_sec_per_gpu": 12.74,
      "tokens/trainable": 169501
    },
    {
      "epoch": 0.17755102040816326,
      "grad_norm": 0.3207368552684784,
      "learning_rate": 9.91162185929904e-05,
      "loss": 0.36425894498825073,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.35,
      "memory/max_allocated (GiB)": 76.35,
      "ppl": 1.43945,
      "step": 87,
      "tokens/total": 21438912,
      "tokens/train_per_sec_per_gpu": 11.2,
      "tokens/trainable": 171575
    },
    {
      "epoch": 0.17959183673469387,
      "grad_norm": 0.30726850032806396,
      "learning_rate": 9.908503245400302e-05,
      "loss": 0.363973468542099,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 75.56,
      "memory/max_allocated (GiB)": 75.56,
      "ppl": 1.43904,
      "step": 88,
      "tokens/total": 21722176,
      "tokens/train_per_sec_per_gpu": 12.12,
      "tokens/trainable": 173659
    },
    {
      "epoch": 0.1816326530612245,
      "grad_norm": 0.3307989835739136,
      "learning_rate": 9.90533106572579e-05,
      "loss": 0.37638866901397705,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 74.11,
      "memory/max_allocated (GiB)": 74.11,
      "ppl": 1.45701,
      "step": 89,
      "tokens/total": 21958848,
      "tokens/train_per_sec_per_gpu": 13.93,
      "tokens/trainable": 175607
    },
    {
      "epoch": 0.1836734693877551,
      "grad_norm": 0.2957727313041687,
      "learning_rate": 9.90210535489303e-05,
      "loss": 0.2558442950248718,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 72.89,
      "memory/max_allocated (GiB)": 72.89,
      "ppl": 1.29155,
      "step": 90,
      "tokens/total": 22210112,
      "tokens/train_per_sec_per_gpu": 13.12,
      "tokens/trainable": 177797
    },
    {
      "epoch": 0.18571428571428572,
      "grad_norm": 0.2863942086696625,
      "learning_rate": 9.898826148103737e-05,
      "loss": 0.3638124465942383,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 75.34,
      "memory/max_allocated (GiB)": 75.34,
      "ppl": 1.4388,
      "step": 91,
      "tokens/total": 22450496,
      "tokens/train_per_sec_per_gpu": 11.32,
      "tokens/trainable": 179665
    },
    {
      "epoch": 0.18775510204081633,
      "grad_norm": 0.2795948088169098,
      "learning_rate": 9.895493481143411e-05,
      "loss": 0.31911370158195496,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 74.03,
      "memory/max_allocated (GiB)": 74.03,
      "ppl": 1.37591,
      "step": 92,
      "tokens/total": 22683776,
      "tokens/train_per_sec_per_gpu": 13.26,
      "tokens/trainable": 181685
    },
    {
      "epoch": 0.18979591836734694,
      "grad_norm": 0.25272026658058167,
      "learning_rate": 9.892107390380958e-05,
      "loss": 0.2508898377418518,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 72.37,
      "memory/max_allocated (GiB)": 72.37,
      "ppl": 1.28517,
      "step": 93,
      "tokens/total": 22917888,
      "tokens/train_per_sec_per_gpu": 12.28,
      "tokens/trainable": 183472
    },
    {
      "epoch": 0.19183673469387755,
      "grad_norm": 0.35054582357406616,
      "learning_rate": 9.888667912768288e-05,
      "loss": 0.35517990589141846,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 74.39,
      "memory/max_allocated (GiB)": 74.39,
      "ppl": 1.42644,
      "step": 94,
      "tokens/total": 23187520,
      "tokens/train_per_sec_per_gpu": 10.77,
      "tokens/trainable": 185479
    },
    {
      "epoch": 0.19387755102040816,
      "grad_norm": 0.26934683322906494,
      "learning_rate": 9.885175085839919e-05,
      "loss": 0.31262412667274475,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 74.47,
      "memory/max_allocated (GiB)": 74.47,
      "ppl": 1.36701,
      "step": 95,
      "tokens/total": 23443008,
      "tokens/train_per_sec_per_gpu": 13.83,
      "tokens/trainable": 187567
    },
    {
      "epoch": 0.19591836734693877,
      "grad_norm": 0.5132613778114319,
      "learning_rate": 9.881628947712556e-05,
      "loss": 0.273453950881958,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 73.79,
      "memory/max_allocated (GiB)": 73.79,
      "ppl": 1.3145,
      "step": 96,
      "tokens/total": 23698432,
      "tokens/train_per_sec_per_gpu": 10.01,
      "tokens/trainable": 189220
    },
    {
      "epoch": 0.19795918367346937,
      "grad_norm": 0.4519254267215729,
      "learning_rate": 9.878029537084684e-05,
      "loss": 0.28466570377349854,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 74.82,
      "memory/max_allocated (GiB)": 74.82,
      "ppl": 1.32932,
      "step": 97,
      "tokens/total": 23947968,
      "tokens/train_per_sec_per_gpu": 10.12,
      "tokens/trainable": 190943
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.2954767048358917,
      "learning_rate": 9.874376893236145e-05,
      "loss": 0.33792048692703247,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 73.79,
      "memory/max_allocated (GiB)": 73.79,
      "ppl": 1.40203,
      "step": 98,
      "tokens/total": 24192704,
      "tokens/train_per_sec_per_gpu": 13.04,
      "tokens/trainable": 192967
    },
    {
      "epoch": 0.20204081632653062,
      "grad_norm": 0.3651937246322632,
      "learning_rate": 9.870671056027705e-05,
      "loss": 0.30805450677871704,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 73.3,
      "memory/max_allocated (GiB)": 73.3,
      "ppl": 1.36078,
      "step": 99,
      "tokens/total": 24420672,
      "tokens/train_per_sec_per_gpu": 13.96,
      "tokens/trainable": 194997
    },
    {
      "epoch": 0.20408163265306123,
      "grad_norm": 0.2739502489566803,
      "learning_rate": 9.866912065900621e-05,
      "loss": 0.2902035117149353,
      "memory/device_reserved (GiB)": 77.53,
      "memory/max_active (GiB)": 76.51,
      "memory/max_allocated (GiB)": 76.51,
      "ppl": 1.3367,
      "step": 100,
      "tokens/total": 24663168,
      "tokens/train_per_sec_per_gpu": 13.35,
      "tokens/trainable": 197125
    }
  ],
  "logging_steps": 1,
  "max_steps": 980,
  "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": 4.5671460522216653e+18,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}