{
  "best_metric": 0.3966299593448639,
  "best_model_checkpoint": "mikhail-panzo/ceb_b128_le4_s8000/checkpoint-2000",
  "epoch": 396.03960396039605,
  "eval_steps": 500,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 3.9603960396039604,
      "grad_norm": 2.95651912689209,
      "learning_rate": 2.5e-06,
      "loss": 0.7707,
      "step": 50
    },
    {
      "epoch": 7.920792079207921,
      "grad_norm": 1.1273679733276367,
      "learning_rate": 5e-06,
      "loss": 0.6934,
      "step": 100
    },
    {
      "epoch": 11.881188118811881,
      "grad_norm": 2.1830034255981445,
      "learning_rate": 7.5e-06,
      "loss": 0.5983,
      "step": 150
    },
    {
      "epoch": 15.841584158415841,
      "grad_norm": 1.0733364820480347,
      "learning_rate": 1e-05,
      "loss": 0.5094,
      "step": 200
    },
    {
      "epoch": 19.801980198019802,
      "grad_norm": 1.2824387550354004,
      "learning_rate": 1.25e-05,
      "loss": 0.4913,
      "step": 250
    },
    {
      "epoch": 23.762376237623762,
      "grad_norm": 0.9164418578147888,
      "learning_rate": 1.5e-05,
      "loss": 0.4764,
      "step": 300
    },
    {
      "epoch": 27.722772277227723,
      "grad_norm": 0.8938279747962952,
      "learning_rate": 1.75e-05,
      "loss": 0.4648,
      "step": 350
    },
    {
      "epoch": 31.683168316831683,
      "grad_norm": 0.7855221033096313,
      "learning_rate": 2e-05,
      "loss": 0.459,
      "step": 400
    },
    {
      "epoch": 35.64356435643565,
      "grad_norm": 0.9247761964797974,
      "learning_rate": 2.25e-05,
      "loss": 0.4519,
      "step": 450
    },
    {
      "epoch": 39.603960396039604,
      "grad_norm": 1.1086606979370117,
      "learning_rate": 2.5e-05,
      "loss": 0.4432,
      "step": 500
    },
    {
      "epoch": 39.603960396039604,
      "eval_loss": 0.4058004915714264,
      "eval_runtime": 7.6054,
      "eval_samples_per_second": 23.667,
      "eval_steps_per_second": 3.024,
      "step": 500
    },
    {
      "epoch": 43.56435643564357,
      "grad_norm": 1.3994853496551514,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.4438,
      "step": 550
    },
    {
      "epoch": 47.524752475247524,
      "grad_norm": 1.1104933023452759,
      "learning_rate": 3e-05,
      "loss": 0.4367,
      "step": 600
    },
    {
      "epoch": 51.48514851485149,
      "grad_norm": 0.7456745505332947,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.4315,
      "step": 650
    },
    {
      "epoch": 55.445544554455445,
      "grad_norm": 1.0349308252334595,
      "learning_rate": 3.5e-05,
      "loss": 0.4294,
      "step": 700
    },
    {
      "epoch": 59.40594059405941,
      "grad_norm": 0.5290626287460327,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.4234,
      "step": 750
    },
    {
      "epoch": 63.366336633663366,
      "grad_norm": 1.054995059967041,
      "learning_rate": 4e-05,
      "loss": 0.4231,
      "step": 800
    },
    {
      "epoch": 67.32673267326733,
      "grad_norm": 1.1774096488952637,
      "learning_rate": 4.25e-05,
      "loss": 0.4198,
      "step": 850
    },
    {
      "epoch": 71.2871287128713,
      "grad_norm": 0.7513113617897034,
      "learning_rate": 4.5e-05,
      "loss": 0.4182,
      "step": 900
    },
    {
      "epoch": 75.24752475247524,
      "grad_norm": 0.8465940952301025,
      "learning_rate": 4.75e-05,
      "loss": 0.4179,
      "step": 950
    },
    {
      "epoch": 79.20792079207921,
      "grad_norm": 1.601589322090149,
      "learning_rate": 5e-05,
      "loss": 0.4129,
      "step": 1000
    },
    {
      "epoch": 79.20792079207921,
      "eval_loss": 0.3995668292045593,
      "eval_runtime": 8.0229,
      "eval_samples_per_second": 22.436,
      "eval_steps_per_second": 2.867,
      "step": 1000
    },
    {
      "epoch": 83.16831683168317,
      "grad_norm": 2.049264669418335,
      "learning_rate": 5.25e-05,
      "loss": 0.4129,
      "step": 1050
    },
    {
      "epoch": 87.12871287128714,
      "grad_norm": 1.0590511560440063,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.4118,
      "step": 1100
    },
    {
      "epoch": 91.08910891089108,
      "grad_norm": 1.015751600265503,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.4074,
      "step": 1150
    },
    {
      "epoch": 95.04950495049505,
      "grad_norm": 1.295837640762329,
      "learning_rate": 6e-05,
      "loss": 0.4056,
      "step": 1200
    },
    {
      "epoch": 99.00990099009901,
      "grad_norm": 1.49947988986969,
      "learning_rate": 6.25e-05,
      "loss": 0.4052,
      "step": 1250
    },
    {
      "epoch": 102.97029702970298,
      "grad_norm": 1.3930383920669556,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.4072,
      "step": 1300
    },
    {
      "epoch": 106.93069306930693,
      "grad_norm": 1.1130273342132568,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.4015,
      "step": 1350
    },
    {
      "epoch": 110.89108910891089,
      "grad_norm": 2.1075382232666016,
      "learning_rate": 7e-05,
      "loss": 0.3994,
      "step": 1400
    },
    {
      "epoch": 114.85148514851485,
      "grad_norm": 1.8470244407653809,
      "learning_rate": 7.25e-05,
      "loss": 0.397,
      "step": 1450
    },
    {
      "epoch": 118.81188118811882,
      "grad_norm": 2.0715270042419434,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.3992,
      "step": 1500
    },
    {
      "epoch": 118.81188118811882,
      "eval_loss": 0.39855822920799255,
      "eval_runtime": 7.8631,
      "eval_samples_per_second": 22.892,
      "eval_steps_per_second": 2.925,
      "step": 1500
    },
    {
      "epoch": 122.77227722772277,
      "grad_norm": 0.9546501636505127,
      "learning_rate": 7.75e-05,
      "loss": 0.3934,
      "step": 1550
    },
    {
      "epoch": 126.73267326732673,
      "grad_norm": 1.1335992813110352,
      "learning_rate": 8e-05,
      "loss": 0.3927,
      "step": 1600
    },
    {
      "epoch": 130.69306930693068,
      "grad_norm": 1.905771017074585,
      "learning_rate": 8.25e-05,
      "loss": 0.3895,
      "step": 1650
    },
    {
      "epoch": 134.65346534653466,
      "grad_norm": 1.129404067993164,
      "learning_rate": 8.5e-05,
      "loss": 0.3897,
      "step": 1700
    },
    {
      "epoch": 138.6138613861386,
      "grad_norm": 0.9640969038009644,
      "learning_rate": 8.75e-05,
      "loss": 0.3894,
      "step": 1750
    },
    {
      "epoch": 142.5742574257426,
      "grad_norm": 0.9840293526649475,
      "learning_rate": 9e-05,
      "loss": 0.3862,
      "step": 1800
    },
    {
      "epoch": 146.53465346534654,
      "grad_norm": 2.770336627960205,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.3861,
      "step": 1850
    },
    {
      "epoch": 150.4950495049505,
      "grad_norm": 1.7189514636993408,
      "learning_rate": 9.5e-05,
      "loss": 0.3914,
      "step": 1900
    },
    {
      "epoch": 154.45544554455446,
      "grad_norm": 2.454707622528076,
      "learning_rate": 9.75e-05,
      "loss": 0.3848,
      "step": 1950
    },
    {
      "epoch": 158.41584158415841,
      "grad_norm": 2.1555542945861816,
      "learning_rate": 0.0001,
      "loss": 0.3814,
      "step": 2000
    },
    {
      "epoch": 158.41584158415841,
      "eval_loss": 0.3966299593448639,
      "eval_runtime": 7.7159,
      "eval_samples_per_second": 23.329,
      "eval_steps_per_second": 2.981,
      "step": 2000
    },
    {
      "epoch": 162.37623762376236,
      "grad_norm": 1.1228975057601929,
      "learning_rate": 9.916666666666667e-05,
      "loss": 0.3817,
      "step": 2050
    },
    {
      "epoch": 166.33663366336634,
      "grad_norm": 1.629294991493225,
      "learning_rate": 9.833333333333333e-05,
      "loss": 0.3813,
      "step": 2100
    },
    {
      "epoch": 170.2970297029703,
      "grad_norm": 0.9008976221084595,
      "learning_rate": 9.75e-05,
      "loss": 0.377,
      "step": 2150
    },
    {
      "epoch": 174.25742574257427,
      "grad_norm": 2.174643039703369,
      "learning_rate": 9.666666666666667e-05,
      "loss": 0.3737,
      "step": 2200
    },
    {
      "epoch": 178.21782178217822,
      "grad_norm": 1.617219090461731,
      "learning_rate": 9.583333333333334e-05,
      "loss": 0.3736,
      "step": 2250
    },
    {
      "epoch": 182.17821782178217,
      "grad_norm": 1.292871117591858,
      "learning_rate": 9.5e-05,
      "loss": 0.3786,
      "step": 2300
    },
    {
      "epoch": 186.13861386138615,
      "grad_norm": 1.5530415773391724,
      "learning_rate": 9.416666666666667e-05,
      "loss": 0.3726,
      "step": 2350
    },
    {
      "epoch": 190.0990099009901,
      "grad_norm": 0.9837754964828491,
      "learning_rate": 9.333333333333334e-05,
      "loss": 0.3713,
      "step": 2400
    },
    {
      "epoch": 194.05940594059405,
      "grad_norm": 0.8343392610549927,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.3727,
      "step": 2450
    },
    {
      "epoch": 198.01980198019803,
      "grad_norm": 2.9330906867980957,
      "learning_rate": 9.166666666666667e-05,
      "loss": 0.3688,
      "step": 2500
    },
    {
      "epoch": 198.01980198019803,
      "eval_loss": 0.40214845538139343,
      "eval_runtime": 7.9273,
      "eval_samples_per_second": 22.706,
      "eval_steps_per_second": 2.901,
      "step": 2500
    },
    {
      "epoch": 201.98019801980197,
      "grad_norm": 2.427091598510742,
      "learning_rate": 9.083333333333334e-05,
      "loss": 0.3699,
      "step": 2550
    },
    {
      "epoch": 205.94059405940595,
      "grad_norm": 0.680686891078949,
      "learning_rate": 9e-05,
      "loss": 0.3659,
      "step": 2600
    },
    {
      "epoch": 209.9009900990099,
      "grad_norm": 2.0860631465911865,
      "learning_rate": 8.916666666666667e-05,
      "loss": 0.3631,
      "step": 2650
    },
    {
      "epoch": 213.86138613861385,
      "grad_norm": 0.7303352355957031,
      "learning_rate": 8.833333333333333e-05,
      "loss": 0.3656,
      "step": 2700
    },
    {
      "epoch": 217.82178217821783,
      "grad_norm": 1.2716140747070312,
      "learning_rate": 8.75e-05,
      "loss": 0.3627,
      "step": 2750
    },
    {
      "epoch": 221.78217821782178,
      "grad_norm": 1.5935007333755493,
      "learning_rate": 8.666666666666667e-05,
      "loss": 0.3607,
      "step": 2800
    },
    {
      "epoch": 225.74257425742573,
      "grad_norm": 1.156733512878418,
      "learning_rate": 8.583333333333334e-05,
      "loss": 0.3594,
      "step": 2850
    },
    {
      "epoch": 229.7029702970297,
      "grad_norm": 1.1708168983459473,
      "learning_rate": 8.5e-05,
      "loss": 0.3596,
      "step": 2900
    },
    {
      "epoch": 233.66336633663366,
      "grad_norm": 1.4882978200912476,
      "learning_rate": 8.416666666666668e-05,
      "loss": 0.3613,
      "step": 2950
    },
    {
      "epoch": 237.62376237623764,
      "grad_norm": 1.0416059494018555,
      "learning_rate": 8.333333333333334e-05,
      "loss": 0.3574,
      "step": 3000
    },
    {
      "epoch": 237.62376237623764,
      "eval_loss": 0.39999938011169434,
      "eval_runtime": 7.4331,
      "eval_samples_per_second": 24.216,
      "eval_steps_per_second": 3.094,
      "step": 3000
    },
    {
      "epoch": 241.58415841584159,
      "grad_norm": 0.9737104773521423,
      "learning_rate": 8.25e-05,
      "loss": 0.3555,
      "step": 3050
    },
    {
      "epoch": 245.54455445544554,
      "grad_norm": 2.4858767986297607,
      "learning_rate": 8.166666666666667e-05,
      "loss": 0.3585,
      "step": 3100
    },
    {
      "epoch": 249.5049504950495,
      "grad_norm": 0.7558372616767883,
      "learning_rate": 8.083333333333334e-05,
      "loss": 0.3562,
      "step": 3150
    },
    {
      "epoch": 253.46534653465346,
      "grad_norm": 0.5738223195075989,
      "learning_rate": 8e-05,
      "loss": 0.3515,
      "step": 3200
    },
    {
      "epoch": 257.4257425742574,
      "grad_norm": 0.8245187401771545,
      "learning_rate": 7.916666666666666e-05,
      "loss": 0.3557,
      "step": 3250
    },
    {
      "epoch": 261.38613861386136,
      "grad_norm": 0.8837783336639404,
      "learning_rate": 7.833333333333333e-05,
      "loss": 0.3556,
      "step": 3300
    },
    {
      "epoch": 265.34653465346537,
      "grad_norm": 1.0626522302627563,
      "learning_rate": 7.75e-05,
      "loss": 0.3531,
      "step": 3350
    },
    {
      "epoch": 269.3069306930693,
      "grad_norm": 0.9158534407615662,
      "learning_rate": 7.666666666666667e-05,
      "loss": 0.3504,
      "step": 3400
    },
    {
      "epoch": 273.26732673267327,
      "grad_norm": 1.1604119539260864,
      "learning_rate": 7.583333333333334e-05,
      "loss": 0.3502,
      "step": 3450
    },
    {
      "epoch": 277.2277227722772,
      "grad_norm": 0.5547149181365967,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.3482,
      "step": 3500
    },
    {
      "epoch": 277.2277227722772,
      "eval_loss": 0.39900416135787964,
      "eval_runtime": 7.7665,
      "eval_samples_per_second": 23.177,
      "eval_steps_per_second": 2.961,
      "step": 3500
    },
    {
      "epoch": 281.18811881188117,
      "grad_norm": 1.213918924331665,
      "learning_rate": 7.416666666666668e-05,
      "loss": 0.3465,
      "step": 3550
    },
    {
      "epoch": 285.1485148514852,
      "grad_norm": 1.0824240446090698,
      "learning_rate": 7.333333333333333e-05,
      "loss": 0.3467,
      "step": 3600
    },
    {
      "epoch": 289.1089108910891,
      "grad_norm": 0.8079837560653687,
      "learning_rate": 7.25e-05,
      "loss": 0.3457,
      "step": 3650
    },
    {
      "epoch": 293.0693069306931,
      "grad_norm": 1.1617168188095093,
      "learning_rate": 7.166666666666667e-05,
      "loss": 0.3457,
      "step": 3700
    },
    {
      "epoch": 297.029702970297,
      "grad_norm": 1.0282728672027588,
      "learning_rate": 7.083333333333334e-05,
      "loss": 0.3491,
      "step": 3750
    },
    {
      "epoch": 300.990099009901,
      "grad_norm": 0.8772862553596497,
      "learning_rate": 7e-05,
      "loss": 0.3494,
      "step": 3800
    },
    {
      "epoch": 304.9504950495049,
      "grad_norm": 1.0405845642089844,
      "learning_rate": 6.916666666666666e-05,
      "loss": 0.3469,
      "step": 3850
    },
    {
      "epoch": 308.91089108910893,
      "grad_norm": 0.7217592000961304,
      "learning_rate": 6.833333333333333e-05,
      "loss": 0.3449,
      "step": 3900
    },
    {
      "epoch": 312.8712871287129,
      "grad_norm": 0.7694669365882874,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.341,
      "step": 3950
    },
    {
      "epoch": 316.83168316831683,
      "grad_norm": 0.6203835010528564,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.3421,
      "step": 4000
    },
    {
      "epoch": 316.83168316831683,
      "eval_loss": 0.4013901650905609,
      "eval_runtime": 8.6905,
      "eval_samples_per_second": 20.712,
      "eval_steps_per_second": 2.647,
      "step": 4000
    },
    {
      "epoch": 320.7920792079208,
      "grad_norm": 0.7539447546005249,
      "learning_rate": 6.583333333333334e-05,
      "loss": 0.3445,
      "step": 4050
    },
    {
      "epoch": 324.7524752475247,
      "grad_norm": 0.9108504056930542,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.341,
      "step": 4100
    },
    {
      "epoch": 328.71287128712873,
      "grad_norm": 0.5291554927825928,
      "learning_rate": 6.416666666666668e-05,
      "loss": 0.3412,
      "step": 4150
    },
    {
      "epoch": 332.6732673267327,
      "grad_norm": 1.2332825660705566,
      "learning_rate": 6.333333333333333e-05,
      "loss": 0.3408,
      "step": 4200
    },
    {
      "epoch": 336.63366336633663,
      "grad_norm": 0.5890567302703857,
      "learning_rate": 6.25e-05,
      "loss": 0.339,
      "step": 4250
    },
    {
      "epoch": 340.5940594059406,
      "grad_norm": 1.05361807346344,
      "learning_rate": 6.168333333333333e-05,
      "loss": 0.3388,
      "step": 4300
    },
    {
      "epoch": 344.55445544554453,
      "grad_norm": 1.6529746055603027,
      "learning_rate": 6.085000000000001e-05,
      "loss": 0.3423,
      "step": 4350
    },
    {
      "epoch": 348.51485148514854,
      "grad_norm": 0.6428557634353638,
      "learning_rate": 6.0016666666666664e-05,
      "loss": 0.3406,
      "step": 4400
    },
    {
      "epoch": 352.4752475247525,
      "grad_norm": 0.832099437713623,
      "learning_rate": 5.918333333333333e-05,
      "loss": 0.3399,
      "step": 4450
    },
    {
      "epoch": 356.43564356435644,
      "grad_norm": 1.4958369731903076,
      "learning_rate": 5.835e-05,
      "loss": 0.3414,
      "step": 4500
    },
    {
      "epoch": 356.43564356435644,
      "eval_loss": 0.4059315621852875,
      "eval_runtime": 7.3212,
      "eval_samples_per_second": 24.586,
      "eval_steps_per_second": 3.142,
      "step": 4500
    },
    {
      "epoch": 360.3960396039604,
      "grad_norm": 0.7380837798118591,
      "learning_rate": 5.751666666666667e-05,
      "loss": 0.3363,
      "step": 4550
    },
    {
      "epoch": 364.35643564356434,
      "grad_norm": 1.1721248626708984,
      "learning_rate": 5.668333333333333e-05,
      "loss": 0.3385,
      "step": 4600
    },
    {
      "epoch": 368.3168316831683,
      "grad_norm": 0.6880260109901428,
      "learning_rate": 5.585e-05,
      "loss": 0.3355,
      "step": 4650
    },
    {
      "epoch": 372.2772277227723,
      "grad_norm": 0.614639401435852,
      "learning_rate": 5.501666666666667e-05,
      "loss": 0.3378,
      "step": 4700
    },
    {
      "epoch": 376.23762376237624,
      "grad_norm": 0.6579806208610535,
      "learning_rate": 5.4183333333333334e-05,
      "loss": 0.3342,
      "step": 4750
    },
    {
      "epoch": 380.1980198019802,
      "grad_norm": 0.7675396203994751,
      "learning_rate": 5.335e-05,
      "loss": 0.3333,
      "step": 4800
    },
    {
      "epoch": 384.15841584158414,
      "grad_norm": 0.6529810428619385,
      "learning_rate": 5.251666666666667e-05,
      "loss": 0.3348,
      "step": 4850
    },
    {
      "epoch": 388.1188118811881,
      "grad_norm": 0.6606137752532959,
      "learning_rate": 5.168333333333334e-05,
      "loss": 0.3351,
      "step": 4900
    },
    {
      "epoch": 392.0792079207921,
      "grad_norm": 0.5794258713722229,
      "learning_rate": 5.0849999999999996e-05,
      "loss": 0.3319,
      "step": 4950
    },
    {
      "epoch": 396.03960396039605,
      "grad_norm": 0.8365175127983093,
      "learning_rate": 5.0016666666666665e-05,
      "loss": 0.3328,
      "step": 5000
    },
    {
      "epoch": 396.03960396039605,
      "eval_loss": 0.40650251507759094,
      "eval_runtime": 7.4758,
      "eval_samples_per_second": 24.078,
      "eval_steps_per_second": 3.077,
      "step": 5000
    }
  ],
  "logging_steps": 50,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 667,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.080405925015103e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}