{
  "best_metric": 0.30681997537612915,
  "best_model_checkpoint": "mikhail_panzo/zlm_b128_le4_s12000/checkpoint-9500",
  "epoch": 16.75392670157068,
  "eval_steps": 500,
  "global_step": 10000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08376963350785341,
      "grad_norm": 2.469545841217041,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 1.0402,
      "step": 50
    },
    {
      "epoch": 0.16753926701570682,
      "grad_norm": 3.3642780780792236,
      "learning_rate": 4.950000000000001e-06,
      "loss": 0.8472,
      "step": 100
    },
    {
      "epoch": 0.2513089005235602,
      "grad_norm": 1.8375087976455688,
      "learning_rate": 7.45e-06,
      "loss": 0.7331,
      "step": 150
    },
    {
      "epoch": 0.33507853403141363,
      "grad_norm": 3.700824737548828,
      "learning_rate": 9.950000000000001e-06,
      "loss": 0.6472,
      "step": 200
    },
    {
      "epoch": 0.418848167539267,
      "grad_norm": 1.7632888555526733,
      "learning_rate": 1.2450000000000001e-05,
      "loss": 0.6112,
      "step": 250
    },
    {
      "epoch": 0.5026178010471204,
      "grad_norm": 3.1861908435821533,
      "learning_rate": 1.4950000000000001e-05,
      "loss": 0.5865,
      "step": 300
    },
    {
      "epoch": 0.5863874345549738,
      "grad_norm": 2.2845046520233154,
      "learning_rate": 1.745e-05,
      "loss": 0.5682,
      "step": 350
    },
    {
      "epoch": 0.6701570680628273,
      "grad_norm": 2.079210042953491,
      "learning_rate": 1.995e-05,
      "loss": 0.5465,
      "step": 400
    },
    {
      "epoch": 0.7539267015706806,
      "grad_norm": 3.669891119003296,
      "learning_rate": 2.245e-05,
      "loss": 0.5302,
      "step": 450
    },
    {
      "epoch": 0.837696335078534,
      "grad_norm": 2.4679417610168457,
      "learning_rate": 2.495e-05,
      "loss": 0.5137,
      "step": 500
    },
    {
      "epoch": 0.837696335078534,
      "eval_loss": 0.4513999819755554,
      "eval_runtime": 277.1074,
      "eval_samples_per_second": 30.634,
      "eval_steps_per_second": 3.832,
      "step": 500
    },
    {
      "epoch": 0.9214659685863874,
      "grad_norm": 2.703871726989746,
      "learning_rate": 2.7450000000000003e-05,
      "loss": 0.5071,
      "step": 550
    },
    {
      "epoch": 1.0052356020942408,
      "grad_norm": 3.8184635639190674,
      "learning_rate": 2.995e-05,
      "loss": 0.4971,
      "step": 600
    },
    {
      "epoch": 1.0890052356020943,
      "grad_norm": 2.2857866287231445,
      "learning_rate": 3.245e-05,
      "loss": 0.4955,
      "step": 650
    },
    {
      "epoch": 1.1727748691099475,
      "grad_norm": 3.5974085330963135,
      "learning_rate": 3.495e-05,
      "loss": 0.4935,
      "step": 700
    },
    {
      "epoch": 1.256544502617801,
      "grad_norm": 1.720818281173706,
      "learning_rate": 3.745e-05,
      "loss": 0.4763,
      "step": 750
    },
    {
      "epoch": 1.3403141361256545,
      "grad_norm": 3.080139636993408,
      "learning_rate": 3.995e-05,
      "loss": 0.4735,
      "step": 800
    },
    {
      "epoch": 1.4240837696335078,
      "grad_norm": 4.877579212188721,
      "learning_rate": 4.245e-05,
      "loss": 0.4654,
      "step": 850
    },
    {
      "epoch": 1.5078534031413613,
      "grad_norm": 3.383965253829956,
      "learning_rate": 4.495e-05,
      "loss": 0.4628,
      "step": 900
    },
    {
      "epoch": 1.5916230366492146,
      "grad_norm": 3.3636982440948486,
      "learning_rate": 4.745e-05,
      "loss": 0.4541,
      "step": 950
    },
    {
      "epoch": 1.675392670157068,
      "grad_norm": 1.666568398475647,
      "learning_rate": 4.995e-05,
      "loss": 0.4565,
      "step": 1000
    },
    {
      "epoch": 1.675392670157068,
      "eval_loss": 0.4093586802482605,
      "eval_runtime": 269.4005,
      "eval_samples_per_second": 31.511,
      "eval_steps_per_second": 3.942,
      "step": 1000
    },
    {
      "epoch": 1.7591623036649215,
      "grad_norm": 1.6522510051727295,
      "learning_rate": 5.245e-05,
      "loss": 0.4541,
      "step": 1050
    },
    {
      "epoch": 1.8429319371727748,
      "grad_norm": 1.6531606912612915,
      "learning_rate": 5.495e-05,
      "loss": 0.4448,
      "step": 1100
    },
    {
      "epoch": 1.9267015706806283,
      "grad_norm": 3.3253750801086426,
      "learning_rate": 5.745e-05,
      "loss": 0.4346,
      "step": 1150
    },
    {
      "epoch": 2.0104712041884816,
      "grad_norm": 2.0393073558807373,
      "learning_rate": 5.995000000000001e-05,
      "loss": 0.4314,
      "step": 1200
    },
    {
      "epoch": 2.094240837696335,
      "grad_norm": 1.906546950340271,
      "learning_rate": 6.245000000000001e-05,
      "loss": 0.4327,
      "step": 1250
    },
    {
      "epoch": 2.1780104712041886,
      "grad_norm": 1.7925021648406982,
      "learning_rate": 6.494999999999999e-05,
      "loss": 0.4285,
      "step": 1300
    },
    {
      "epoch": 2.261780104712042,
      "grad_norm": 2.5238988399505615,
      "learning_rate": 6.745e-05,
      "loss": 0.4251,
      "step": 1350
    },
    {
      "epoch": 2.345549738219895,
      "grad_norm": 2.53450345993042,
      "learning_rate": 6.995e-05,
      "loss": 0.4284,
      "step": 1400
    },
    {
      "epoch": 2.4293193717277486,
      "grad_norm": 1.9077616930007935,
      "learning_rate": 7.245000000000001e-05,
      "loss": 0.4244,
      "step": 1450
    },
    {
      "epoch": 2.513089005235602,
      "grad_norm": 1.5720113515853882,
      "learning_rate": 7.495e-05,
      "loss": 0.4171,
      "step": 1500
    },
    {
      "epoch": 2.513089005235602,
      "eval_loss": 0.384281188249588,
      "eval_runtime": 274.8763,
      "eval_samples_per_second": 30.883,
      "eval_steps_per_second": 3.864,
      "step": 1500
    },
    {
      "epoch": 2.5968586387434556,
      "grad_norm": 2.272549867630005,
      "learning_rate": 7.745e-05,
      "loss": 0.4175,
      "step": 1550
    },
    {
      "epoch": 2.680628272251309,
      "grad_norm": 1.3440821170806885,
      "learning_rate": 7.995e-05,
      "loss": 0.4134,
      "step": 1600
    },
    {
      "epoch": 2.7643979057591626,
      "grad_norm": 1.99045729637146,
      "learning_rate": 8.245e-05,
      "loss": 0.4117,
      "step": 1650
    },
    {
      "epoch": 2.8481675392670156,
      "grad_norm": 1.3599165678024292,
      "learning_rate": 8.495e-05,
      "loss": 0.4101,
      "step": 1700
    },
    {
      "epoch": 2.931937172774869,
      "grad_norm": 2.3722105026245117,
      "learning_rate": 8.745000000000001e-05,
      "loss": 0.413,
      "step": 1750
    },
    {
      "epoch": 3.0157068062827226,
      "grad_norm": 1.5704238414764404,
      "learning_rate": 8.995e-05,
      "loss": 0.4086,
      "step": 1800
    },
    {
      "epoch": 3.099476439790576,
      "grad_norm": 3.272968053817749,
      "learning_rate": 9.245e-05,
      "loss": 0.4057,
      "step": 1850
    },
    {
      "epoch": 3.183246073298429,
      "grad_norm": 2.658064126968384,
      "learning_rate": 9.495e-05,
      "loss": 0.4071,
      "step": 1900
    },
    {
      "epoch": 3.2670157068062826,
      "grad_norm": 1.688971757888794,
      "learning_rate": 9.745000000000001e-05,
      "loss": 0.3984,
      "step": 1950
    },
    {
      "epoch": 3.350785340314136,
      "grad_norm": 1.4563082456588745,
      "learning_rate": 9.995e-05,
      "loss": 0.404,
      "step": 2000
    },
    {
      "epoch": 3.350785340314136,
      "eval_loss": 0.36596959829330444,
      "eval_runtime": 268.3638,
      "eval_samples_per_second": 31.632,
      "eval_steps_per_second": 3.957,
      "step": 2000
    },
    {
      "epoch": 3.4345549738219896,
      "grad_norm": 1.6672757863998413,
      "learning_rate": 9.951e-05,
      "loss": 0.4032,
      "step": 2050
    },
    {
      "epoch": 3.518324607329843,
      "grad_norm": 2.0084328651428223,
      "learning_rate": 9.901e-05,
      "loss": 0.4003,
      "step": 2100
    },
    {
      "epoch": 3.6020942408376966,
      "grad_norm": 1.8289756774902344,
      "learning_rate": 9.851e-05,
      "loss": 0.3986,
      "step": 2150
    },
    {
      "epoch": 3.6858638743455496,
      "grad_norm": 1.791348934173584,
      "learning_rate": 9.801e-05,
      "loss": 0.3928,
      "step": 2200
    },
    {
      "epoch": 3.769633507853403,
      "grad_norm": 1.9079582691192627,
      "learning_rate": 9.751e-05,
      "loss": 0.3899,
      "step": 2250
    },
    {
      "epoch": 3.8534031413612566,
      "grad_norm": 7.154463768005371,
      "learning_rate": 9.701e-05,
      "loss": 0.3894,
      "step": 2300
    },
    {
      "epoch": 3.93717277486911,
      "grad_norm": 1.9638899564743042,
      "learning_rate": 9.651e-05,
      "loss": 0.399,
      "step": 2350
    },
    {
      "epoch": 4.020942408376963,
      "grad_norm": 2.147676706314087,
      "learning_rate": 9.601e-05,
      "loss": 0.3902,
      "step": 2400
    },
    {
      "epoch": 4.104712041884817,
      "grad_norm": 2.282815456390381,
      "learning_rate": 9.551e-05,
      "loss": 0.3906,
      "step": 2450
    },
    {
      "epoch": 4.18848167539267,
      "grad_norm": 1.2255228757858276,
      "learning_rate": 9.501e-05,
      "loss": 0.3851,
      "step": 2500
    },
    {
      "epoch": 4.18848167539267,
      "eval_loss": 0.354680597782135,
      "eval_runtime": 269.5668,
      "eval_samples_per_second": 31.491,
      "eval_steps_per_second": 3.94,
      "step": 2500
    },
    {
      "epoch": 4.272251308900524,
      "grad_norm": 1.2088522911071777,
      "learning_rate": 9.451000000000002e-05,
      "loss": 0.3891,
      "step": 2550
    },
    {
      "epoch": 4.356020942408377,
      "grad_norm": 1.3467260599136353,
      "learning_rate": 9.402000000000001e-05,
      "loss": 0.3854,
      "step": 2600
    },
    {
      "epoch": 4.439790575916231,
      "grad_norm": 1.2227791547775269,
      "learning_rate": 9.352000000000001e-05,
      "loss": 0.3818,
      "step": 2650
    },
    {
      "epoch": 4.523560209424084,
      "grad_norm": 1.0672763586044312,
      "learning_rate": 9.302e-05,
      "loss": 0.389,
      "step": 2700
    },
    {
      "epoch": 4.607329842931938,
      "grad_norm": 1.6198370456695557,
      "learning_rate": 9.252e-05,
      "loss": 0.3811,
      "step": 2750
    },
    {
      "epoch": 4.69109947643979,
      "grad_norm": 1.296200156211853,
      "learning_rate": 9.202e-05,
      "loss": 0.382,
      "step": 2800
    },
    {
      "epoch": 4.774869109947644,
      "grad_norm": 1.533375859260559,
      "learning_rate": 9.152e-05,
      "loss": 0.381,
      "step": 2850
    },
    {
      "epoch": 4.858638743455497,
      "grad_norm": 2.161743640899658,
      "learning_rate": 9.102e-05,
      "loss": 0.3829,
      "step": 2900
    },
    {
      "epoch": 4.942408376963351,
      "grad_norm": 1.2308694124221802,
      "learning_rate": 9.052000000000001e-05,
      "loss": 0.3806,
      "step": 2950
    },
    {
      "epoch": 5.026178010471204,
      "grad_norm": 1.2448184490203857,
      "learning_rate": 9.002000000000001e-05,
      "loss": 0.3808,
      "step": 3000
    },
    {
      "epoch": 5.026178010471204,
      "eval_loss": 0.3451012670993805,
      "eval_runtime": 274.5252,
      "eval_samples_per_second": 30.922,
      "eval_steps_per_second": 3.868,
      "step": 3000
    },
    {
      "epoch": 5.109947643979058,
      "grad_norm": 1.4961705207824707,
      "learning_rate": 8.952000000000001e-05,
      "loss": 0.3755,
      "step": 3050
    },
    {
      "epoch": 5.193717277486911,
      "grad_norm": 1.2276802062988281,
      "learning_rate": 8.902e-05,
      "loss": 0.3771,
      "step": 3100
    },
    {
      "epoch": 5.277486910994765,
      "grad_norm": 1.3186664581298828,
      "learning_rate": 8.852e-05,
      "loss": 0.3766,
      "step": 3150
    },
    {
      "epoch": 5.361256544502618,
      "grad_norm": 1.4056960344314575,
      "learning_rate": 8.802e-05,
      "loss": 0.3761,
      "step": 3200
    },
    {
      "epoch": 5.445026178010472,
      "grad_norm": 1.4609332084655762,
      "learning_rate": 8.752e-05,
      "loss": 0.377,
      "step": 3250
    },
    {
      "epoch": 5.528795811518324,
      "grad_norm": 1.2348676919937134,
      "learning_rate": 8.702e-05,
      "loss": 0.3761,
      "step": 3300
    },
    {
      "epoch": 5.612565445026178,
      "grad_norm": 1.2290446758270264,
      "learning_rate": 8.652e-05,
      "loss": 0.3735,
      "step": 3350
    },
    {
      "epoch": 5.696335078534031,
      "grad_norm": 1.0205050706863403,
      "learning_rate": 8.602e-05,
      "loss": 0.3709,
      "step": 3400
    },
    {
      "epoch": 5.780104712041885,
      "grad_norm": 1.83589768409729,
      "learning_rate": 8.552e-05,
      "loss": 0.3709,
      "step": 3450
    },
    {
      "epoch": 5.863874345549738,
      "grad_norm": 1.8419536352157593,
      "learning_rate": 8.502e-05,
      "loss": 0.3704,
      "step": 3500
    },
    {
      "epoch": 5.863874345549738,
      "eval_loss": 0.33645617961883545,
      "eval_runtime": 267.7957,
      "eval_samples_per_second": 31.7,
      "eval_steps_per_second": 3.966,
      "step": 3500
    },
    {
      "epoch": 5.947643979057592,
      "grad_norm": 0.9079850316047668,
      "learning_rate": 8.452e-05,
      "loss": 0.3674,
      "step": 3550
    },
    {
      "epoch": 6.031413612565445,
      "grad_norm": 1.6563224792480469,
      "learning_rate": 8.402e-05,
      "loss": 0.368,
      "step": 3600
    },
    {
      "epoch": 6.115183246073299,
      "grad_norm": 0.9500299096107483,
      "learning_rate": 8.352000000000001e-05,
      "loss": 0.3692,
      "step": 3650
    },
    {
      "epoch": 6.198952879581152,
      "grad_norm": 1.219301700592041,
      "learning_rate": 8.302000000000001e-05,
      "loss": 0.3664,
      "step": 3700
    },
    {
      "epoch": 6.282722513089006,
      "grad_norm": 1.466737985610962,
      "learning_rate": 8.252e-05,
      "loss": 0.3697,
      "step": 3750
    },
    {
      "epoch": 6.366492146596858,
      "grad_norm": 1.329934000968933,
      "learning_rate": 8.202e-05,
      "loss": 0.3676,
      "step": 3800
    },
    {
      "epoch": 6.450261780104712,
      "grad_norm": 1.0545940399169922,
      "learning_rate": 8.152e-05,
      "loss": 0.3673,
      "step": 3850
    },
    {
      "epoch": 6.534031413612565,
      "grad_norm": 1.7130441665649414,
      "learning_rate": 8.102000000000001e-05,
      "loss": 0.3658,
      "step": 3900
    },
    {
      "epoch": 6.617801047120419,
      "grad_norm": 2.0463407039642334,
      "learning_rate": 8.052000000000001e-05,
      "loss": 0.3681,
      "step": 3950
    },
    {
      "epoch": 6.701570680628272,
      "grad_norm": 1.0734078884124756,
      "learning_rate": 8.002000000000001e-05,
      "loss": 0.3664,
      "step": 4000
    },
    {
      "epoch": 6.701570680628272,
      "eval_loss": 0.3334435820579529,
      "eval_runtime": 269.2566,
      "eval_samples_per_second": 31.528,
      "eval_steps_per_second": 3.944,
      "step": 4000
    },
    {
      "epoch": 6.785340314136126,
      "grad_norm": 0.9216086864471436,
      "learning_rate": 7.952000000000001e-05,
      "loss": 0.3619,
      "step": 4050
    },
    {
      "epoch": 6.869109947643979,
      "grad_norm": 1.802753210067749,
      "learning_rate": 7.902e-05,
      "loss": 0.3628,
      "step": 4100
    },
    {
      "epoch": 6.952879581151833,
      "grad_norm": 1.3391004800796509,
      "learning_rate": 7.852e-05,
      "loss": 0.3626,
      "step": 4150
    },
    {
      "epoch": 7.036649214659686,
      "grad_norm": 0.9989505410194397,
      "learning_rate": 7.802e-05,
      "loss": 0.3641,
      "step": 4200
    },
    {
      "epoch": 7.12041884816754,
      "grad_norm": 1.318490743637085,
      "learning_rate": 7.752e-05,
      "loss": 0.3661,
      "step": 4250
    },
    {
      "epoch": 7.204188481675392,
      "grad_norm": 1.4725680351257324,
      "learning_rate": 7.702e-05,
      "loss": 0.359,
      "step": 4300
    },
    {
      "epoch": 7.287958115183246,
      "grad_norm": 1.3948692083358765,
      "learning_rate": 7.652e-05,
      "loss": 0.3622,
      "step": 4350
    },
    {
      "epoch": 7.371727748691099,
      "grad_norm": 1.4226194620132446,
      "learning_rate": 7.602000000000001e-05,
      "loss": 0.3641,
      "step": 4400
    },
    {
      "epoch": 7.455497382198953,
      "grad_norm": 1.008616328239441,
      "learning_rate": 7.552e-05,
      "loss": 0.3643,
      "step": 4450
    },
    {
      "epoch": 7.539267015706806,
      "grad_norm": 1.3399937152862549,
      "learning_rate": 7.502e-05,
      "loss": 0.3598,
      "step": 4500
    },
    {
      "epoch": 7.539267015706806,
      "eval_loss": 0.32835692167282104,
      "eval_runtime": 275.5633,
      "eval_samples_per_second": 30.806,
      "eval_steps_per_second": 3.854,
      "step": 4500
    },
    {
      "epoch": 7.62303664921466,
      "grad_norm": 1.2946890592575073,
      "learning_rate": 7.452e-05,
      "loss": 0.3595,
      "step": 4550
    },
    {
      "epoch": 7.706806282722513,
      "grad_norm": 1.672792911529541,
      "learning_rate": 7.402e-05,
      "loss": 0.3651,
      "step": 4600
    },
    {
      "epoch": 7.790575916230367,
      "grad_norm": 1.3620250225067139,
      "learning_rate": 7.352e-05,
      "loss": 0.3666,
      "step": 4650
    },
    {
      "epoch": 7.87434554973822,
      "grad_norm": 2.236588954925537,
      "learning_rate": 7.302e-05,
      "loss": 0.365,
      "step": 4700
    },
    {
      "epoch": 7.958115183246074,
      "grad_norm": 2.052438259124756,
      "learning_rate": 7.252e-05,
      "loss": 0.3557,
      "step": 4750
    },
    {
      "epoch": 8.041884816753926,
      "grad_norm": 1.1328259706497192,
      "learning_rate": 7.202e-05,
      "loss": 0.3573,
      "step": 4800
    },
    {
      "epoch": 8.12565445026178,
      "grad_norm": 1.4987516403198242,
      "learning_rate": 7.151999999999999e-05,
      "loss": 0.3547,
      "step": 4850
    },
    {
      "epoch": 8.209424083769633,
      "grad_norm": 1.502505898475647,
      "learning_rate": 7.102000000000001e-05,
      "loss": 0.3668,
      "step": 4900
    },
    {
      "epoch": 8.293193717277488,
      "grad_norm": 1.4446626901626587,
      "learning_rate": 7.052000000000001e-05,
      "loss": 0.3566,
      "step": 4950
    },
    {
      "epoch": 8.37696335078534,
      "grad_norm": 0.9949765205383301,
      "learning_rate": 7.002000000000001e-05,
      "loss": 0.3612,
      "step": 5000
    },
    {
      "epoch": 8.37696335078534,
      "eval_loss": 0.3273167312145233,
      "eval_runtime": 277.1246,
      "eval_samples_per_second": 30.632,
      "eval_steps_per_second": 3.832,
      "step": 5000
    },
    {
      "epoch": 8.460732984293193,
      "grad_norm": 1.5191736221313477,
      "learning_rate": 6.952000000000001e-05,
      "loss": 0.355,
      "step": 5050
    },
    {
      "epoch": 8.544502617801047,
      "grad_norm": 2.3525736331939697,
      "learning_rate": 6.902000000000001e-05,
      "loss": 0.3554,
      "step": 5100
    },
    {
      "epoch": 8.6282722513089,
      "grad_norm": 1.3651047945022583,
      "learning_rate": 6.852e-05,
      "loss": 0.36,
      "step": 5150
    },
    {
      "epoch": 8.712041884816754,
      "grad_norm": 1.8448572158813477,
      "learning_rate": 6.802e-05,
      "loss": 0.3535,
      "step": 5200
    },
    {
      "epoch": 8.795811518324607,
      "grad_norm": 1.2485219240188599,
      "learning_rate": 6.752e-05,
      "loss": 0.3564,
      "step": 5250
    },
    {
      "epoch": 8.879581151832461,
      "grad_norm": 1.3560365438461304,
      "learning_rate": 6.702e-05,
      "loss": 0.3536,
      "step": 5300
    },
    {
      "epoch": 8.963350785340314,
      "grad_norm": 1.9484964609146118,
      "learning_rate": 6.652000000000001e-05,
      "loss": 0.3554,
      "step": 5350
    },
    {
      "epoch": 9.047120418848168,
      "grad_norm": 1.1570740938186646,
      "learning_rate": 6.602000000000001e-05,
      "loss": 0.3608,
      "step": 5400
    },
    {
      "epoch": 9.13089005235602,
      "grad_norm": 2.5024497509002686,
      "learning_rate": 6.552000000000001e-05,
      "loss": 0.3519,
      "step": 5450
    },
    {
      "epoch": 9.214659685863875,
      "grad_norm": 1.2458699941635132,
      "learning_rate": 6.502e-05,
      "loss": 0.3523,
      "step": 5500
    },
    {
      "epoch": 9.214659685863875,
      "eval_loss": 0.32777872681617737,
      "eval_runtime": 273.5101,
      "eval_samples_per_second": 31.037,
      "eval_steps_per_second": 3.883,
      "step": 5500
    },
    {
      "epoch": 9.298429319371728,
      "grad_norm": 1.3453543186187744,
      "learning_rate": 6.452e-05,
      "loss": 0.3534,
      "step": 5550
    },
    {
      "epoch": 9.38219895287958,
      "grad_norm": 1.2104579210281372,
      "learning_rate": 6.402e-05,
      "loss": 0.3553,
      "step": 5600
    },
    {
      "epoch": 9.465968586387435,
      "grad_norm": 1.1557406187057495,
      "learning_rate": 6.352e-05,
      "loss": 0.3514,
      "step": 5650
    },
    {
      "epoch": 9.549738219895287,
      "grad_norm": 1.621117353439331,
      "learning_rate": 6.302e-05,
      "loss": 0.3542,
      "step": 5700
    },
    {
      "epoch": 9.633507853403142,
      "grad_norm": 0.9746065139770508,
      "learning_rate": 6.252e-05,
      "loss": 0.3538,
      "step": 5750
    },
    {
      "epoch": 9.717277486910994,
      "grad_norm": 1.445143222808838,
      "learning_rate": 6.202e-05,
      "loss": 0.3536,
      "step": 5800
    },
    {
      "epoch": 9.801047120418849,
      "grad_norm": 1.8927617073059082,
      "learning_rate": 6.152e-05,
      "loss": 0.3532,
      "step": 5850
    },
    {
      "epoch": 9.884816753926701,
      "grad_norm": 1.3752628564834595,
      "learning_rate": 6.102e-05,
      "loss": 0.3479,
      "step": 5900
    },
    {
      "epoch": 9.968586387434556,
      "grad_norm": 1.8939998149871826,
      "learning_rate": 6.0519999999999997e-05,
      "loss": 0.3533,
      "step": 5950
    },
    {
      "epoch": 10.052356020942408,
      "grad_norm": 1.564810037612915,
      "learning_rate": 6.002e-05,
      "loss": 0.3484,
      "step": 6000
    },
    {
      "epoch": 10.052356020942408,
      "eval_loss": 0.3196151554584503,
      "eval_runtime": 270.4466,
      "eval_samples_per_second": 31.389,
      "eval_steps_per_second": 3.927,
      "step": 6000
    },
    {
      "epoch": 10.136125654450261,
      "grad_norm": 1.3735418319702148,
      "learning_rate": 5.952e-05,
      "loss": 0.3504,
      "step": 6050
    },
    {
      "epoch": 10.219895287958115,
      "grad_norm": 1.1693004369735718,
      "learning_rate": 5.902e-05,
      "loss": 0.3535,
      "step": 6100
    },
    {
      "epoch": 10.303664921465968,
      "grad_norm": 1.23040771484375,
      "learning_rate": 5.852000000000001e-05,
      "loss": 0.3511,
      "step": 6150
    },
    {
      "epoch": 10.387434554973822,
      "grad_norm": 1.6076780557632446,
      "learning_rate": 5.802000000000001e-05,
      "loss": 0.35,
      "step": 6200
    },
    {
      "epoch": 10.471204188481675,
      "grad_norm": 0.9413685202598572,
      "learning_rate": 5.7520000000000005e-05,
      "loss": 0.3492,
      "step": 6250
    },
    {
      "epoch": 10.55497382198953,
      "grad_norm": 1.239015817642212,
      "learning_rate": 5.7020000000000006e-05,
      "loss": 0.3489,
      "step": 6300
    },
    {
      "epoch": 10.638743455497382,
      "grad_norm": 0.9552774429321289,
      "learning_rate": 5.652000000000001e-05,
      "loss": 0.3491,
      "step": 6350
    },
    {
      "epoch": 10.722513089005236,
      "grad_norm": 1.378158450126648,
      "learning_rate": 5.602000000000001e-05,
      "loss": 0.3492,
      "step": 6400
    },
    {
      "epoch": 10.806282722513089,
      "grad_norm": 1.4036709070205688,
      "learning_rate": 5.5520000000000004e-05,
      "loss": 0.3464,
      "step": 6450
    },
    {
      "epoch": 10.890052356020943,
      "grad_norm": 1.134458303451538,
      "learning_rate": 5.5020000000000005e-05,
      "loss": 0.3485,
      "step": 6500
    },
    {
      "epoch": 10.890052356020943,
      "eval_loss": 0.3166483938694,
      "eval_runtime": 270.1007,
      "eval_samples_per_second": 31.429,
      "eval_steps_per_second": 3.932,
      "step": 6500
    },
    {
      "epoch": 10.973821989528796,
      "grad_norm": 1.485079288482666,
      "learning_rate": 5.4520000000000007e-05,
      "loss": 0.3473,
      "step": 6550
    },
    {
      "epoch": 11.057591623036648,
      "grad_norm": 1.483330488204956,
      "learning_rate": 5.402e-05,
      "loss": 0.3501,
      "step": 6600
    },
    {
      "epoch": 11.141361256544503,
      "grad_norm": 2.135457992553711,
      "learning_rate": 5.352e-05,
      "loss": 0.3496,
      "step": 6650
    },
    {
      "epoch": 11.225130890052355,
      "grad_norm": 1.0900447368621826,
      "learning_rate": 5.3020000000000004e-05,
      "loss": 0.3452,
      "step": 6700
    },
    {
      "epoch": 11.30890052356021,
      "grad_norm": 2.021146297454834,
      "learning_rate": 5.2520000000000005e-05,
      "loss": 0.3462,
      "step": 6750
    },
    {
      "epoch": 11.392670157068062,
      "grad_norm": 1.9676167964935303,
      "learning_rate": 5.202e-05,
      "loss": 0.3472,
      "step": 6800
    },
    {
      "epoch": 11.476439790575917,
      "grad_norm": 1.6358522176742554,
      "learning_rate": 5.152e-05,
      "loss": 0.3458,
      "step": 6850
    },
    {
      "epoch": 11.56020942408377,
      "grad_norm": 1.9463181495666504,
      "learning_rate": 5.102e-05,
      "loss": 0.3452,
      "step": 6900
    },
    {
      "epoch": 11.643979057591624,
      "grad_norm": 1.2650079727172852,
      "learning_rate": 5.052e-05,
      "loss": 0.343,
      "step": 6950
    },
    {
      "epoch": 11.727748691099476,
      "grad_norm": 1.8479430675506592,
      "learning_rate": 5.002e-05,
      "loss": 0.3468,
      "step": 7000
    },
    {
      "epoch": 11.727748691099476,
      "eval_loss": 0.3152253031730652,
      "eval_runtime": 277.1498,
      "eval_samples_per_second": 30.63,
      "eval_steps_per_second": 3.832,
      "step": 7000
    },
    {
      "epoch": 11.81151832460733,
      "grad_norm": 1.1885603666305542,
      "learning_rate": 4.952e-05,
      "loss": 0.3461,
      "step": 7050
    },
    {
      "epoch": 11.895287958115183,
      "grad_norm": 1.5886296033859253,
      "learning_rate": 4.902e-05,
      "loss": 0.348,
      "step": 7100
    },
    {
      "epoch": 11.979057591623036,
      "grad_norm": 1.7292935848236084,
      "learning_rate": 4.852e-05,
      "loss": 0.3431,
      "step": 7150
    },
    {
      "epoch": 12.06282722513089,
      "grad_norm": 1.188632845878601,
      "learning_rate": 4.8030000000000006e-05,
      "loss": 0.3531,
      "step": 7200
    },
    {
      "epoch": 12.146596858638743,
      "grad_norm": 1.0749746561050415,
      "learning_rate": 4.753e-05,
      "loss": 0.3446,
      "step": 7250
    },
    {
      "epoch": 12.230366492146597,
      "grad_norm": 1.6439704895019531,
      "learning_rate": 4.703e-05,
      "loss": 0.3436,
      "step": 7300
    },
    {
      "epoch": 12.31413612565445,
      "grad_norm": 1.0988739728927612,
      "learning_rate": 4.6530000000000003e-05,
      "loss": 0.344,
      "step": 7350
    },
    {
      "epoch": 12.397905759162304,
      "grad_norm": 1.7305316925048828,
      "learning_rate": 4.603e-05,
      "loss": 0.3485,
      "step": 7400
    },
    {
      "epoch": 12.481675392670157,
      "grad_norm": 1.0579546689987183,
      "learning_rate": 4.553e-05,
      "loss": 0.3455,
      "step": 7450
    },
    {
      "epoch": 12.565445026178011,
      "grad_norm": 1.1896628141403198,
      "learning_rate": 4.503e-05,
      "loss": 0.3434,
      "step": 7500
    },
    {
      "epoch": 12.565445026178011,
      "eval_loss": 0.31436365842819214,
      "eval_runtime": 277.7357,
      "eval_samples_per_second": 30.565,
      "eval_steps_per_second": 3.824,
      "step": 7500
    },
    {
      "epoch": 12.649214659685864,
      "grad_norm": 2.54182505607605,
      "learning_rate": 4.453e-05,
      "loss": 0.3425,
      "step": 7550
    },
    {
      "epoch": 12.732984293193716,
      "grad_norm": 0.9473859071731567,
      "learning_rate": 4.4030000000000004e-05,
      "loss": 0.3464,
      "step": 7600
    },
    {
      "epoch": 12.81675392670157,
      "grad_norm": 1.2491322755813599,
      "learning_rate": 4.3530000000000005e-05,
      "loss": 0.3473,
      "step": 7650
    },
    {
      "epoch": 12.900523560209423,
      "grad_norm": 1.4746623039245605,
      "learning_rate": 4.3030000000000006e-05,
      "loss": 0.3446,
      "step": 7700
    },
    {
      "epoch": 12.984293193717278,
      "grad_norm": 1.4437618255615234,
      "learning_rate": 4.253e-05,
      "loss": 0.3476,
      "step": 7750
    },
    {
      "epoch": 13.06806282722513,
      "grad_norm": 1.1064411401748657,
      "learning_rate": 4.203e-05,
      "loss": 0.3404,
      "step": 7800
    },
    {
      "epoch": 13.151832460732985,
      "grad_norm": 1.2159862518310547,
      "learning_rate": 4.1530000000000004e-05,
      "loss": 0.3432,
      "step": 7850
    },
    {
      "epoch": 13.235602094240837,
      "grad_norm": 1.931052327156067,
      "learning_rate": 4.103e-05,
      "loss": 0.3449,
      "step": 7900
    },
    {
      "epoch": 13.319371727748692,
      "grad_norm": 1.412427544593811,
      "learning_rate": 4.053e-05,
      "loss": 0.343,
      "step": 7950
    },
    {
      "epoch": 13.403141361256544,
      "grad_norm": 1.385571837425232,
      "learning_rate": 4.003e-05,
      "loss": 0.3417,
      "step": 8000
    },
    {
      "epoch": 13.403141361256544,
      "eval_loss": 0.3115607798099518,
      "eval_runtime": 278.4935,
      "eval_samples_per_second": 30.482,
      "eval_steps_per_second": 3.813,
      "step": 8000
    },
    {
      "epoch": 13.486910994764397,
      "grad_norm": 1.0727407932281494,
      "learning_rate": 3.953e-05,
      "loss": 0.3401,
      "step": 8050
    },
    {
      "epoch": 13.570680628272251,
      "grad_norm": 1.1382709741592407,
      "learning_rate": 3.903e-05,
      "loss": 0.3423,
      "step": 8100
    },
    {
      "epoch": 13.654450261780104,
      "grad_norm": 0.8184741139411926,
      "learning_rate": 3.853e-05,
      "loss": 0.3414,
      "step": 8150
    },
    {
      "epoch": 13.738219895287958,
      "grad_norm": 1.5440157651901245,
      "learning_rate": 3.803000000000001e-05,
      "loss": 0.3405,
      "step": 8200
    },
    {
      "epoch": 13.821989528795811,
      "grad_norm": 1.2176579236984253,
      "learning_rate": 3.753e-05,
      "loss": 0.343,
      "step": 8250
    },
    {
      "epoch": 13.905759162303665,
      "grad_norm": 1.1099356412887573,
      "learning_rate": 3.703e-05,
      "loss": 0.3391,
      "step": 8300
    },
    {
      "epoch": 13.989528795811518,
      "grad_norm": 1.5049606561660767,
      "learning_rate": 3.6530000000000004e-05,
      "loss": 0.3406,
      "step": 8350
    },
    {
      "epoch": 14.073298429319372,
      "grad_norm": 1.3164118528366089,
      "learning_rate": 3.6030000000000006e-05,
      "loss": 0.3456,
      "step": 8400
    },
    {
      "epoch": 14.157068062827225,
      "grad_norm": 1.478460669517517,
      "learning_rate": 3.553e-05,
      "loss": 0.338,
      "step": 8450
    },
    {
      "epoch": 14.24083769633508,
      "grad_norm": 1.039398193359375,
      "learning_rate": 3.503e-05,
      "loss": 0.3402,
      "step": 8500
    },
    {
      "epoch": 14.24083769633508,
      "eval_loss": 0.31324246525764465,
      "eval_runtime": 274.0353,
      "eval_samples_per_second": 30.978,
      "eval_steps_per_second": 3.875,
      "step": 8500
    },
    {
      "epoch": 14.324607329842932,
      "grad_norm": 1.5117567777633667,
      "learning_rate": 3.453e-05,
      "loss": 0.3416,
      "step": 8550
    },
    {
      "epoch": 14.408376963350785,
      "grad_norm": 1.1644045114517212,
      "learning_rate": 3.403e-05,
      "loss": 0.3411,
      "step": 8600
    },
    {
      "epoch": 14.492146596858639,
      "grad_norm": 1.09773850440979,
      "learning_rate": 3.353e-05,
      "loss": 0.3398,
      "step": 8650
    },
    {
      "epoch": 14.575916230366492,
      "grad_norm": 1.1257084608078003,
      "learning_rate": 3.303e-05,
      "loss": 0.3398,
      "step": 8700
    },
    {
      "epoch": 14.659685863874346,
      "grad_norm": 1.0074257850646973,
      "learning_rate": 3.253e-05,
      "loss": 0.3375,
      "step": 8750
    },
    {
      "epoch": 14.743455497382199,
      "grad_norm": 2.0023674964904785,
      "learning_rate": 3.2029999999999997e-05,
      "loss": 0.3411,
      "step": 8800
    },
    {
      "epoch": 14.827225130890053,
      "grad_norm": 2.224597930908203,
      "learning_rate": 3.1530000000000005e-05,
      "loss": 0.3427,
      "step": 8850
    },
    {
      "epoch": 14.910994764397905,
      "grad_norm": 1.1201759576797485,
      "learning_rate": 3.1030000000000006e-05,
      "loss": 0.3398,
      "step": 8900
    },
    {
      "epoch": 14.99476439790576,
      "grad_norm": 1.0911871194839478,
      "learning_rate": 3.053e-05,
      "loss": 0.3397,
      "step": 8950
    },
    {
      "epoch": 15.078534031413612,
      "grad_norm": 1.871634602546692,
      "learning_rate": 3.0030000000000002e-05,
      "loss": 0.3407,
      "step": 9000
    },
    {
      "epoch": 15.078534031413612,
      "eval_loss": 0.3115617036819458,
      "eval_runtime": 273.4516,
      "eval_samples_per_second": 31.044,
      "eval_steps_per_second": 3.884,
      "step": 9000
    },
    {
      "epoch": 15.162303664921467,
      "grad_norm": 1.8591896295547485,
      "learning_rate": 2.9530000000000004e-05,
      "loss": 0.3392,
      "step": 9050
    },
    {
      "epoch": 15.24607329842932,
      "grad_norm": 0.8903402090072632,
      "learning_rate": 2.903e-05,
      "loss": 0.3376,
      "step": 9100
    },
    {
      "epoch": 15.329842931937172,
      "grad_norm": 1.0485785007476807,
      "learning_rate": 2.853e-05,
      "loss": 0.3365,
      "step": 9150
    },
    {
      "epoch": 15.413612565445026,
      "grad_norm": 1.551017165184021,
      "learning_rate": 2.803e-05,
      "loss": 0.3372,
      "step": 9200
    },
    {
      "epoch": 15.497382198952879,
      "grad_norm": 0.8951018452644348,
      "learning_rate": 2.753e-05,
      "loss": 0.3397,
      "step": 9250
    },
    {
      "epoch": 15.581151832460733,
      "grad_norm": 1.5708497762680054,
      "learning_rate": 2.703e-05,
      "loss": 0.3416,
      "step": 9300
    },
    {
      "epoch": 15.664921465968586,
      "grad_norm": 2.1250946521759033,
      "learning_rate": 2.653e-05,
      "loss": 0.3385,
      "step": 9350
    },
    {
      "epoch": 15.74869109947644,
      "grad_norm": 1.4175411462783813,
      "learning_rate": 2.603e-05,
      "loss": 0.3358,
      "step": 9400
    },
    {
      "epoch": 15.832460732984293,
      "grad_norm": 1.6622357368469238,
      "learning_rate": 2.5530000000000005e-05,
      "loss": 0.3387,
      "step": 9450
    },
    {
      "epoch": 15.916230366492147,
      "grad_norm": 2.2202370166778564,
      "learning_rate": 2.504e-05,
      "loss": 0.3387,
      "step": 9500
    },
    {
      "epoch": 15.916230366492147,
      "eval_loss": 0.30681997537612915,
      "eval_runtime": 274.6544,
      "eval_samples_per_second": 30.908,
      "eval_steps_per_second": 3.867,
      "step": 9500
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.8703222870826721,
      "learning_rate": 2.4540000000000002e-05,
      "loss": 0.339,
      "step": 9550
    },
    {
      "epoch": 16.083769633507853,
      "grad_norm": 0.8840942978858948,
      "learning_rate": 2.404e-05,
      "loss": 0.3358,
      "step": 9600
    },
    {
      "epoch": 16.167539267015705,
      "grad_norm": 1.5786060094833374,
      "learning_rate": 2.354e-05,
      "loss": 0.3375,
      "step": 9650
    },
    {
      "epoch": 16.25130890052356,
      "grad_norm": 0.7372541427612305,
      "learning_rate": 2.304e-05,
      "loss": 0.3351,
      "step": 9700
    },
    {
      "epoch": 16.335078534031414,
      "grad_norm": 1.6219487190246582,
      "learning_rate": 2.254e-05,
      "loss": 0.3348,
      "step": 9750
    },
    {
      "epoch": 16.418848167539267,
      "grad_norm": 1.5188941955566406,
      "learning_rate": 2.2040000000000002e-05,
      "loss": 0.341,
      "step": 9800
    },
    {
      "epoch": 16.50261780104712,
      "grad_norm": 1.0383074283599854,
      "learning_rate": 2.154e-05,
      "loss": 0.3409,
      "step": 9850
    },
    {
      "epoch": 16.586387434554975,
      "grad_norm": 1.1702338457107544,
      "learning_rate": 2.1040000000000002e-05,
      "loss": 0.3353,
      "step": 9900
    },
    {
      "epoch": 16.670157068062828,
      "grad_norm": 1.9785561561584473,
      "learning_rate": 2.054e-05,
      "loss": 0.3389,
      "step": 9950
    },
    {
      "epoch": 16.75392670157068,
      "grad_norm": 0.9374827742576599,
      "learning_rate": 2.004e-05,
      "loss": 0.3375,
      "step": 10000
    },
    {
      "epoch": 16.75392670157068,
      "eval_loss": 0.3070985674858093,
      "eval_runtime": 272.242,
      "eval_samples_per_second": 31.182,
      "eval_steps_per_second": 3.901,
      "step": 10000
    }
  ],
  "logging_steps": 50,
  "max_steps": 12000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 21,
  "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": 1.7975109075829805e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}