{
  "best_metric": 2.2837178707122803,
  "best_model_checkpoint": "/exports/eddie/scratch/s1970716/models/longt5_xl_sfd_memsum_40/checkpoint-57",
  "epoch": 38.95652173913044,
  "eval_steps": 500,
  "global_step": 1120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "grad_norm": 1.5414308309555054,
      "learning_rate": 0.001,
      "loss": 3.1508,
      "step": 2
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.1113317012786865,
      "learning_rate": 0.001,
      "loss": 3.2124,
      "step": 4
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.7978044152259827,
      "learning_rate": 0.001,
      "loss": 2.9466,
      "step": 6
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.6611266136169434,
      "learning_rate": 0.001,
      "loss": 2.8914,
      "step": 8
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.6737671494483948,
      "learning_rate": 0.001,
      "loss": 2.7507,
      "step": 10
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.41579362750053406,
      "learning_rate": 0.001,
      "loss": 2.6653,
      "step": 12
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.37621861696243286,
      "learning_rate": 0.001,
      "loss": 2.579,
      "step": 14
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.581123411655426,
      "learning_rate": 0.001,
      "loss": 2.631,
      "step": 16
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.6653873920440674,
      "learning_rate": 0.001,
      "loss": 2.6978,
      "step": 18
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.45471808314323425,
      "learning_rate": 0.001,
      "loss": 2.6541,
      "step": 20
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.38716837763786316,
      "learning_rate": 0.001,
      "loss": 2.5325,
      "step": 22
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.40614280104637146,
      "learning_rate": 0.001,
      "loss": 2.5584,
      "step": 24
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.3157476484775543,
      "learning_rate": 0.001,
      "loss": 2.5054,
      "step": 26
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.3763532042503357,
      "learning_rate": 0.001,
      "loss": 2.5238,
      "step": 28
    },
    {
      "epoch": 0.97,
      "eval_loss": 2.314685106277466,
      "eval_runtime": 14.2455,
      "eval_samples_per_second": 23.727,
      "eval_steps_per_second": 3.018,
      "step": 28
    },
    {
      "epoch": 1.04,
      "grad_norm": 0.3202365040779114,
      "learning_rate": 0.001,
      "loss": 2.3018,
      "step": 30
    },
    {
      "epoch": 1.11,
      "grad_norm": 0.4395117461681366,
      "learning_rate": 0.001,
      "loss": 2.1284,
      "step": 32
    },
    {
      "epoch": 1.18,
      "grad_norm": 0.47938525676727295,
      "learning_rate": 0.001,
      "loss": 2.1782,
      "step": 34
    },
    {
      "epoch": 1.25,
      "grad_norm": 1.0489932298660278,
      "learning_rate": 0.001,
      "loss": 2.1666,
      "step": 36
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.34978240728378296,
      "learning_rate": 0.001,
      "loss": 2.1598,
      "step": 38
    },
    {
      "epoch": 1.39,
      "grad_norm": 0.28448227047920227,
      "learning_rate": 0.001,
      "loss": 2.023,
      "step": 40
    },
    {
      "epoch": 1.46,
      "grad_norm": 0.3449482023715973,
      "learning_rate": 0.001,
      "loss": 2.0545,
      "step": 42
    },
    {
      "epoch": 1.53,
      "grad_norm": 0.6303105354309082,
      "learning_rate": 0.001,
      "loss": 2.1264,
      "step": 44
    },
    {
      "epoch": 1.6,
      "grad_norm": 0.3750413656234741,
      "learning_rate": 0.001,
      "loss": 2.1176,
      "step": 46
    },
    {
      "epoch": 1.67,
      "grad_norm": 0.44775936007499695,
      "learning_rate": 0.001,
      "loss": 2.1624,
      "step": 48
    },
    {
      "epoch": 1.74,
      "grad_norm": 0.34849441051483154,
      "learning_rate": 0.001,
      "loss": 2.0809,
      "step": 50
    },
    {
      "epoch": 1.81,
      "grad_norm": 0.28750795125961304,
      "learning_rate": 0.001,
      "loss": 2.0913,
      "step": 52
    },
    {
      "epoch": 1.88,
      "grad_norm": 0.2710261940956116,
      "learning_rate": 0.001,
      "loss": 2.0855,
      "step": 54
    },
    {
      "epoch": 1.95,
      "grad_norm": 0.28888803720474243,
      "learning_rate": 0.001,
      "loss": 2.1298,
      "step": 56
    },
    {
      "epoch": 1.98,
      "eval_loss": 2.2837178707122803,
      "eval_runtime": 14.2329,
      "eval_samples_per_second": 23.748,
      "eval_steps_per_second": 3.021,
      "step": 57
    },
    {
      "epoch": 2.02,
      "grad_norm": 0.2953917682170868,
      "learning_rate": 0.001,
      "loss": 1.9808,
      "step": 58
    },
    {
      "epoch": 2.09,
      "grad_norm": 2.0295698642730713,
      "learning_rate": 0.001,
      "loss": 1.7208,
      "step": 60
    },
    {
      "epoch": 2.16,
      "grad_norm": 0.4198295772075653,
      "learning_rate": 0.001,
      "loss": 1.7108,
      "step": 62
    },
    {
      "epoch": 2.23,
      "grad_norm": 0.3806304335594177,
      "learning_rate": 0.001,
      "loss": 1.677,
      "step": 64
    },
    {
      "epoch": 2.3,
      "grad_norm": 0.36914414167404175,
      "learning_rate": 0.001,
      "loss": 1.6717,
      "step": 66
    },
    {
      "epoch": 2.37,
      "grad_norm": 0.34834641218185425,
      "learning_rate": 0.001,
      "loss": 1.6516,
      "step": 68
    },
    {
      "epoch": 2.43,
      "grad_norm": 0.3333311080932617,
      "learning_rate": 0.001,
      "loss": 1.6677,
      "step": 70
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.41852736473083496,
      "learning_rate": 0.001,
      "loss": 1.6135,
      "step": 72
    },
    {
      "epoch": 2.57,
      "grad_norm": 0.40690872073173523,
      "learning_rate": 0.001,
      "loss": 1.6447,
      "step": 74
    },
    {
      "epoch": 2.64,
      "grad_norm": 0.4816026985645294,
      "learning_rate": 0.001,
      "loss": 1.7096,
      "step": 76
    },
    {
      "epoch": 2.71,
      "grad_norm": 0.8278970122337341,
      "learning_rate": 0.001,
      "loss": 1.6721,
      "step": 78
    },
    {
      "epoch": 2.78,
      "grad_norm": 0.27179405093193054,
      "learning_rate": 0.001,
      "loss": 1.6351,
      "step": 80
    },
    {
      "epoch": 2.85,
      "grad_norm": 0.2935534417629242,
      "learning_rate": 0.001,
      "loss": 1.6402,
      "step": 82
    },
    {
      "epoch": 2.92,
      "grad_norm": 0.3557533919811249,
      "learning_rate": 0.001,
      "loss": 1.6755,
      "step": 84
    },
    {
      "epoch": 2.99,
      "grad_norm": 0.3663090169429779,
      "learning_rate": 0.001,
      "loss": 1.7525,
      "step": 86
    },
    {
      "epoch": 2.99,
      "eval_loss": 2.333530902862549,
      "eval_runtime": 14.2215,
      "eval_samples_per_second": 23.767,
      "eval_steps_per_second": 3.024,
      "step": 86
    },
    {
      "epoch": 3.06,
      "grad_norm": 0.37869980931282043,
      "learning_rate": 0.001,
      "loss": 1.3103,
      "step": 88
    },
    {
      "epoch": 3.13,
      "grad_norm": 0.32068952918052673,
      "learning_rate": 0.001,
      "loss": 1.2763,
      "step": 90
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.33284562826156616,
      "learning_rate": 0.001,
      "loss": 1.2235,
      "step": 92
    },
    {
      "epoch": 3.27,
      "grad_norm": 0.4027257561683655,
      "learning_rate": 0.001,
      "loss": 1.2606,
      "step": 94
    },
    {
      "epoch": 3.34,
      "grad_norm": 0.8390576839447021,
      "learning_rate": 0.001,
      "loss": 1.2426,
      "step": 96
    },
    {
      "epoch": 3.41,
      "grad_norm": 0.3938799798488617,
      "learning_rate": 0.001,
      "loss": 1.2499,
      "step": 98
    },
    {
      "epoch": 3.48,
      "grad_norm": 0.39702850580215454,
      "learning_rate": 0.001,
      "loss": 1.3165,
      "step": 100
    },
    {
      "epoch": 3.55,
      "grad_norm": 0.4196929633617401,
      "learning_rate": 0.001,
      "loss": 1.2884,
      "step": 102
    },
    {
      "epoch": 3.62,
      "grad_norm": 0.4435763657093048,
      "learning_rate": 0.001,
      "loss": 1.3407,
      "step": 104
    },
    {
      "epoch": 3.69,
      "grad_norm": 0.3989652395248413,
      "learning_rate": 0.001,
      "loss": 1.3119,
      "step": 106
    },
    {
      "epoch": 3.76,
      "grad_norm": 0.35742542147636414,
      "learning_rate": 0.001,
      "loss": 1.2882,
      "step": 108
    },
    {
      "epoch": 3.83,
      "grad_norm": 0.3512905538082123,
      "learning_rate": 0.001,
      "loss": 1.276,
      "step": 110
    },
    {
      "epoch": 3.9,
      "grad_norm": 0.35384494066238403,
      "learning_rate": 0.001,
      "loss": 1.31,
      "step": 112
    },
    {
      "epoch": 3.97,
      "grad_norm": 0.38055282831192017,
      "learning_rate": 0.001,
      "loss": 1.2954,
      "step": 114
    },
    {
      "epoch": 4.0,
      "eval_loss": 2.4995365142822266,
      "eval_runtime": 14.2261,
      "eval_samples_per_second": 23.759,
      "eval_steps_per_second": 3.023,
      "step": 115
    },
    {
      "epoch": 4.03,
      "grad_norm": 0.33426040410995483,
      "learning_rate": 0.001,
      "loss": 1.0702,
      "step": 116
    },
    {
      "epoch": 4.1,
      "grad_norm": 0.3915540277957916,
      "learning_rate": 0.001,
      "loss": 0.93,
      "step": 118
    },
    {
      "epoch": 4.17,
      "grad_norm": 0.3885715901851654,
      "learning_rate": 0.001,
      "loss": 0.8923,
      "step": 120
    },
    {
      "epoch": 4.24,
      "grad_norm": 0.3450148105621338,
      "learning_rate": 0.001,
      "loss": 0.9367,
      "step": 122
    },
    {
      "epoch": 4.31,
      "grad_norm": 0.34307780861854553,
      "learning_rate": 0.001,
      "loss": 0.9471,
      "step": 124
    },
    {
      "epoch": 4.38,
      "grad_norm": 0.3358236253261566,
      "learning_rate": 0.001,
      "loss": 0.9597,
      "step": 126
    },
    {
      "epoch": 4.45,
      "grad_norm": 0.32632964849472046,
      "learning_rate": 0.001,
      "loss": 0.9678,
      "step": 128
    },
    {
      "epoch": 4.52,
      "grad_norm": 0.3462707996368408,
      "learning_rate": 0.001,
      "loss": 0.975,
      "step": 130
    },
    {
      "epoch": 4.59,
      "grad_norm": 0.4402617812156677,
      "learning_rate": 0.001,
      "loss": 0.9908,
      "step": 132
    },
    {
      "epoch": 4.66,
      "grad_norm": 0.4284449815750122,
      "learning_rate": 0.001,
      "loss": 0.9721,
      "step": 134
    },
    {
      "epoch": 4.73,
      "grad_norm": 0.3820173144340515,
      "learning_rate": 0.001,
      "loss": 0.983,
      "step": 136
    },
    {
      "epoch": 4.8,
      "grad_norm": 0.5288845300674438,
      "learning_rate": 0.001,
      "loss": 0.9376,
      "step": 138
    },
    {
      "epoch": 4.87,
      "grad_norm": 0.489346981048584,
      "learning_rate": 0.001,
      "loss": 0.9801,
      "step": 140
    },
    {
      "epoch": 4.94,
      "grad_norm": 0.4548284709453583,
      "learning_rate": 0.001,
      "loss": 1.0518,
      "step": 142
    },
    {
      "epoch": 4.97,
      "eval_loss": 2.832628011703491,
      "eval_runtime": 14.24,
      "eval_samples_per_second": 23.736,
      "eval_steps_per_second": 3.02,
      "step": 143
    },
    {
      "epoch": 5.01,
      "grad_norm": 0.4093753397464752,
      "learning_rate": 0.001,
      "loss": 0.9381,
      "step": 144
    },
    {
      "epoch": 5.08,
      "grad_norm": 0.5986877679824829,
      "learning_rate": 0.001,
      "loss": 0.6649,
      "step": 146
    },
    {
      "epoch": 5.15,
      "grad_norm": 0.4789908230304718,
      "learning_rate": 0.001,
      "loss": 0.6925,
      "step": 148
    },
    {
      "epoch": 5.22,
      "grad_norm": 0.4392305612564087,
      "learning_rate": 0.001,
      "loss": 0.6901,
      "step": 150
    },
    {
      "epoch": 5.29,
      "grad_norm": 0.49618610739707947,
      "learning_rate": 0.001,
      "loss": 0.7173,
      "step": 152
    },
    {
      "epoch": 5.36,
      "grad_norm": 0.513729453086853,
      "learning_rate": 0.001,
      "loss": 0.6951,
      "step": 154
    },
    {
      "epoch": 5.43,
      "grad_norm": 0.4622480571269989,
      "learning_rate": 0.001,
      "loss": 0.6757,
      "step": 156
    },
    {
      "epoch": 5.5,
      "grad_norm": 0.4192960560321808,
      "learning_rate": 0.001,
      "loss": 0.7125,
      "step": 158
    },
    {
      "epoch": 5.57,
      "grad_norm": 0.3887766897678375,
      "learning_rate": 0.001,
      "loss": 0.6905,
      "step": 160
    },
    {
      "epoch": 5.63,
      "grad_norm": 0.3633786141872406,
      "learning_rate": 0.001,
      "loss": 0.73,
      "step": 162
    },
    {
      "epoch": 5.7,
      "grad_norm": 0.4010884761810303,
      "learning_rate": 0.001,
      "loss": 0.7473,
      "step": 164
    },
    {
      "epoch": 5.77,
      "grad_norm": 0.4541136920452118,
      "learning_rate": 0.001,
      "loss": 0.7659,
      "step": 166
    },
    {
      "epoch": 5.84,
      "grad_norm": 0.4243777394294739,
      "learning_rate": 0.001,
      "loss": 0.7742,
      "step": 168
    },
    {
      "epoch": 5.91,
      "grad_norm": 0.40389642119407654,
      "learning_rate": 0.001,
      "loss": 0.7312,
      "step": 170
    },
    {
      "epoch": 5.98,
      "grad_norm": 0.39572134613990784,
      "learning_rate": 0.001,
      "loss": 0.7083,
      "step": 172
    },
    {
      "epoch": 5.98,
      "eval_loss": 2.9095335006713867,
      "eval_runtime": 14.2201,
      "eval_samples_per_second": 23.769,
      "eval_steps_per_second": 3.024,
      "step": 172
    },
    {
      "epoch": 6.05,
      "grad_norm": 0.34924399852752686,
      "learning_rate": 0.001,
      "loss": 0.5277,
      "step": 174
    },
    {
      "epoch": 6.12,
      "grad_norm": 0.42162197828292847,
      "learning_rate": 0.001,
      "loss": 0.5141,
      "step": 176
    },
    {
      "epoch": 6.19,
      "grad_norm": 0.43131959438323975,
      "learning_rate": 0.001,
      "loss": 0.5295,
      "step": 178
    },
    {
      "epoch": 6.26,
      "grad_norm": 0.3414030075073242,
      "learning_rate": 0.001,
      "loss": 0.5293,
      "step": 180
    },
    {
      "epoch": 6.33,
      "grad_norm": 0.2947409451007843,
      "learning_rate": 0.001,
      "loss": 0.5094,
      "step": 182
    },
    {
      "epoch": 6.4,
      "grad_norm": 0.3493882119655609,
      "learning_rate": 0.001,
      "loss": 0.5103,
      "step": 184
    },
    {
      "epoch": 6.47,
      "grad_norm": 0.32564929127693176,
      "learning_rate": 0.001,
      "loss": 0.4955,
      "step": 186
    },
    {
      "epoch": 6.54,
      "grad_norm": 0.3580136299133301,
      "learning_rate": 0.001,
      "loss": 0.4923,
      "step": 188
    },
    {
      "epoch": 6.61,
      "grad_norm": 0.40878280997276306,
      "learning_rate": 0.001,
      "loss": 0.5312,
      "step": 190
    },
    {
      "epoch": 6.68,
      "grad_norm": 0.42687222361564636,
      "learning_rate": 0.001,
      "loss": 0.5977,
      "step": 192
    },
    {
      "epoch": 6.75,
      "grad_norm": 0.3438556492328644,
      "learning_rate": 0.001,
      "loss": 0.5811,
      "step": 194
    },
    {
      "epoch": 6.82,
      "grad_norm": 0.3497261703014374,
      "learning_rate": 0.001,
      "loss": 0.542,
      "step": 196
    },
    {
      "epoch": 6.89,
      "grad_norm": 0.5538045167922974,
      "learning_rate": 0.001,
      "loss": 0.5392,
      "step": 198
    },
    {
      "epoch": 6.96,
      "grad_norm": 0.40168508887290955,
      "learning_rate": 0.001,
      "loss": 0.5124,
      "step": 200
    },
    {
      "epoch": 6.99,
      "eval_loss": 3.4108099937438965,
      "eval_runtime": 14.2162,
      "eval_samples_per_second": 23.776,
      "eval_steps_per_second": 3.025,
      "step": 201
    },
    {
      "epoch": 7.03,
      "grad_norm": 0.45531949400901794,
      "learning_rate": 0.001,
      "loss": 0.4621,
      "step": 202
    },
    {
      "epoch": 7.1,
      "grad_norm": 0.420217901468277,
      "learning_rate": 0.001,
      "loss": 0.3526,
      "step": 204
    },
    {
      "epoch": 7.17,
      "grad_norm": 0.39295145869255066,
      "learning_rate": 0.001,
      "loss": 0.3522,
      "step": 206
    },
    {
      "epoch": 7.23,
      "grad_norm": 0.42002934217453003,
      "learning_rate": 0.001,
      "loss": 0.37,
      "step": 208
    },
    {
      "epoch": 7.3,
      "grad_norm": 0.3919942378997803,
      "learning_rate": 0.001,
      "loss": 0.4123,
      "step": 210
    },
    {
      "epoch": 7.37,
      "grad_norm": 0.34964632987976074,
      "learning_rate": 0.001,
      "loss": 0.4155,
      "step": 212
    },
    {
      "epoch": 7.44,
      "grad_norm": 0.3752292990684509,
      "learning_rate": 0.001,
      "loss": 0.392,
      "step": 214
    },
    {
      "epoch": 7.51,
      "grad_norm": 0.31345558166503906,
      "learning_rate": 0.001,
      "loss": 0.3973,
      "step": 216
    },
    {
      "epoch": 7.58,
      "grad_norm": 0.32503190636634827,
      "learning_rate": 0.001,
      "loss": 0.3788,
      "step": 218
    },
    {
      "epoch": 7.65,
      "grad_norm": 0.40439680218696594,
      "learning_rate": 0.001,
      "loss": 0.3997,
      "step": 220
    },
    {
      "epoch": 7.72,
      "grad_norm": 0.5034092664718628,
      "learning_rate": 0.001,
      "loss": 0.3934,
      "step": 222
    },
    {
      "epoch": 7.79,
      "grad_norm": 0.4480283260345459,
      "learning_rate": 0.001,
      "loss": 0.4095,
      "step": 224
    },
    {
      "epoch": 7.86,
      "grad_norm": 0.44010230898857117,
      "learning_rate": 0.001,
      "loss": 0.3963,
      "step": 226
    },
    {
      "epoch": 7.93,
      "grad_norm": 0.4241428077220917,
      "learning_rate": 0.001,
      "loss": 0.4081,
      "step": 228
    },
    {
      "epoch": 8.0,
      "grad_norm": 1.4331179857254028,
      "learning_rate": 0.001,
      "loss": 0.4503,
      "step": 230
    },
    {
      "epoch": 8.0,
      "eval_loss": 3.445927858352661,
      "eval_runtime": 14.2138,
      "eval_samples_per_second": 23.78,
      "eval_steps_per_second": 3.025,
      "step": 230
    },
    {
      "epoch": 8.07,
      "grad_norm": 0.26638248562812805,
      "learning_rate": 0.001,
      "loss": 0.2491,
      "step": 232
    },
    {
      "epoch": 8.14,
      "grad_norm": 0.30395275354385376,
      "learning_rate": 0.001,
      "loss": 0.2501,
      "step": 234
    },
    {
      "epoch": 8.21,
      "grad_norm": 0.36178329586982727,
      "learning_rate": 0.001,
      "loss": 0.2818,
      "step": 236
    },
    {
      "epoch": 8.28,
      "grad_norm": 0.33713603019714355,
      "learning_rate": 0.001,
      "loss": 0.3005,
      "step": 238
    },
    {
      "epoch": 8.35,
      "grad_norm": 0.2999133765697479,
      "learning_rate": 0.001,
      "loss": 0.2863,
      "step": 240
    },
    {
      "epoch": 8.42,
      "grad_norm": 0.4105747938156128,
      "learning_rate": 0.001,
      "loss": 0.2946,
      "step": 242
    },
    {
      "epoch": 8.49,
      "grad_norm": 0.3367477357387543,
      "learning_rate": 0.001,
      "loss": 0.3052,
      "step": 244
    },
    {
      "epoch": 8.56,
      "grad_norm": 0.3373625874519348,
      "learning_rate": 0.001,
      "loss": 0.3166,
      "step": 246
    },
    {
      "epoch": 8.63,
      "grad_norm": 0.2816782295703888,
      "learning_rate": 0.001,
      "loss": 0.2819,
      "step": 248
    },
    {
      "epoch": 8.7,
      "grad_norm": 0.30990245938301086,
      "learning_rate": 0.001,
      "loss": 0.3023,
      "step": 250
    },
    {
      "epoch": 8.77,
      "grad_norm": 0.317903608083725,
      "learning_rate": 0.001,
      "loss": 0.3128,
      "step": 252
    },
    {
      "epoch": 8.83,
      "grad_norm": 0.3024023473262787,
      "learning_rate": 0.001,
      "loss": 0.3167,
      "step": 254
    },
    {
      "epoch": 8.9,
      "grad_norm": 0.3257545828819275,
      "learning_rate": 0.001,
      "loss": 0.3121,
      "step": 256
    },
    {
      "epoch": 8.97,
      "grad_norm": 0.34182751178741455,
      "learning_rate": 0.001,
      "loss": 0.3145,
      "step": 258
    },
    {
      "epoch": 8.97,
      "eval_loss": 3.52162766456604,
      "eval_runtime": 14.1996,
      "eval_samples_per_second": 23.804,
      "eval_steps_per_second": 3.028,
      "step": 258
    },
    {
      "epoch": 9.04,
      "grad_norm": 0.2942521274089813,
      "learning_rate": 0.001,
      "loss": 0.2462,
      "step": 260
    },
    {
      "epoch": 9.11,
      "grad_norm": 0.2955419719219208,
      "learning_rate": 0.001,
      "loss": 0.205,
      "step": 262
    },
    {
      "epoch": 9.18,
      "grad_norm": 0.29875773191452026,
      "learning_rate": 0.001,
      "loss": 0.2185,
      "step": 264
    },
    {
      "epoch": 9.25,
      "grad_norm": 0.9682367444038391,
      "learning_rate": 0.001,
      "loss": 0.5828,
      "step": 266
    },
    {
      "epoch": 9.32,
      "grad_norm": 0.32209742069244385,
      "learning_rate": 0.001,
      "loss": 0.2255,
      "step": 268
    },
    {
      "epoch": 9.39,
      "grad_norm": 0.2810392379760742,
      "learning_rate": 0.001,
      "loss": 0.2157,
      "step": 270
    },
    {
      "epoch": 9.46,
      "grad_norm": 0.26448503136634827,
      "learning_rate": 0.001,
      "loss": 0.2136,
      "step": 272
    },
    {
      "epoch": 9.53,
      "grad_norm": 0.2602413296699524,
      "learning_rate": 0.001,
      "loss": 0.2264,
      "step": 274
    },
    {
      "epoch": 9.6,
      "grad_norm": 0.30358317494392395,
      "learning_rate": 0.001,
      "loss": 0.2349,
      "step": 276
    },
    {
      "epoch": 9.67,
      "grad_norm": 0.2755320966243744,
      "learning_rate": 0.001,
      "loss": 0.2346,
      "step": 278
    },
    {
      "epoch": 9.74,
      "grad_norm": 0.257522314786911,
      "learning_rate": 0.001,
      "loss": 0.2312,
      "step": 280
    },
    {
      "epoch": 9.81,
      "grad_norm": 0.3185376524925232,
      "learning_rate": 0.001,
      "loss": 0.2519,
      "step": 282
    },
    {
      "epoch": 9.88,
      "grad_norm": 0.352522611618042,
      "learning_rate": 0.001,
      "loss": 0.2678,
      "step": 284
    },
    {
      "epoch": 9.95,
      "grad_norm": 0.3061072826385498,
      "learning_rate": 0.001,
      "loss": 0.2573,
      "step": 286
    },
    {
      "epoch": 9.98,
      "eval_loss": 4.012668609619141,
      "eval_runtime": 14.2083,
      "eval_samples_per_second": 23.789,
      "eval_steps_per_second": 3.026,
      "step": 287
    },
    {
      "epoch": 10.02,
      "grad_norm": 0.34469902515411377,
      "learning_rate": 0.001,
      "loss": 0.2394,
      "step": 288
    },
    {
      "epoch": 10.09,
      "grad_norm": 0.26758918166160583,
      "learning_rate": 0.001,
      "loss": 0.1749,
      "step": 290
    },
    {
      "epoch": 10.16,
      "grad_norm": 0.24024738371372223,
      "learning_rate": 0.001,
      "loss": 0.1687,
      "step": 292
    },
    {
      "epoch": 10.23,
      "grad_norm": 0.2440188080072403,
      "learning_rate": 0.001,
      "loss": 0.1634,
      "step": 294
    },
    {
      "epoch": 10.3,
      "grad_norm": 0.2564130425453186,
      "learning_rate": 0.001,
      "loss": 0.1759,
      "step": 296
    },
    {
      "epoch": 10.37,
      "grad_norm": 0.27317214012145996,
      "learning_rate": 0.001,
      "loss": 0.19,
      "step": 298
    },
    {
      "epoch": 10.43,
      "grad_norm": 0.3196158707141876,
      "learning_rate": 0.001,
      "loss": 0.1903,
      "step": 300
    },
    {
      "epoch": 10.5,
      "grad_norm": 0.3399882912635803,
      "learning_rate": 0.001,
      "loss": 0.2016,
      "step": 302
    },
    {
      "epoch": 10.57,
      "grad_norm": 0.28635892271995544,
      "learning_rate": 0.001,
      "loss": 0.189,
      "step": 304
    },
    {
      "epoch": 10.64,
      "grad_norm": 0.2602458596229553,
      "learning_rate": 0.001,
      "loss": 0.1792,
      "step": 306
    },
    {
      "epoch": 10.71,
      "grad_norm": 0.29866456985473633,
      "learning_rate": 0.001,
      "loss": 0.1974,
      "step": 308
    },
    {
      "epoch": 10.78,
      "grad_norm": 0.2757103443145752,
      "learning_rate": 0.001,
      "loss": 0.2025,
      "step": 310
    },
    {
      "epoch": 10.85,
      "grad_norm": 0.2643924355506897,
      "learning_rate": 0.001,
      "loss": 0.2106,
      "step": 312
    },
    {
      "epoch": 10.92,
      "grad_norm": 0.27770230174064636,
      "learning_rate": 0.001,
      "loss": 0.1999,
      "step": 314
    },
    {
      "epoch": 10.99,
      "grad_norm": 0.2809506356716156,
      "learning_rate": 0.001,
      "loss": 0.213,
      "step": 316
    },
    {
      "epoch": 10.99,
      "eval_loss": 3.9713540077209473,
      "eval_runtime": 14.2091,
      "eval_samples_per_second": 23.788,
      "eval_steps_per_second": 3.026,
      "step": 316
    },
    {
      "epoch": 11.06,
      "grad_norm": 0.19931960105895996,
      "learning_rate": 0.001,
      "loss": 0.1467,
      "step": 318
    },
    {
      "epoch": 11.13,
      "grad_norm": 0.20454713702201843,
      "learning_rate": 0.001,
      "loss": 0.1357,
      "step": 320
    },
    {
      "epoch": 11.2,
      "grad_norm": 0.22833150625228882,
      "learning_rate": 0.001,
      "loss": 0.1357,
      "step": 322
    },
    {
      "epoch": 11.27,
      "grad_norm": 0.45052453875541687,
      "learning_rate": 0.001,
      "loss": 0.1401,
      "step": 324
    },
    {
      "epoch": 11.34,
      "grad_norm": 0.2621672451496124,
      "learning_rate": 0.001,
      "loss": 0.1402,
      "step": 326
    },
    {
      "epoch": 11.41,
      "grad_norm": 0.2570265829563141,
      "learning_rate": 0.001,
      "loss": 0.1743,
      "step": 328
    },
    {
      "epoch": 11.48,
      "grad_norm": 0.2333604246377945,
      "learning_rate": 0.001,
      "loss": 0.1572,
      "step": 330
    },
    {
      "epoch": 11.55,
      "grad_norm": 0.2325439304113388,
      "learning_rate": 0.001,
      "loss": 0.1505,
      "step": 332
    },
    {
      "epoch": 11.62,
      "grad_norm": 0.2524455487728119,
      "learning_rate": 0.001,
      "loss": 0.1574,
      "step": 334
    },
    {
      "epoch": 11.69,
      "grad_norm": 0.2821509838104248,
      "learning_rate": 0.001,
      "loss": 0.162,
      "step": 336
    },
    {
      "epoch": 11.76,
      "grad_norm": 0.2627921402454376,
      "learning_rate": 0.001,
      "loss": 0.1823,
      "step": 338
    },
    {
      "epoch": 11.83,
      "grad_norm": 0.20858174562454224,
      "learning_rate": 0.001,
      "loss": 0.1581,
      "step": 340
    },
    {
      "epoch": 11.9,
      "grad_norm": 0.24444954097270966,
      "learning_rate": 0.001,
      "loss": 0.1528,
      "step": 342
    },
    {
      "epoch": 11.97,
      "grad_norm": 0.23985397815704346,
      "learning_rate": 0.001,
      "loss": 0.1682,
      "step": 344
    },
    {
      "epoch": 12.0,
      "eval_loss": 3.842698097229004,
      "eval_runtime": 14.2204,
      "eval_samples_per_second": 23.769,
      "eval_steps_per_second": 3.024,
      "step": 345
    },
    {
      "epoch": 12.03,
      "grad_norm": 0.20599780976772308,
      "learning_rate": 0.001,
      "loss": 0.1472,
      "step": 346
    },
    {
      "epoch": 12.1,
      "grad_norm": 0.20421436429023743,
      "learning_rate": 0.001,
      "loss": 0.1136,
      "step": 348
    },
    {
      "epoch": 12.17,
      "grad_norm": 0.21933647990226746,
      "learning_rate": 0.001,
      "loss": 0.1137,
      "step": 350
    },
    {
      "epoch": 12.24,
      "grad_norm": 0.2488287091255188,
      "learning_rate": 0.001,
      "loss": 0.1223,
      "step": 352
    },
    {
      "epoch": 12.31,
      "grad_norm": 0.21850407123565674,
      "learning_rate": 0.001,
      "loss": 0.1266,
      "step": 354
    },
    {
      "epoch": 12.38,
      "grad_norm": 0.219721257686615,
      "learning_rate": 0.001,
      "loss": 0.1266,
      "step": 356
    },
    {
      "epoch": 12.45,
      "grad_norm": 0.44860273599624634,
      "learning_rate": 0.001,
      "loss": 0.1271,
      "step": 358
    },
    {
      "epoch": 12.52,
      "grad_norm": 0.20995499193668365,
      "learning_rate": 0.001,
      "loss": 0.1201,
      "step": 360
    },
    {
      "epoch": 12.59,
      "grad_norm": 0.2180211991071701,
      "learning_rate": 0.001,
      "loss": 0.1325,
      "step": 362
    },
    {
      "epoch": 12.66,
      "grad_norm": 0.23486308753490448,
      "learning_rate": 0.001,
      "loss": 0.1316,
      "step": 364
    },
    {
      "epoch": 12.73,
      "grad_norm": 0.26361599564552307,
      "learning_rate": 0.001,
      "loss": 0.1417,
      "step": 366
    },
    {
      "epoch": 12.8,
      "grad_norm": 0.27585896849632263,
      "learning_rate": 0.001,
      "loss": 0.1514,
      "step": 368
    },
    {
      "epoch": 12.87,
      "grad_norm": 0.23249991238117218,
      "learning_rate": 0.001,
      "loss": 0.1387,
      "step": 370
    },
    {
      "epoch": 12.94,
      "grad_norm": 0.2378188967704773,
      "learning_rate": 0.001,
      "loss": 0.1396,
      "step": 372
    },
    {
      "epoch": 12.97,
      "eval_loss": 4.202455520629883,
      "eval_runtime": 14.2176,
      "eval_samples_per_second": 23.773,
      "eval_steps_per_second": 3.024,
      "step": 373
    },
    {
      "epoch": 13.01,
      "grad_norm": 0.25618231296539307,
      "learning_rate": 0.001,
      "loss": 0.1403,
      "step": 374
    },
    {
      "epoch": 13.08,
      "grad_norm": 0.19359038770198822,
      "learning_rate": 0.001,
      "loss": 0.1048,
      "step": 376
    },
    {
      "epoch": 13.15,
      "grad_norm": 0.20456379652023315,
      "learning_rate": 0.001,
      "loss": 0.102,
      "step": 378
    },
    {
      "epoch": 13.22,
      "grad_norm": 0.20489871501922607,
      "learning_rate": 0.001,
      "loss": 0.1039,
      "step": 380
    },
    {
      "epoch": 13.29,
      "grad_norm": 0.20058828592300415,
      "learning_rate": 0.001,
      "loss": 0.1066,
      "step": 382
    },
    {
      "epoch": 13.36,
      "grad_norm": 0.20939727127552032,
      "learning_rate": 0.001,
      "loss": 0.1017,
      "step": 384
    },
    {
      "epoch": 13.43,
      "grad_norm": 0.20905180275440216,
      "learning_rate": 0.001,
      "loss": 0.1074,
      "step": 386
    },
    {
      "epoch": 13.5,
      "grad_norm": 0.2082565426826477,
      "learning_rate": 0.001,
      "loss": 0.1122,
      "step": 388
    },
    {
      "epoch": 13.57,
      "grad_norm": 0.20671486854553223,
      "learning_rate": 0.001,
      "loss": 0.1137,
      "step": 390
    },
    {
      "epoch": 13.63,
      "grad_norm": 0.20714151859283447,
      "learning_rate": 0.001,
      "loss": 0.1085,
      "step": 392
    },
    {
      "epoch": 13.7,
      "grad_norm": 0.21117734909057617,
      "learning_rate": 0.001,
      "loss": 0.1225,
      "step": 394
    },
    {
      "epoch": 13.77,
      "grad_norm": 0.20374175906181335,
      "learning_rate": 0.001,
      "loss": 0.1135,
      "step": 396
    },
    {
      "epoch": 13.84,
      "grad_norm": 0.2223602831363678,
      "learning_rate": 0.001,
      "loss": 0.1188,
      "step": 398
    },
    {
      "epoch": 13.91,
      "grad_norm": 0.2405446618795395,
      "learning_rate": 0.001,
      "loss": 0.1329,
      "step": 400
    },
    {
      "epoch": 13.98,
      "grad_norm": 0.25538983941078186,
      "learning_rate": 0.001,
      "loss": 0.1363,
      "step": 402
    },
    {
      "epoch": 13.98,
      "eval_loss": 4.401235103607178,
      "eval_runtime": 14.2163,
      "eval_samples_per_second": 23.776,
      "eval_steps_per_second": 3.025,
      "step": 402
    },
    {
      "epoch": 14.05,
      "grad_norm": 0.20007029175758362,
      "learning_rate": 0.001,
      "loss": 0.0995,
      "step": 404
    },
    {
      "epoch": 14.12,
      "grad_norm": 0.17897407710552216,
      "learning_rate": 0.001,
      "loss": 0.0927,
      "step": 406
    },
    {
      "epoch": 14.19,
      "grad_norm": 0.18746937811374664,
      "learning_rate": 0.001,
      "loss": 0.0898,
      "step": 408
    },
    {
      "epoch": 14.26,
      "grad_norm": 0.1855156570672989,
      "learning_rate": 0.001,
      "loss": 0.0913,
      "step": 410
    },
    {
      "epoch": 14.33,
      "grad_norm": 0.1712156981229782,
      "learning_rate": 0.001,
      "loss": 0.0891,
      "step": 412
    },
    {
      "epoch": 14.4,
      "grad_norm": 0.193171888589859,
      "learning_rate": 0.001,
      "loss": 0.0871,
      "step": 414
    },
    {
      "epoch": 14.47,
      "grad_norm": 0.19764076173305511,
      "learning_rate": 0.001,
      "loss": 0.0892,
      "step": 416
    },
    {
      "epoch": 14.54,
      "grad_norm": 0.20630787312984467,
      "learning_rate": 0.001,
      "loss": 0.0984,
      "step": 418
    },
    {
      "epoch": 14.61,
      "grad_norm": 0.18784281611442566,
      "learning_rate": 0.001,
      "loss": 0.1002,
      "step": 420
    },
    {
      "epoch": 14.68,
      "grad_norm": 0.2134091705083847,
      "learning_rate": 0.001,
      "loss": 0.0994,
      "step": 422
    },
    {
      "epoch": 14.75,
      "grad_norm": 0.20559023320674896,
      "learning_rate": 0.001,
      "loss": 0.1036,
      "step": 424
    },
    {
      "epoch": 14.82,
      "grad_norm": 0.22063368558883667,
      "learning_rate": 0.001,
      "loss": 0.1008,
      "step": 426
    },
    {
      "epoch": 14.89,
      "grad_norm": 0.21584555506706238,
      "learning_rate": 0.001,
      "loss": 0.1137,
      "step": 428
    },
    {
      "epoch": 14.96,
      "grad_norm": 0.20185333490371704,
      "learning_rate": 0.001,
      "loss": 0.1148,
      "step": 430
    },
    {
      "epoch": 14.99,
      "eval_loss": 4.717379093170166,
      "eval_runtime": 14.2112,
      "eval_samples_per_second": 23.784,
      "eval_steps_per_second": 3.026,
      "step": 431
    },
    {
      "epoch": 15.03,
      "grad_norm": 0.22225050628185272,
      "learning_rate": 0.001,
      "loss": 0.0942,
      "step": 432
    },
    {
      "epoch": 15.1,
      "grad_norm": 0.2037811130285263,
      "learning_rate": 0.001,
      "loss": 0.0776,
      "step": 434
    },
    {
      "epoch": 15.17,
      "grad_norm": 0.1820273995399475,
      "learning_rate": 0.001,
      "loss": 0.0849,
      "step": 436
    },
    {
      "epoch": 15.23,
      "grad_norm": 0.18018700182437897,
      "learning_rate": 0.001,
      "loss": 0.0744,
      "step": 438
    },
    {
      "epoch": 15.3,
      "grad_norm": 0.21271073818206787,
      "learning_rate": 0.001,
      "loss": 0.0836,
      "step": 440
    },
    {
      "epoch": 15.37,
      "grad_norm": 0.19622232019901276,
      "learning_rate": 0.001,
      "loss": 0.0841,
      "step": 442
    },
    {
      "epoch": 15.44,
      "grad_norm": 0.19101227819919586,
      "learning_rate": 0.001,
      "loss": 0.085,
      "step": 444
    },
    {
      "epoch": 15.51,
      "grad_norm": 0.18909965455532074,
      "learning_rate": 0.001,
      "loss": 0.093,
      "step": 446
    },
    {
      "epoch": 15.58,
      "grad_norm": 0.2003660649061203,
      "learning_rate": 0.001,
      "loss": 0.0877,
      "step": 448
    },
    {
      "epoch": 15.65,
      "grad_norm": 0.2101544439792633,
      "learning_rate": 0.001,
      "loss": 0.0908,
      "step": 450
    },
    {
      "epoch": 15.72,
      "grad_norm": 0.20344135165214539,
      "learning_rate": 0.001,
      "loss": 0.0943,
      "step": 452
    },
    {
      "epoch": 15.79,
      "grad_norm": 0.28948044776916504,
      "learning_rate": 0.001,
      "loss": 0.1094,
      "step": 454
    },
    {
      "epoch": 15.86,
      "grad_norm": 0.29758819937705994,
      "learning_rate": 0.001,
      "loss": 0.1247,
      "step": 456
    },
    {
      "epoch": 15.93,
      "grad_norm": 0.24994027614593506,
      "learning_rate": 0.001,
      "loss": 0.1106,
      "step": 458
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.2185268998146057,
      "learning_rate": 0.001,
      "loss": 0.0907,
      "step": 460
    },
    {
      "epoch": 16.0,
      "eval_loss": 4.497994899749756,
      "eval_runtime": 14.2114,
      "eval_samples_per_second": 23.784,
      "eval_steps_per_second": 3.026,
      "step": 460
    },
    {
      "epoch": 16.07,
      "grad_norm": 0.1781373769044876,
      "learning_rate": 0.001,
      "loss": 0.0698,
      "step": 462
    },
    {
      "epoch": 16.14,
      "grad_norm": 0.20170117914676666,
      "learning_rate": 0.001,
      "loss": 0.0763,
      "step": 464
    },
    {
      "epoch": 16.21,
      "grad_norm": 0.20545266568660736,
      "learning_rate": 0.001,
      "loss": 0.0858,
      "step": 466
    },
    {
      "epoch": 16.28,
      "grad_norm": 0.18228977918624878,
      "learning_rate": 0.001,
      "loss": 0.0802,
      "step": 468
    },
    {
      "epoch": 16.35,
      "grad_norm": 0.17187969386577606,
      "learning_rate": 0.001,
      "loss": 0.0803,
      "step": 470
    },
    {
      "epoch": 16.42,
      "grad_norm": 0.17641624808311462,
      "learning_rate": 0.001,
      "loss": 0.0796,
      "step": 472
    },
    {
      "epoch": 16.49,
      "grad_norm": 0.19069235026836395,
      "learning_rate": 0.001,
      "loss": 0.0798,
      "step": 474
    },
    {
      "epoch": 16.56,
      "grad_norm": 0.1968899667263031,
      "learning_rate": 0.001,
      "loss": 0.081,
      "step": 476
    },
    {
      "epoch": 16.63,
      "grad_norm": 0.196567103266716,
      "learning_rate": 0.001,
      "loss": 0.0891,
      "step": 478
    },
    {
      "epoch": 16.7,
      "grad_norm": 0.18010011315345764,
      "learning_rate": 0.001,
      "loss": 0.0795,
      "step": 480
    },
    {
      "epoch": 16.77,
      "grad_norm": 0.31678447127342224,
      "learning_rate": 0.001,
      "loss": 0.0819,
      "step": 482
    },
    {
      "epoch": 16.83,
      "grad_norm": 0.20116592943668365,
      "learning_rate": 0.001,
      "loss": 0.0873,
      "step": 484
    },
    {
      "epoch": 16.9,
      "grad_norm": 0.19425255060195923,
      "learning_rate": 0.001,
      "loss": 0.0898,
      "step": 486
    },
    {
      "epoch": 16.97,
      "grad_norm": 0.21240267157554626,
      "learning_rate": 0.001,
      "loss": 0.0942,
      "step": 488
    },
    {
      "epoch": 16.97,
      "eval_loss": 4.70243501663208,
      "eval_runtime": 14.2157,
      "eval_samples_per_second": 23.777,
      "eval_steps_per_second": 3.025,
      "step": 488
    },
    {
      "epoch": 17.04,
      "grad_norm": 0.16432495415210724,
      "learning_rate": 0.001,
      "loss": 0.0762,
      "step": 490
    },
    {
      "epoch": 17.11,
      "grad_norm": 0.19645223021507263,
      "learning_rate": 0.001,
      "loss": 0.0638,
      "step": 492
    },
    {
      "epoch": 17.18,
      "grad_norm": 0.1714513599872589,
      "learning_rate": 0.001,
      "loss": 0.0669,
      "step": 494
    },
    {
      "epoch": 17.25,
      "grad_norm": 0.16747209429740906,
      "learning_rate": 0.001,
      "loss": 0.064,
      "step": 496
    },
    {
      "epoch": 17.32,
      "grad_norm": 0.17507390677928925,
      "learning_rate": 0.001,
      "loss": 0.0675,
      "step": 498
    },
    {
      "epoch": 17.39,
      "grad_norm": 0.1752365380525589,
      "learning_rate": 0.001,
      "loss": 0.0716,
      "step": 500
    },
    {
      "epoch": 17.46,
      "grad_norm": 0.1760423481464386,
      "learning_rate": 0.001,
      "loss": 0.0736,
      "step": 502
    },
    {
      "epoch": 17.53,
      "grad_norm": 0.17969931662082672,
      "learning_rate": 0.001,
      "loss": 0.0725,
      "step": 504
    },
    {
      "epoch": 17.6,
      "grad_norm": 0.20572717487812042,
      "learning_rate": 0.001,
      "loss": 0.0751,
      "step": 506
    },
    {
      "epoch": 17.67,
      "grad_norm": 0.20842702686786652,
      "learning_rate": 0.001,
      "loss": 0.0796,
      "step": 508
    },
    {
      "epoch": 17.74,
      "grad_norm": 0.18587857484817505,
      "learning_rate": 0.001,
      "loss": 0.0757,
      "step": 510
    },
    {
      "epoch": 17.81,
      "grad_norm": 0.17455922067165375,
      "learning_rate": 0.001,
      "loss": 0.0792,
      "step": 512
    },
    {
      "epoch": 17.88,
      "grad_norm": 0.181496262550354,
      "learning_rate": 0.001,
      "loss": 0.0711,
      "step": 514
    },
    {
      "epoch": 17.95,
      "grad_norm": 0.1918957382440567,
      "learning_rate": 0.001,
      "loss": 0.0765,
      "step": 516
    },
    {
      "epoch": 17.98,
      "eval_loss": 4.348242282867432,
      "eval_runtime": 14.2133,
      "eval_samples_per_second": 23.78,
      "eval_steps_per_second": 3.025,
      "step": 517
    },
    {
      "epoch": 18.02,
      "grad_norm": 0.23215439915657043,
      "learning_rate": 0.001,
      "loss": 0.0854,
      "step": 518
    },
    {
      "epoch": 18.09,
      "grad_norm": 0.16362574696540833,
      "learning_rate": 0.001,
      "loss": 0.0636,
      "step": 520
    },
    {
      "epoch": 18.16,
      "grad_norm": 0.18604953587055206,
      "learning_rate": 0.001,
      "loss": 0.0632,
      "step": 522
    },
    {
      "epoch": 18.23,
      "grad_norm": 0.20401595532894135,
      "learning_rate": 0.001,
      "loss": 0.077,
      "step": 524
    },
    {
      "epoch": 18.3,
      "grad_norm": 0.1713324785232544,
      "learning_rate": 0.001,
      "loss": 0.0743,
      "step": 526
    },
    {
      "epoch": 18.37,
      "grad_norm": 0.1794094294309616,
      "learning_rate": 0.001,
      "loss": 0.0693,
      "step": 528
    },
    {
      "epoch": 18.43,
      "grad_norm": 0.1892070323228836,
      "learning_rate": 0.001,
      "loss": 0.0663,
      "step": 530
    },
    {
      "epoch": 18.5,
      "grad_norm": 0.1843532919883728,
      "learning_rate": 0.001,
      "loss": 0.0721,
      "step": 532
    },
    {
      "epoch": 18.57,
      "grad_norm": 0.17645902931690216,
      "learning_rate": 0.001,
      "loss": 0.0746,
      "step": 534
    },
    {
      "epoch": 18.64,
      "grad_norm": 0.16752366721630096,
      "learning_rate": 0.001,
      "loss": 0.0712,
      "step": 536
    },
    {
      "epoch": 18.71,
      "grad_norm": 0.16078142821788788,
      "learning_rate": 0.001,
      "loss": 0.0729,
      "step": 538
    },
    {
      "epoch": 18.78,
      "grad_norm": 0.1775059700012207,
      "learning_rate": 0.001,
      "loss": 0.0745,
      "step": 540
    },
    {
      "epoch": 18.85,
      "grad_norm": 0.18710005283355713,
      "learning_rate": 0.001,
      "loss": 0.0743,
      "step": 542
    },
    {
      "epoch": 18.92,
      "grad_norm": 0.20217396318912506,
      "learning_rate": 0.001,
      "loss": 0.0761,
      "step": 544
    },
    {
      "epoch": 18.99,
      "grad_norm": 0.19417209923267365,
      "learning_rate": 0.001,
      "loss": 0.0799,
      "step": 546
    },
    {
      "epoch": 18.99,
      "eval_loss": 4.538631916046143,
      "eval_runtime": 14.2104,
      "eval_samples_per_second": 23.785,
      "eval_steps_per_second": 3.026,
      "step": 546
    },
    {
      "epoch": 19.06,
      "grad_norm": 0.16049686074256897,
      "learning_rate": 0.001,
      "loss": 0.0633,
      "step": 548
    },
    {
      "epoch": 19.13,
      "grad_norm": 0.15315844118595123,
      "learning_rate": 0.001,
      "loss": 0.0614,
      "step": 550
    },
    {
      "epoch": 19.2,
      "grad_norm": 0.16441303491592407,
      "learning_rate": 0.001,
      "loss": 0.0658,
      "step": 552
    },
    {
      "epoch": 19.27,
      "grad_norm": 0.16347071528434753,
      "learning_rate": 0.001,
      "loss": 0.0551,
      "step": 554
    },
    {
      "epoch": 19.34,
      "grad_norm": 0.16774040460586548,
      "learning_rate": 0.001,
      "loss": 0.0698,
      "step": 556
    },
    {
      "epoch": 19.41,
      "grad_norm": 0.15695886313915253,
      "learning_rate": 0.001,
      "loss": 0.063,
      "step": 558
    },
    {
      "epoch": 19.48,
      "grad_norm": 0.1639881581068039,
      "learning_rate": 0.001,
      "loss": 0.0649,
      "step": 560
    },
    {
      "epoch": 19.55,
      "grad_norm": 0.1653515100479126,
      "learning_rate": 0.001,
      "loss": 0.0592,
      "step": 562
    },
    {
      "epoch": 19.62,
      "grad_norm": 0.1697942465543747,
      "learning_rate": 0.001,
      "loss": 0.0619,
      "step": 564
    },
    {
      "epoch": 19.69,
      "grad_norm": 0.18345074355602264,
      "learning_rate": 0.001,
      "loss": 0.0674,
      "step": 566
    },
    {
      "epoch": 19.76,
      "grad_norm": 0.17060339450836182,
      "learning_rate": 0.001,
      "loss": 0.0603,
      "step": 568
    },
    {
      "epoch": 19.83,
      "grad_norm": 0.18731752038002014,
      "learning_rate": 0.001,
      "loss": 0.0668,
      "step": 570
    },
    {
      "epoch": 19.9,
      "grad_norm": 0.1727019101381302,
      "learning_rate": 0.001,
      "loss": 0.0717,
      "step": 572
    },
    {
      "epoch": 19.97,
      "grad_norm": 0.18209876120090485,
      "learning_rate": 0.001,
      "loss": 0.073,
      "step": 574
    },
    {
      "epoch": 20.0,
      "eval_loss": 4.588906764984131,
      "eval_runtime": 14.2111,
      "eval_samples_per_second": 23.784,
      "eval_steps_per_second": 3.026,
      "step": 575
    },
    {
      "epoch": 20.03,
      "grad_norm": 0.15836042165756226,
      "learning_rate": 0.001,
      "loss": 0.0655,
      "step": 576
    },
    {
      "epoch": 20.1,
      "grad_norm": 0.1645493507385254,
      "learning_rate": 0.001,
      "loss": 0.0609,
      "step": 578
    },
    {
      "epoch": 20.17,
      "grad_norm": 0.14691436290740967,
      "learning_rate": 0.001,
      "loss": 0.0533,
      "step": 580
    },
    {
      "epoch": 20.24,
      "grad_norm": 0.15472760796546936,
      "learning_rate": 0.001,
      "loss": 0.0596,
      "step": 582
    },
    {
      "epoch": 20.31,
      "grad_norm": 0.14597873389720917,
      "learning_rate": 0.001,
      "loss": 0.0548,
      "step": 584
    },
    {
      "epoch": 20.38,
      "grad_norm": 0.14974112808704376,
      "learning_rate": 0.001,
      "loss": 0.057,
      "step": 586
    },
    {
      "epoch": 20.45,
      "grad_norm": 0.16359882056713104,
      "learning_rate": 0.001,
      "loss": 0.0597,
      "step": 588
    },
    {
      "epoch": 20.52,
      "grad_norm": 0.15623408555984497,
      "learning_rate": 0.001,
      "loss": 0.0581,
      "step": 590
    },
    {
      "epoch": 20.59,
      "grad_norm": 0.16072715818881989,
      "learning_rate": 0.001,
      "loss": 0.0577,
      "step": 592
    },
    {
      "epoch": 20.66,
      "grad_norm": 0.17041383683681488,
      "learning_rate": 0.001,
      "loss": 0.0601,
      "step": 594
    },
    {
      "epoch": 20.73,
      "grad_norm": 0.182576522231102,
      "learning_rate": 0.001,
      "loss": 0.0614,
      "step": 596
    },
    {
      "epoch": 20.8,
      "grad_norm": 0.17546887695789337,
      "learning_rate": 0.001,
      "loss": 0.0658,
      "step": 598
    },
    {
      "epoch": 20.87,
      "grad_norm": 0.19615040719509125,
      "learning_rate": 0.001,
      "loss": 0.061,
      "step": 600
    },
    {
      "epoch": 20.94,
      "grad_norm": 0.2895524799823761,
      "learning_rate": 0.001,
      "loss": 0.0825,
      "step": 602
    },
    {
      "epoch": 20.97,
      "eval_loss": 4.681673049926758,
      "eval_runtime": 14.2047,
      "eval_samples_per_second": 23.795,
      "eval_steps_per_second": 3.027,
      "step": 603
    },
    {
      "epoch": 21.01,
      "grad_norm": 0.3042687177658081,
      "learning_rate": 0.001,
      "loss": 0.102,
      "step": 604
    },
    {
      "epoch": 21.08,
      "grad_norm": 0.14991188049316406,
      "learning_rate": 0.001,
      "loss": 0.0585,
      "step": 606
    },
    {
      "epoch": 21.15,
      "grad_norm": 0.13689693808555603,
      "learning_rate": 0.001,
      "loss": 0.0546,
      "step": 608
    },
    {
      "epoch": 21.22,
      "grad_norm": 0.14480513334274292,
      "learning_rate": 0.001,
      "loss": 0.0533,
      "step": 610
    },
    {
      "epoch": 21.29,
      "grad_norm": 0.148710235953331,
      "learning_rate": 0.001,
      "loss": 0.0543,
      "step": 612
    },
    {
      "epoch": 21.36,
      "grad_norm": 0.14531894028186798,
      "learning_rate": 0.001,
      "loss": 0.0528,
      "step": 614
    },
    {
      "epoch": 21.43,
      "grad_norm": 0.14758220314979553,
      "learning_rate": 0.001,
      "loss": 0.0536,
      "step": 616
    },
    {
      "epoch": 21.5,
      "grad_norm": 0.1422356367111206,
      "learning_rate": 0.001,
      "loss": 0.0537,
      "step": 618
    },
    {
      "epoch": 21.57,
      "grad_norm": 0.16927896440029144,
      "learning_rate": 0.001,
      "loss": 0.0551,
      "step": 620
    },
    {
      "epoch": 21.63,
      "grad_norm": 0.18368251621723175,
      "learning_rate": 0.001,
      "loss": 0.0589,
      "step": 622
    },
    {
      "epoch": 21.7,
      "grad_norm": 0.16162264347076416,
      "learning_rate": 0.001,
      "loss": 0.0639,
      "step": 624
    },
    {
      "epoch": 21.77,
      "grad_norm": 0.1610013097524643,
      "learning_rate": 0.001,
      "loss": 0.0604,
      "step": 626
    },
    {
      "epoch": 21.84,
      "grad_norm": 0.15656866133213043,
      "learning_rate": 0.001,
      "loss": 0.0603,
      "step": 628
    },
    {
      "epoch": 21.91,
      "grad_norm": 0.15664397180080414,
      "learning_rate": 0.001,
      "loss": 0.0589,
      "step": 630
    },
    {
      "epoch": 21.98,
      "grad_norm": 0.1484573483467102,
      "learning_rate": 0.001,
      "loss": 0.0616,
      "step": 632
    },
    {
      "epoch": 21.98,
      "eval_loss": 5.026252746582031,
      "eval_runtime": 14.2151,
      "eval_samples_per_second": 23.778,
      "eval_steps_per_second": 3.025,
      "step": 632
    },
    {
      "epoch": 22.05,
      "grad_norm": 0.16003428399562836,
      "learning_rate": 0.001,
      "loss": 0.0579,
      "step": 634
    },
    {
      "epoch": 22.12,
      "grad_norm": 0.14746786653995514,
      "learning_rate": 0.001,
      "loss": 0.0503,
      "step": 636
    },
    {
      "epoch": 22.19,
      "grad_norm": 0.1616966724395752,
      "learning_rate": 0.001,
      "loss": 0.0518,
      "step": 638
    },
    {
      "epoch": 22.26,
      "grad_norm": 0.14375324547290802,
      "learning_rate": 0.001,
      "loss": 0.0501,
      "step": 640
    },
    {
      "epoch": 22.33,
      "grad_norm": 0.14273105561733246,
      "learning_rate": 0.001,
      "loss": 0.0543,
      "step": 642
    },
    {
      "epoch": 22.4,
      "grad_norm": 0.14281895756721497,
      "learning_rate": 0.001,
      "loss": 0.0572,
      "step": 644
    },
    {
      "epoch": 22.47,
      "grad_norm": 0.1508861482143402,
      "learning_rate": 0.001,
      "loss": 0.0538,
      "step": 646
    },
    {
      "epoch": 22.54,
      "grad_norm": 0.1677195429801941,
      "learning_rate": 0.001,
      "loss": 0.052,
      "step": 648
    },
    {
      "epoch": 22.61,
      "grad_norm": 0.16616381704807281,
      "learning_rate": 0.001,
      "loss": 0.0569,
      "step": 650
    },
    {
      "epoch": 22.68,
      "grad_norm": 0.159612774848938,
      "learning_rate": 0.001,
      "loss": 0.0585,
      "step": 652
    },
    {
      "epoch": 22.75,
      "grad_norm": 0.14692574739456177,
      "learning_rate": 0.001,
      "loss": 0.0559,
      "step": 654
    },
    {
      "epoch": 22.82,
      "grad_norm": 0.14404280483722687,
      "learning_rate": 0.001,
      "loss": 0.0614,
      "step": 656
    },
    {
      "epoch": 22.89,
      "grad_norm": 0.16935443878173828,
      "learning_rate": 0.001,
      "loss": 0.0531,
      "step": 658
    },
    {
      "epoch": 22.96,
      "grad_norm": 0.16444338858127594,
      "learning_rate": 0.001,
      "loss": 0.0677,
      "step": 660
    },
    {
      "epoch": 22.99,
      "eval_loss": 4.580361843109131,
      "eval_runtime": 14.2193,
      "eval_samples_per_second": 23.77,
      "eval_steps_per_second": 3.024,
      "step": 661
    },
    {
      "epoch": 23.03,
      "grad_norm": 0.1523355394601822,
      "learning_rate": 0.001,
      "loss": 0.056,
      "step": 662
    },
    {
      "epoch": 23.1,
      "grad_norm": 0.13197599351406097,
      "learning_rate": 0.001,
      "loss": 0.0505,
      "step": 664
    },
    {
      "epoch": 23.17,
      "grad_norm": 0.1436968594789505,
      "learning_rate": 0.001,
      "loss": 0.0459,
      "step": 666
    },
    {
      "epoch": 23.23,
      "grad_norm": 0.15447324514389038,
      "learning_rate": 0.001,
      "loss": 0.0484,
      "step": 668
    },
    {
      "epoch": 23.3,
      "grad_norm": 0.1536988466978073,
      "learning_rate": 0.001,
      "loss": 0.0523,
      "step": 670
    },
    {
      "epoch": 23.37,
      "grad_norm": 0.14964550733566284,
      "learning_rate": 0.001,
      "loss": 0.0523,
      "step": 672
    },
    {
      "epoch": 23.44,
      "grad_norm": 0.14874936640262604,
      "learning_rate": 0.001,
      "loss": 0.053,
      "step": 674
    },
    {
      "epoch": 23.51,
      "grad_norm": 0.14731401205062866,
      "learning_rate": 0.001,
      "loss": 0.0534,
      "step": 676
    },
    {
      "epoch": 23.58,
      "grad_norm": 0.15508398413658142,
      "learning_rate": 0.001,
      "loss": 0.0538,
      "step": 678
    },
    {
      "epoch": 23.65,
      "grad_norm": 0.16380004584789276,
      "learning_rate": 0.001,
      "loss": 0.0539,
      "step": 680
    },
    {
      "epoch": 23.72,
      "grad_norm": 0.38461077213287354,
      "learning_rate": 0.001,
      "loss": 0.0609,
      "step": 682
    },
    {
      "epoch": 23.79,
      "grad_norm": 0.1609809696674347,
      "learning_rate": 0.001,
      "loss": 0.0553,
      "step": 684
    },
    {
      "epoch": 23.86,
      "grad_norm": 0.1482686847448349,
      "learning_rate": 0.001,
      "loss": 0.0576,
      "step": 686
    },
    {
      "epoch": 23.93,
      "grad_norm": 0.16069763898849487,
      "learning_rate": 0.001,
      "loss": 0.0569,
      "step": 688
    },
    {
      "epoch": 24.0,
      "grad_norm": 0.16612602770328522,
      "learning_rate": 0.001,
      "loss": 0.0571,
      "step": 690
    },
    {
      "epoch": 24.0,
      "eval_loss": 4.839931488037109,
      "eval_runtime": 14.2054,
      "eval_samples_per_second": 23.794,
      "eval_steps_per_second": 3.027,
      "step": 690
    },
    {
      "epoch": 24.07,
      "grad_norm": 0.14117293059825897,
      "learning_rate": 0.001,
      "loss": 0.0473,
      "step": 692
    },
    {
      "epoch": 24.14,
      "grad_norm": 0.13660480082035065,
      "learning_rate": 0.001,
      "loss": 0.0462,
      "step": 694
    },
    {
      "epoch": 24.21,
      "grad_norm": 0.13649675250053406,
      "learning_rate": 0.001,
      "loss": 0.0436,
      "step": 696
    },
    {
      "epoch": 24.28,
      "grad_norm": 0.13140219449996948,
      "learning_rate": 0.001,
      "loss": 0.0496,
      "step": 698
    },
    {
      "epoch": 24.35,
      "grad_norm": 0.13666340708732605,
      "learning_rate": 0.001,
      "loss": 0.0453,
      "step": 700
    },
    {
      "epoch": 24.42,
      "grad_norm": 0.14786505699157715,
      "learning_rate": 0.001,
      "loss": 0.0458,
      "step": 702
    },
    {
      "epoch": 24.49,
      "grad_norm": 0.14177779853343964,
      "learning_rate": 0.001,
      "loss": 0.0454,
      "step": 704
    },
    {
      "epoch": 24.56,
      "grad_norm": 0.15088017284870148,
      "learning_rate": 0.001,
      "loss": 0.0486,
      "step": 706
    },
    {
      "epoch": 24.63,
      "grad_norm": 0.1343182623386383,
      "learning_rate": 0.001,
      "loss": 0.0466,
      "step": 708
    },
    {
      "epoch": 24.7,
      "grad_norm": 0.14765577018260956,
      "learning_rate": 0.001,
      "loss": 0.051,
      "step": 710
    },
    {
      "epoch": 24.77,
      "grad_norm": 0.1491304486989975,
      "learning_rate": 0.001,
      "loss": 0.0497,
      "step": 712
    },
    {
      "epoch": 24.83,
      "grad_norm": 0.1482744961977005,
      "learning_rate": 0.001,
      "loss": 0.0525,
      "step": 714
    },
    {
      "epoch": 24.9,
      "grad_norm": 0.15700866281986237,
      "learning_rate": 0.001,
      "loss": 0.0497,
      "step": 716
    },
    {
      "epoch": 24.97,
      "grad_norm": 0.15650177001953125,
      "learning_rate": 0.001,
      "loss": 0.0525,
      "step": 718
    },
    {
      "epoch": 24.97,
      "eval_loss": 4.934982776641846,
      "eval_runtime": 14.2072,
      "eval_samples_per_second": 23.791,
      "eval_steps_per_second": 3.027,
      "step": 718
    },
    {
      "epoch": 25.04,
      "grad_norm": 0.15254813432693481,
      "learning_rate": 0.001,
      "loss": 0.0501,
      "step": 720
    },
    {
      "epoch": 25.11,
      "grad_norm": 0.1353287696838379,
      "learning_rate": 0.001,
      "loss": 0.0471,
      "step": 722
    },
    {
      "epoch": 25.18,
      "grad_norm": 0.155950665473938,
      "learning_rate": 0.001,
      "loss": 0.0478,
      "step": 724
    },
    {
      "epoch": 25.25,
      "grad_norm": 0.14119938015937805,
      "learning_rate": 0.001,
      "loss": 0.0484,
      "step": 726
    },
    {
      "epoch": 25.32,
      "grad_norm": 0.14679110050201416,
      "learning_rate": 0.001,
      "loss": 0.0432,
      "step": 728
    },
    {
      "epoch": 25.39,
      "grad_norm": 0.13392671942710876,
      "learning_rate": 0.001,
      "loss": 0.0486,
      "step": 730
    },
    {
      "epoch": 25.46,
      "grad_norm": 0.13733159005641937,
      "learning_rate": 0.001,
      "loss": 0.0451,
      "step": 732
    },
    {
      "epoch": 25.53,
      "grad_norm": 0.13778570294380188,
      "learning_rate": 0.001,
      "loss": 0.0452,
      "step": 734
    },
    {
      "epoch": 25.6,
      "grad_norm": 0.15224093198776245,
      "learning_rate": 0.001,
      "loss": 0.0469,
      "step": 736
    },
    {
      "epoch": 25.67,
      "grad_norm": 0.15857195854187012,
      "learning_rate": 0.001,
      "loss": 0.0489,
      "step": 738
    },
    {
      "epoch": 25.74,
      "grad_norm": 0.15234865248203278,
      "learning_rate": 0.001,
      "loss": 0.0536,
      "step": 740
    },
    {
      "epoch": 25.81,
      "grad_norm": 0.17310313880443573,
      "learning_rate": 0.001,
      "loss": 0.0518,
      "step": 742
    },
    {
      "epoch": 25.88,
      "grad_norm": 0.23332546651363373,
      "learning_rate": 0.001,
      "loss": 0.0601,
      "step": 744
    },
    {
      "epoch": 25.95,
      "grad_norm": 0.22406938672065735,
      "learning_rate": 0.001,
      "loss": 0.081,
      "step": 746
    },
    {
      "epoch": 25.98,
      "eval_loss": 4.690272331237793,
      "eval_runtime": 14.2141,
      "eval_samples_per_second": 23.779,
      "eval_steps_per_second": 3.025,
      "step": 747
    },
    {
      "epoch": 26.02,
      "grad_norm": 0.14596611261367798,
      "learning_rate": 0.001,
      "loss": 0.0551,
      "step": 748
    },
    {
      "epoch": 26.09,
      "grad_norm": 0.11857102066278458,
      "learning_rate": 0.001,
      "loss": 0.0398,
      "step": 750
    },
    {
      "epoch": 26.16,
      "grad_norm": 0.13823005557060242,
      "learning_rate": 0.001,
      "loss": 0.0424,
      "step": 752
    },
    {
      "epoch": 26.23,
      "grad_norm": 0.12586718797683716,
      "learning_rate": 0.001,
      "loss": 0.0422,
      "step": 754
    },
    {
      "epoch": 26.3,
      "grad_norm": 0.12402280420064926,
      "learning_rate": 0.001,
      "loss": 0.04,
      "step": 756
    },
    {
      "epoch": 26.37,
      "grad_norm": 0.1546175628900528,
      "learning_rate": 0.001,
      "loss": 0.0434,
      "step": 758
    },
    {
      "epoch": 26.43,
      "grad_norm": 0.15831632912158966,
      "learning_rate": 0.001,
      "loss": 0.041,
      "step": 760
    },
    {
      "epoch": 26.5,
      "grad_norm": 0.1380414068698883,
      "learning_rate": 0.001,
      "loss": 0.0411,
      "step": 762
    },
    {
      "epoch": 26.57,
      "grad_norm": 0.13040709495544434,
      "learning_rate": 0.001,
      "loss": 0.0412,
      "step": 764
    },
    {
      "epoch": 26.64,
      "grad_norm": 0.14285586774349213,
      "learning_rate": 0.001,
      "loss": 0.0426,
      "step": 766
    },
    {
      "epoch": 26.71,
      "grad_norm": 0.14807988703250885,
      "learning_rate": 0.001,
      "loss": 0.045,
      "step": 768
    },
    {
      "epoch": 26.78,
      "grad_norm": 0.14204862713813782,
      "learning_rate": 0.001,
      "loss": 0.0517,
      "step": 770
    },
    {
      "epoch": 26.85,
      "grad_norm": 0.13837137818336487,
      "learning_rate": 0.001,
      "loss": 0.0452,
      "step": 772
    },
    {
      "epoch": 26.92,
      "grad_norm": 0.14084497094154358,
      "learning_rate": 0.001,
      "loss": 0.0475,
      "step": 774
    },
    {
      "epoch": 26.99,
      "grad_norm": 0.13428126275539398,
      "learning_rate": 0.001,
      "loss": 0.0505,
      "step": 776
    },
    {
      "epoch": 26.99,
      "eval_loss": 5.000490665435791,
      "eval_runtime": 14.2197,
      "eval_samples_per_second": 23.77,
      "eval_steps_per_second": 3.024,
      "step": 776
    },
    {
      "epoch": 27.06,
      "grad_norm": 0.1252642720937729,
      "learning_rate": 0.001,
      "loss": 0.0372,
      "step": 778
    },
    {
      "epoch": 27.13,
      "grad_norm": 0.1609179526567459,
      "learning_rate": 0.001,
      "loss": 0.0422,
      "step": 780
    },
    {
      "epoch": 27.2,
      "grad_norm": 0.15185533463954926,
      "learning_rate": 0.001,
      "loss": 0.0427,
      "step": 782
    },
    {
      "epoch": 27.27,
      "grad_norm": 0.1398567408323288,
      "learning_rate": 0.001,
      "loss": 0.0452,
      "step": 784
    },
    {
      "epoch": 27.34,
      "grad_norm": 0.13516364991664886,
      "learning_rate": 0.001,
      "loss": 0.0431,
      "step": 786
    },
    {
      "epoch": 27.41,
      "grad_norm": 0.1388152688741684,
      "learning_rate": 0.001,
      "loss": 0.0424,
      "step": 788
    },
    {
      "epoch": 27.48,
      "grad_norm": 0.14043325185775757,
      "learning_rate": 0.001,
      "loss": 0.0428,
      "step": 790
    },
    {
      "epoch": 27.55,
      "grad_norm": 0.14878694713115692,
      "learning_rate": 0.001,
      "loss": 0.0473,
      "step": 792
    },
    {
      "epoch": 27.62,
      "grad_norm": 0.15647174417972565,
      "learning_rate": 0.001,
      "loss": 0.0448,
      "step": 794
    },
    {
      "epoch": 27.69,
      "grad_norm": 0.1453557014465332,
      "learning_rate": 0.001,
      "loss": 0.0441,
      "step": 796
    },
    {
      "epoch": 27.76,
      "grad_norm": 0.13821879029273987,
      "learning_rate": 0.001,
      "loss": 0.0473,
      "step": 798
    },
    {
      "epoch": 27.83,
      "grad_norm": 0.1780141144990921,
      "learning_rate": 0.001,
      "loss": 0.048,
      "step": 800
    },
    {
      "epoch": 27.9,
      "grad_norm": 0.16689875721931458,
      "learning_rate": 0.001,
      "loss": 0.0568,
      "step": 802
    },
    {
      "epoch": 27.97,
      "grad_norm": 0.16479206085205078,
      "learning_rate": 0.001,
      "loss": 0.0576,
      "step": 804
    },
    {
      "epoch": 28.0,
      "eval_loss": 5.019590377807617,
      "eval_runtime": 14.2181,
      "eval_samples_per_second": 23.772,
      "eval_steps_per_second": 3.024,
      "step": 805
    },
    {
      "epoch": 28.03,
      "grad_norm": 0.17976313829421997,
      "learning_rate": 0.001,
      "loss": 0.0586,
      "step": 806
    },
    {
      "epoch": 28.1,
      "grad_norm": 0.15373440086841583,
      "learning_rate": 0.001,
      "loss": 0.0479,
      "step": 808
    },
    {
      "epoch": 28.17,
      "grad_norm": 0.13314908742904663,
      "learning_rate": 0.001,
      "loss": 0.044,
      "step": 810
    },
    {
      "epoch": 28.24,
      "grad_norm": 0.12571971118450165,
      "learning_rate": 0.001,
      "loss": 0.0382,
      "step": 812
    },
    {
      "epoch": 28.31,
      "grad_norm": 0.1362116038799286,
      "learning_rate": 0.001,
      "loss": 0.0412,
      "step": 814
    },
    {
      "epoch": 28.38,
      "grad_norm": 0.1244189664721489,
      "learning_rate": 0.001,
      "loss": 0.0422,
      "step": 816
    },
    {
      "epoch": 28.45,
      "grad_norm": 0.1321047842502594,
      "learning_rate": 0.001,
      "loss": 0.0419,
      "step": 818
    },
    {
      "epoch": 28.52,
      "grad_norm": 0.14995931088924408,
      "learning_rate": 0.001,
      "loss": 0.0453,
      "step": 820
    },
    {
      "epoch": 28.59,
      "grad_norm": 0.1319851577281952,
      "learning_rate": 0.001,
      "loss": 0.0428,
      "step": 822
    },
    {
      "epoch": 28.66,
      "grad_norm": 0.14392071962356567,
      "learning_rate": 0.001,
      "loss": 0.0417,
      "step": 824
    },
    {
      "epoch": 28.73,
      "grad_norm": 0.13848307728767395,
      "learning_rate": 0.001,
      "loss": 0.0445,
      "step": 826
    },
    {
      "epoch": 28.8,
      "grad_norm": 0.1495485007762909,
      "learning_rate": 0.001,
      "loss": 0.0451,
      "step": 828
    },
    {
      "epoch": 28.87,
      "grad_norm": 0.13798055052757263,
      "learning_rate": 0.001,
      "loss": 0.0456,
      "step": 830
    },
    {
      "epoch": 28.94,
      "grad_norm": 0.1452612578868866,
      "learning_rate": 0.001,
      "loss": 0.0448,
      "step": 832
    },
    {
      "epoch": 28.97,
      "eval_loss": 5.109987735748291,
      "eval_runtime": 14.2104,
      "eval_samples_per_second": 23.785,
      "eval_steps_per_second": 3.026,
      "step": 833
    },
    {
      "epoch": 29.01,
      "grad_norm": 0.17332157492637634,
      "learning_rate": 0.001,
      "loss": 0.0438,
      "step": 834
    },
    {
      "epoch": 29.08,
      "grad_norm": 0.12759405374526978,
      "learning_rate": 0.001,
      "loss": 0.0391,
      "step": 836
    },
    {
      "epoch": 29.15,
      "grad_norm": 0.13717012107372284,
      "learning_rate": 0.001,
      "loss": 0.0392,
      "step": 838
    },
    {
      "epoch": 29.22,
      "grad_norm": 0.12544453144073486,
      "learning_rate": 0.001,
      "loss": 0.0373,
      "step": 840
    },
    {
      "epoch": 29.29,
      "grad_norm": 0.12997865676879883,
      "learning_rate": 0.001,
      "loss": 0.0403,
      "step": 842
    },
    {
      "epoch": 29.36,
      "grad_norm": 0.1252465397119522,
      "learning_rate": 0.001,
      "loss": 0.0355,
      "step": 844
    },
    {
      "epoch": 29.43,
      "grad_norm": 0.13186001777648926,
      "learning_rate": 0.001,
      "loss": 0.046,
      "step": 846
    },
    {
      "epoch": 29.5,
      "grad_norm": 0.1309279203414917,
      "learning_rate": 0.001,
      "loss": 0.0423,
      "step": 848
    },
    {
      "epoch": 29.57,
      "grad_norm": 0.13001705706119537,
      "learning_rate": 0.001,
      "loss": 0.0401,
      "step": 850
    },
    {
      "epoch": 29.63,
      "grad_norm": 0.13039255142211914,
      "learning_rate": 0.001,
      "loss": 0.0396,
      "step": 852
    },
    {
      "epoch": 29.7,
      "grad_norm": 0.13635265827178955,
      "learning_rate": 0.001,
      "loss": 0.0426,
      "step": 854
    },
    {
      "epoch": 29.77,
      "grad_norm": 0.16096022725105286,
      "learning_rate": 0.001,
      "loss": 0.0444,
      "step": 856
    },
    {
      "epoch": 29.84,
      "grad_norm": 0.12575644254684448,
      "learning_rate": 0.001,
      "loss": 0.0427,
      "step": 858
    },
    {
      "epoch": 29.91,
      "grad_norm": 0.13646484911441803,
      "learning_rate": 0.001,
      "loss": 0.0442,
      "step": 860
    },
    {
      "epoch": 29.98,
      "grad_norm": 0.13415445387363434,
      "learning_rate": 0.001,
      "loss": 0.0457,
      "step": 862
    },
    {
      "epoch": 29.98,
      "eval_loss": 5.0008440017700195,
      "eval_runtime": 14.2056,
      "eval_samples_per_second": 23.793,
      "eval_steps_per_second": 3.027,
      "step": 862
    },
    {
      "epoch": 30.05,
      "grad_norm": 0.12284772843122482,
      "learning_rate": 0.001,
      "loss": 0.0331,
      "step": 864
    },
    {
      "epoch": 30.12,
      "grad_norm": 0.12438986450433731,
      "learning_rate": 0.001,
      "loss": 0.0372,
      "step": 866
    },
    {
      "epoch": 30.19,
      "grad_norm": 0.12020161747932434,
      "learning_rate": 0.001,
      "loss": 0.0366,
      "step": 868
    },
    {
      "epoch": 30.26,
      "grad_norm": 0.12868288159370422,
      "learning_rate": 0.001,
      "loss": 0.0384,
      "step": 870
    },
    {
      "epoch": 30.33,
      "grad_norm": 0.1276613026857376,
      "learning_rate": 0.001,
      "loss": 0.0415,
      "step": 872
    },
    {
      "epoch": 30.4,
      "grad_norm": 0.13801202178001404,
      "learning_rate": 0.001,
      "loss": 0.0403,
      "step": 874
    },
    {
      "epoch": 30.47,
      "grad_norm": 0.1174420416355133,
      "learning_rate": 0.001,
      "loss": 0.0389,
      "step": 876
    },
    {
      "epoch": 30.54,
      "grad_norm": 0.11008256673812866,
      "learning_rate": 0.001,
      "loss": 0.0342,
      "step": 878
    },
    {
      "epoch": 30.61,
      "grad_norm": 0.12188894301652908,
      "learning_rate": 0.001,
      "loss": 0.0382,
      "step": 880
    },
    {
      "epoch": 30.68,
      "grad_norm": 0.12297231703996658,
      "learning_rate": 0.001,
      "loss": 0.0382,
      "step": 882
    },
    {
      "epoch": 30.75,
      "grad_norm": 0.12277786433696747,
      "learning_rate": 0.001,
      "loss": 0.0381,
      "step": 884
    },
    {
      "epoch": 30.82,
      "grad_norm": 0.13909707963466644,
      "learning_rate": 0.001,
      "loss": 0.0385,
      "step": 886
    },
    {
      "epoch": 30.89,
      "grad_norm": 0.13425001502037048,
      "learning_rate": 0.001,
      "loss": 0.0424,
      "step": 888
    },
    {
      "epoch": 30.96,
      "grad_norm": 0.14270208775997162,
      "learning_rate": 0.001,
      "loss": 0.0442,
      "step": 890
    },
    {
      "epoch": 30.99,
      "eval_loss": 5.50925874710083,
      "eval_runtime": 14.2033,
      "eval_samples_per_second": 23.797,
      "eval_steps_per_second": 3.027,
      "step": 891
    },
    {
      "epoch": 31.03,
      "grad_norm": 0.14988122880458832,
      "learning_rate": 0.001,
      "loss": 0.0391,
      "step": 892
    },
    {
      "epoch": 31.1,
      "grad_norm": 0.12136206030845642,
      "learning_rate": 0.001,
      "loss": 0.0368,
      "step": 894
    },
    {
      "epoch": 31.17,
      "grad_norm": 0.1474607139825821,
      "learning_rate": 0.001,
      "loss": 0.0366,
      "step": 896
    },
    {
      "epoch": 31.23,
      "grad_norm": 0.1664348542690277,
      "learning_rate": 0.001,
      "loss": 0.0364,
      "step": 898
    },
    {
      "epoch": 31.3,
      "grad_norm": 0.12352883070707321,
      "learning_rate": 0.001,
      "loss": 0.0368,
      "step": 900
    },
    {
      "epoch": 31.37,
      "grad_norm": 0.1317415088415146,
      "learning_rate": 0.001,
      "loss": 0.0344,
      "step": 902
    },
    {
      "epoch": 31.44,
      "grad_norm": 0.12790362536907196,
      "learning_rate": 0.001,
      "loss": 0.034,
      "step": 904
    },
    {
      "epoch": 31.51,
      "grad_norm": 0.13639821112155914,
      "learning_rate": 0.001,
      "loss": 0.0385,
      "step": 906
    },
    {
      "epoch": 31.58,
      "grad_norm": 0.14075782895088196,
      "learning_rate": 0.001,
      "loss": 0.0414,
      "step": 908
    },
    {
      "epoch": 31.65,
      "grad_norm": 0.1388426125049591,
      "learning_rate": 0.001,
      "loss": 0.0423,
      "step": 910
    },
    {
      "epoch": 31.72,
      "grad_norm": 0.1304369866847992,
      "learning_rate": 0.001,
      "loss": 0.0418,
      "step": 912
    },
    {
      "epoch": 31.79,
      "grad_norm": 0.13760504126548767,
      "learning_rate": 0.001,
      "loss": 0.0438,
      "step": 914
    },
    {
      "epoch": 31.86,
      "grad_norm": 0.13388106226921082,
      "learning_rate": 0.001,
      "loss": 0.0442,
      "step": 916
    },
    {
      "epoch": 31.93,
      "grad_norm": 0.13234779238700867,
      "learning_rate": 0.001,
      "loss": 0.0403,
      "step": 918
    },
    {
      "epoch": 32.0,
      "grad_norm": 0.12893211841583252,
      "learning_rate": 0.001,
      "loss": 0.0391,
      "step": 920
    },
    {
      "epoch": 32.0,
      "eval_loss": 5.429612159729004,
      "eval_runtime": 14.2128,
      "eval_samples_per_second": 23.781,
      "eval_steps_per_second": 3.025,
      "step": 920
    },
    {
      "epoch": 32.07,
      "grad_norm": 0.12455160915851593,
      "learning_rate": 0.001,
      "loss": 0.0343,
      "step": 922
    },
    {
      "epoch": 32.14,
      "grad_norm": 0.13458769023418427,
      "learning_rate": 0.001,
      "loss": 0.0358,
      "step": 924
    },
    {
      "epoch": 32.21,
      "grad_norm": 0.11662837117910385,
      "learning_rate": 0.001,
      "loss": 0.0359,
      "step": 926
    },
    {
      "epoch": 32.28,
      "grad_norm": 0.1250811070203781,
      "learning_rate": 0.001,
      "loss": 0.0366,
      "step": 928
    },
    {
      "epoch": 32.35,
      "grad_norm": 0.1264219582080841,
      "learning_rate": 0.001,
      "loss": 0.0374,
      "step": 930
    },
    {
      "epoch": 32.42,
      "grad_norm": 0.12272246927022934,
      "learning_rate": 0.001,
      "loss": 0.0378,
      "step": 932
    },
    {
      "epoch": 32.49,
      "grad_norm": 0.12346846610307693,
      "learning_rate": 0.001,
      "loss": 0.0387,
      "step": 934
    },
    {
      "epoch": 32.56,
      "grad_norm": 0.12550891935825348,
      "learning_rate": 0.001,
      "loss": 0.0355,
      "step": 936
    },
    {
      "epoch": 32.63,
      "grad_norm": 0.13145896792411804,
      "learning_rate": 0.001,
      "loss": 0.0371,
      "step": 938
    },
    {
      "epoch": 32.7,
      "grad_norm": 0.12549975514411926,
      "learning_rate": 0.001,
      "loss": 0.038,
      "step": 940
    },
    {
      "epoch": 32.77,
      "grad_norm": 0.13180097937583923,
      "learning_rate": 0.001,
      "loss": 0.0361,
      "step": 942
    },
    {
      "epoch": 32.83,
      "grad_norm": 0.1240687221288681,
      "learning_rate": 0.001,
      "loss": 0.0364,
      "step": 944
    },
    {
      "epoch": 32.9,
      "grad_norm": 0.14110788702964783,
      "learning_rate": 0.001,
      "loss": 0.0421,
      "step": 946
    },
    {
      "epoch": 32.97,
      "grad_norm": 0.14410488307476044,
      "learning_rate": 0.001,
      "loss": 0.0392,
      "step": 948
    },
    {
      "epoch": 32.97,
      "eval_loss": 5.23565673828125,
      "eval_runtime": 14.2111,
      "eval_samples_per_second": 23.784,
      "eval_steps_per_second": 3.026,
      "step": 948
    },
    {
      "epoch": 33.04,
      "grad_norm": 0.12806980311870575,
      "learning_rate": 0.001,
      "loss": 0.041,
      "step": 950
    },
    {
      "epoch": 33.11,
      "grad_norm": 0.12071588635444641,
      "learning_rate": 0.001,
      "loss": 0.0401,
      "step": 952
    },
    {
      "epoch": 33.18,
      "grad_norm": 0.11720431596040726,
      "learning_rate": 0.001,
      "loss": 0.035,
      "step": 954
    },
    {
      "epoch": 33.25,
      "grad_norm": 0.12089748680591583,
      "learning_rate": 0.001,
      "loss": 0.0389,
      "step": 956
    },
    {
      "epoch": 33.32,
      "grad_norm": 0.1227785050868988,
      "learning_rate": 0.001,
      "loss": 0.0389,
      "step": 958
    },
    {
      "epoch": 33.39,
      "grad_norm": 0.12859608232975006,
      "learning_rate": 0.001,
      "loss": 0.0342,
      "step": 960
    },
    {
      "epoch": 33.46,
      "grad_norm": 0.2180158495903015,
      "learning_rate": 0.001,
      "loss": 0.041,
      "step": 962
    },
    {
      "epoch": 33.53,
      "grad_norm": 0.2960141897201538,
      "learning_rate": 0.001,
      "loss": 0.0989,
      "step": 964
    },
    {
      "epoch": 33.6,
      "grad_norm": 0.15140074491500854,
      "learning_rate": 0.001,
      "loss": 0.0503,
      "step": 966
    },
    {
      "epoch": 33.67,
      "grad_norm": 0.13876180350780487,
      "learning_rate": 0.001,
      "loss": 0.0418,
      "step": 968
    },
    {
      "epoch": 33.74,
      "grad_norm": 0.1290924996137619,
      "learning_rate": 0.001,
      "loss": 0.0413,
      "step": 970
    },
    {
      "epoch": 33.81,
      "grad_norm": 0.13152842223644257,
      "learning_rate": 0.001,
      "loss": 0.0398,
      "step": 972
    },
    {
      "epoch": 33.88,
      "grad_norm": 0.1308111995458603,
      "learning_rate": 0.001,
      "loss": 0.0384,
      "step": 974
    },
    {
      "epoch": 33.95,
      "grad_norm": 0.12206979840993881,
      "learning_rate": 0.001,
      "loss": 0.0376,
      "step": 976
    },
    {
      "epoch": 33.98,
      "eval_loss": 5.226629734039307,
      "eval_runtime": 14.2113,
      "eval_samples_per_second": 23.784,
      "eval_steps_per_second": 3.026,
      "step": 977
    },
    {
      "epoch": 34.02,
      "grad_norm": 0.15338914096355438,
      "learning_rate": 0.001,
      "loss": 0.036,
      "step": 978
    },
    {
      "epoch": 34.09,
      "grad_norm": 0.12300385534763336,
      "learning_rate": 0.001,
      "loss": 0.0357,
      "step": 980
    },
    {
      "epoch": 34.16,
      "grad_norm": 0.10930658131837845,
      "learning_rate": 0.001,
      "loss": 0.03,
      "step": 982
    },
    {
      "epoch": 34.23,
      "grad_norm": 0.12387796491384506,
      "learning_rate": 0.001,
      "loss": 0.0393,
      "step": 984
    },
    {
      "epoch": 34.3,
      "grad_norm": 0.1442791372537613,
      "learning_rate": 0.001,
      "loss": 0.0327,
      "step": 986
    },
    {
      "epoch": 34.37,
      "grad_norm": 0.12000349164009094,
      "learning_rate": 0.001,
      "loss": 0.0356,
      "step": 988
    },
    {
      "epoch": 34.43,
      "grad_norm": 0.12804967164993286,
      "learning_rate": 0.001,
      "loss": 0.0415,
      "step": 990
    },
    {
      "epoch": 34.5,
      "grad_norm": 0.12344588339328766,
      "learning_rate": 0.001,
      "loss": 0.0348,
      "step": 992
    },
    {
      "epoch": 34.57,
      "grad_norm": 0.12841206789016724,
      "learning_rate": 0.001,
      "loss": 0.036,
      "step": 994
    },
    {
      "epoch": 34.64,
      "grad_norm": 0.11577708274126053,
      "learning_rate": 0.001,
      "loss": 0.0349,
      "step": 996
    },
    {
      "epoch": 34.71,
      "grad_norm": 0.12154724448919296,
      "learning_rate": 0.001,
      "loss": 0.0384,
      "step": 998
    },
    {
      "epoch": 34.78,
      "grad_norm": 0.12926706671714783,
      "learning_rate": 0.001,
      "loss": 0.0376,
      "step": 1000
    },
    {
      "epoch": 34.85,
      "grad_norm": 0.11031629890203476,
      "learning_rate": 0.001,
      "loss": 0.0341,
      "step": 1002
    },
    {
      "epoch": 34.92,
      "grad_norm": 0.13376009464263916,
      "learning_rate": 0.001,
      "loss": 0.0352,
      "step": 1004
    },
    {
      "epoch": 34.99,
      "grad_norm": 0.1240614578127861,
      "learning_rate": 0.001,
      "loss": 0.0381,
      "step": 1006
    },
    {
      "epoch": 34.99,
      "eval_loss": 5.263000965118408,
      "eval_runtime": 14.2061,
      "eval_samples_per_second": 23.793,
      "eval_steps_per_second": 3.027,
      "step": 1006
    },
    {
      "epoch": 35.06,
      "grad_norm": 0.12797795236110687,
      "learning_rate": 0.001,
      "loss": 0.0339,
      "step": 1008
    },
    {
      "epoch": 35.13,
      "grad_norm": 0.13305290043354034,
      "learning_rate": 0.001,
      "loss": 0.0333,
      "step": 1010
    },
    {
      "epoch": 35.2,
      "grad_norm": 0.12507498264312744,
      "learning_rate": 0.001,
      "loss": 0.0316,
      "step": 1012
    },
    {
      "epoch": 35.27,
      "grad_norm": 0.1163344755768776,
      "learning_rate": 0.001,
      "loss": 0.0323,
      "step": 1014
    },
    {
      "epoch": 35.34,
      "grad_norm": 0.11232735961675644,
      "learning_rate": 0.001,
      "loss": 0.035,
      "step": 1016
    },
    {
      "epoch": 35.41,
      "grad_norm": 0.11937879770994186,
      "learning_rate": 0.001,
      "loss": 0.0343,
      "step": 1018
    },
    {
      "epoch": 35.48,
      "grad_norm": 0.1227354183793068,
      "learning_rate": 0.001,
      "loss": 0.0316,
      "step": 1020
    },
    {
      "epoch": 35.55,
      "grad_norm": 0.1201673150062561,
      "learning_rate": 0.001,
      "loss": 0.0345,
      "step": 1022
    },
    {
      "epoch": 35.62,
      "grad_norm": 0.11625051498413086,
      "learning_rate": 0.001,
      "loss": 0.0305,
      "step": 1024
    },
    {
      "epoch": 35.69,
      "grad_norm": 0.1144721657037735,
      "learning_rate": 0.001,
      "loss": 0.03,
      "step": 1026
    },
    {
      "epoch": 35.76,
      "grad_norm": 0.12341687083244324,
      "learning_rate": 0.001,
      "loss": 0.0374,
      "step": 1028
    },
    {
      "epoch": 35.83,
      "grad_norm": 0.12577305734157562,
      "learning_rate": 0.001,
      "loss": 0.0347,
      "step": 1030
    },
    {
      "epoch": 35.9,
      "grad_norm": 0.11321111768484116,
      "learning_rate": 0.001,
      "loss": 0.0336,
      "step": 1032
    },
    {
      "epoch": 35.97,
      "grad_norm": 0.11875444650650024,
      "learning_rate": 0.001,
      "loss": 0.0339,
      "step": 1034
    },
    {
      "epoch": 36.0,
      "eval_loss": 5.353184700012207,
      "eval_runtime": 14.1981,
      "eval_samples_per_second": 23.806,
      "eval_steps_per_second": 3.029,
      "step": 1035
    },
    {
      "epoch": 36.03,
      "grad_norm": 0.122185118496418,
      "learning_rate": 0.001,
      "loss": 0.0285,
      "step": 1036
    },
    {
      "epoch": 36.1,
      "grad_norm": 0.1075131967663765,
      "learning_rate": 0.001,
      "loss": 0.0312,
      "step": 1038
    },
    {
      "epoch": 36.17,
      "grad_norm": 0.12375488132238388,
      "learning_rate": 0.001,
      "loss": 0.0291,
      "step": 1040
    },
    {
      "epoch": 36.24,
      "grad_norm": 0.12728258967399597,
      "learning_rate": 0.001,
      "loss": 0.0333,
      "step": 1042
    },
    {
      "epoch": 36.31,
      "grad_norm": 0.1169193834066391,
      "learning_rate": 0.001,
      "loss": 0.0309,
      "step": 1044
    },
    {
      "epoch": 36.38,
      "grad_norm": 0.11048964411020279,
      "learning_rate": 0.001,
      "loss": 0.0299,
      "step": 1046
    },
    {
      "epoch": 36.45,
      "grad_norm": 0.11923568695783615,
      "learning_rate": 0.001,
      "loss": 0.0301,
      "step": 1048
    },
    {
      "epoch": 36.52,
      "grad_norm": 0.1160750761628151,
      "learning_rate": 0.001,
      "loss": 0.0287,
      "step": 1050
    },
    {
      "epoch": 36.59,
      "grad_norm": 0.12535539269447327,
      "learning_rate": 0.001,
      "loss": 0.0335,
      "step": 1052
    },
    {
      "epoch": 36.66,
      "grad_norm": 0.13712206482887268,
      "learning_rate": 0.001,
      "loss": 0.0358,
      "step": 1054
    },
    {
      "epoch": 36.73,
      "grad_norm": 0.12085587531328201,
      "learning_rate": 0.001,
      "loss": 0.0362,
      "step": 1056
    },
    {
      "epoch": 36.8,
      "grad_norm": 0.1010272428393364,
      "learning_rate": 0.001,
      "loss": 0.0324,
      "step": 1058
    },
    {
      "epoch": 36.87,
      "grad_norm": 0.11820247769355774,
      "learning_rate": 0.001,
      "loss": 0.0315,
      "step": 1060
    },
    {
      "epoch": 36.94,
      "grad_norm": 0.1259121298789978,
      "learning_rate": 0.001,
      "loss": 0.0377,
      "step": 1062
    },
    {
      "epoch": 36.97,
      "eval_loss": 5.44431734085083,
      "eval_runtime": 14.2065,
      "eval_samples_per_second": 23.792,
      "eval_steps_per_second": 3.027,
      "step": 1063
    },
    {
      "epoch": 37.01,
      "grad_norm": 0.12725719809532166,
      "learning_rate": 0.001,
      "loss": 0.0277,
      "step": 1064
    },
    {
      "epoch": 37.08,
      "grad_norm": 0.11410412192344666,
      "learning_rate": 0.001,
      "loss": 0.0282,
      "step": 1066
    },
    {
      "epoch": 37.15,
      "grad_norm": 0.11205872148275375,
      "learning_rate": 0.001,
      "loss": 0.0302,
      "step": 1068
    },
    {
      "epoch": 37.22,
      "grad_norm": 0.09593572467565536,
      "learning_rate": 0.001,
      "loss": 0.0263,
      "step": 1070
    },
    {
      "epoch": 37.29,
      "grad_norm": 0.12462112307548523,
      "learning_rate": 0.001,
      "loss": 0.0328,
      "step": 1072
    },
    {
      "epoch": 37.36,
      "grad_norm": 0.11873652786016464,
      "learning_rate": 0.001,
      "loss": 0.0328,
      "step": 1074
    },
    {
      "epoch": 37.43,
      "grad_norm": 0.1105048656463623,
      "learning_rate": 0.001,
      "loss": 0.0298,
      "step": 1076
    },
    {
      "epoch": 37.5,
      "grad_norm": 0.10692736506462097,
      "learning_rate": 0.001,
      "loss": 0.0328,
      "step": 1078
    },
    {
      "epoch": 37.57,
      "grad_norm": 0.12755931913852692,
      "learning_rate": 0.001,
      "loss": 0.0321,
      "step": 1080
    },
    {
      "epoch": 37.63,
      "grad_norm": 0.13307899236679077,
      "learning_rate": 0.001,
      "loss": 0.0329,
      "step": 1082
    },
    {
      "epoch": 37.7,
      "grad_norm": 0.12125247716903687,
      "learning_rate": 0.001,
      "loss": 0.034,
      "step": 1084
    },
    {
      "epoch": 37.77,
      "grad_norm": 0.12050389498472214,
      "learning_rate": 0.001,
      "loss": 0.031,
      "step": 1086
    },
    {
      "epoch": 37.84,
      "grad_norm": 0.11437363177537918,
      "learning_rate": 0.001,
      "loss": 0.0308,
      "step": 1088
    },
    {
      "epoch": 37.91,
      "grad_norm": 0.13162565231323242,
      "learning_rate": 0.001,
      "loss": 0.037,
      "step": 1090
    },
    {
      "epoch": 37.98,
      "grad_norm": 0.13537472486495972,
      "learning_rate": 0.001,
      "loss": 0.0336,
      "step": 1092
    },
    {
      "epoch": 37.98,
      "eval_loss": 5.080876350402832,
      "eval_runtime": 14.2123,
      "eval_samples_per_second": 23.782,
      "eval_steps_per_second": 3.026,
      "step": 1092
    },
    {
      "epoch": 38.05,
      "grad_norm": 0.10324922204017639,
      "learning_rate": 0.001,
      "loss": 0.0311,
      "step": 1094
    },
    {
      "epoch": 38.12,
      "grad_norm": 0.10196162760257721,
      "learning_rate": 0.001,
      "loss": 0.0274,
      "step": 1096
    },
    {
      "epoch": 38.19,
      "grad_norm": 0.10867214947938919,
      "learning_rate": 0.001,
      "loss": 0.0281,
      "step": 1098
    },
    {
      "epoch": 38.26,
      "grad_norm": 0.10762563347816467,
      "learning_rate": 0.001,
      "loss": 0.0301,
      "step": 1100
    },
    {
      "epoch": 38.33,
      "grad_norm": 0.10636897385120392,
      "learning_rate": 0.001,
      "loss": 0.0305,
      "step": 1102
    },
    {
      "epoch": 38.4,
      "grad_norm": 0.10994244366884232,
      "learning_rate": 0.001,
      "loss": 0.0328,
      "step": 1104
    },
    {
      "epoch": 38.47,
      "grad_norm": 0.11419042944908142,
      "learning_rate": 0.001,
      "loss": 0.0328,
      "step": 1106
    },
    {
      "epoch": 38.54,
      "grad_norm": 0.10719581693410873,
      "learning_rate": 0.001,
      "loss": 0.0359,
      "step": 1108
    },
    {
      "epoch": 38.61,
      "grad_norm": 0.13429002463817596,
      "learning_rate": 0.001,
      "loss": 0.0302,
      "step": 1110
    },
    {
      "epoch": 38.68,
      "grad_norm": 0.11462274938821793,
      "learning_rate": 0.001,
      "loss": 0.0311,
      "step": 1112
    },
    {
      "epoch": 38.75,
      "grad_norm": 0.12587036192417145,
      "learning_rate": 0.001,
      "loss": 0.0303,
      "step": 1114
    },
    {
      "epoch": 38.82,
      "grad_norm": 0.11628645658493042,
      "learning_rate": 0.001,
      "loss": 0.0294,
      "step": 1116
    },
    {
      "epoch": 38.89,
      "grad_norm": 0.1132420152425766,
      "learning_rate": 0.001,
      "loss": 0.0344,
      "step": 1118
    },
    {
      "epoch": 38.96,
      "grad_norm": 0.12212924659252167,
      "learning_rate": 0.001,
      "loss": 0.0316,
      "step": 1120
    },
    {
      "epoch": 38.96,
      "eval_loss": 5.267875671386719,
      "eval_runtime": 14.2092,
      "eval_samples_per_second": 23.787,
      "eval_steps_per_second": 3.026,
      "step": 1120
    },
    {
      "epoch": 38.96,
      "step": 1120,
      "total_flos": 1.0970359673435505e+18,
      "train_loss": 0.32971865670822026,
      "train_runtime": 25027.7089,
      "train_samples_per_second": 5.87,
      "train_steps_per_second": 0.045
    }
  ],
  "logging_steps": 2,
  "max_steps": 1120,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 500,
  "total_flos": 1.0970359673435505e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}