{
  "best_metric": 0.3439213037490845,
  "best_model_checkpoint": "mikhail_panzo/zlm_b64_le4_s12000/checkpoint-3500",
  "epoch": 2.932565445026178,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.041884816753926704,
      "grad_norm": 2.697540283203125,
      "learning_rate": 2.4000000000000003e-06,
      "loss": 1.0526,
      "step": 50
    },
    {
      "epoch": 0.08376963350785341,
      "grad_norm": 4.0351996421813965,
      "learning_rate": 4.9000000000000005e-06,
      "loss": 0.8484,
      "step": 100
    },
    {
      "epoch": 0.1256544502617801,
      "grad_norm": 4.858840465545654,
      "learning_rate": 7.4e-06,
      "loss": 0.7505,
      "step": 150
    },
    {
      "epoch": 0.16753926701570682,
      "grad_norm": 2.396157741546631,
      "learning_rate": 9.900000000000002e-06,
      "loss": 0.6889,
      "step": 200
    },
    {
      "epoch": 0.2094240837696335,
      "grad_norm": 2.660891056060791,
      "learning_rate": 1.24e-05,
      "loss": 0.6363,
      "step": 250
    },
    {
      "epoch": 0.2513089005235602,
      "grad_norm": 2.848365068435669,
      "learning_rate": 1.49e-05,
      "loss": 0.6107,
      "step": 300
    },
    {
      "epoch": 0.2931937172774869,
      "grad_norm": 2.0031681060791016,
      "learning_rate": 1.74e-05,
      "loss": 0.5673,
      "step": 350
    },
    {
      "epoch": 0.33507853403141363,
      "grad_norm": 4.771584510803223,
      "learning_rate": 1.9900000000000003e-05,
      "loss": 0.5566,
      "step": 400
    },
    {
      "epoch": 0.3769633507853403,
      "grad_norm": 9.876974105834961,
      "learning_rate": 2.2400000000000002e-05,
      "loss": 0.5487,
      "step": 450
    },
    {
      "epoch": 0.418848167539267,
      "grad_norm": 2.113546848297119,
      "learning_rate": 2.4900000000000002e-05,
      "loss": 0.5268,
      "step": 500
    },
    {
      "epoch": 0.418848167539267,
      "eval_loss": 0.46557939052581787,
      "eval_runtime": 297.9193,
      "eval_samples_per_second": 28.494,
      "eval_steps_per_second": 3.565,
      "step": 500
    },
    {
      "epoch": 0.4607329842931937,
      "grad_norm": 2.7003655433654785,
      "learning_rate": 2.7400000000000002e-05,
      "loss": 0.5212,
      "step": 550
    },
    {
      "epoch": 0.5026178010471204,
      "grad_norm": 5.649191379547119,
      "learning_rate": 2.9900000000000002e-05,
      "loss": 0.5138,
      "step": 600
    },
    {
      "epoch": 0.5445026178010471,
      "grad_norm": 2.272181749343872,
      "learning_rate": 3.24e-05,
      "loss": 0.5121,
      "step": 650
    },
    {
      "epoch": 0.5863874345549738,
      "grad_norm": 2.957671642303467,
      "learning_rate": 3.49e-05,
      "loss": 0.496,
      "step": 700
    },
    {
      "epoch": 0.6282722513089005,
      "grad_norm": 3.577693462371826,
      "learning_rate": 3.74e-05,
      "loss": 0.4886,
      "step": 750
    },
    {
      "epoch": 0.6701570680628273,
      "grad_norm": 2.466113328933716,
      "learning_rate": 3.99e-05,
      "loss": 0.4879,
      "step": 800
    },
    {
      "epoch": 0.7120418848167539,
      "grad_norm": 2.941208600997925,
      "learning_rate": 4.24e-05,
      "loss": 0.4865,
      "step": 850
    },
    {
      "epoch": 0.7539267015706806,
      "grad_norm": 2.3843142986297607,
      "learning_rate": 4.49e-05,
      "loss": 0.4689,
      "step": 900
    },
    {
      "epoch": 0.7958115183246073,
      "grad_norm": 2.232208728790283,
      "learning_rate": 4.74e-05,
      "loss": 0.4768,
      "step": 950
    },
    {
      "epoch": 0.837696335078534,
      "grad_norm": 2.1702969074249268,
      "learning_rate": 4.99e-05,
      "loss": 0.4583,
      "step": 1000
    },
    {
      "epoch": 0.837696335078534,
      "eval_loss": 0.4044812321662903,
      "eval_runtime": 296.3081,
      "eval_samples_per_second": 28.649,
      "eval_steps_per_second": 3.584,
      "step": 1000
    },
    {
      "epoch": 0.8795811518324608,
      "grad_norm": 2.7613942623138428,
      "learning_rate": 5.2400000000000007e-05,
      "loss": 0.4664,
      "step": 1050
    },
    {
      "epoch": 0.9214659685863874,
      "grad_norm": 2.4752321243286133,
      "learning_rate": 5.4900000000000006e-05,
      "loss": 0.442,
      "step": 1100
    },
    {
      "epoch": 0.9633507853403142,
      "grad_norm": 2.27407169342041,
      "learning_rate": 5.74e-05,
      "loss": 0.4498,
      "step": 1150
    },
    {
      "epoch": 1.0052356020942408,
      "grad_norm": 1.8998862504959106,
      "learning_rate": 5.99e-05,
      "loss": 0.4433,
      "step": 1200
    },
    {
      "epoch": 1.0471204188481675,
      "grad_norm": 2.8479738235473633,
      "learning_rate": 6.24e-05,
      "loss": 0.4417,
      "step": 1250
    },
    {
      "epoch": 1.0890052356020943,
      "grad_norm": 1.9095953702926636,
      "learning_rate": 6.49e-05,
      "loss": 0.444,
      "step": 1300
    },
    {
      "epoch": 1.130890052356021,
      "grad_norm": 1.368077278137207,
      "learning_rate": 6.740000000000001e-05,
      "loss": 0.4519,
      "step": 1350
    },
    {
      "epoch": 1.1727748691099475,
      "grad_norm": 2.0456721782684326,
      "learning_rate": 6.99e-05,
      "loss": 0.4411,
      "step": 1400
    },
    {
      "epoch": 1.2146596858638743,
      "grad_norm": 1.6217619180679321,
      "learning_rate": 7.24e-05,
      "loss": 0.4324,
      "step": 1450
    },
    {
      "epoch": 1.256544502617801,
      "grad_norm": 2.961646795272827,
      "learning_rate": 7.49e-05,
      "loss": 0.4321,
      "step": 1500
    },
    {
      "epoch": 1.256544502617801,
      "eval_loss": 0.3930450975894928,
      "eval_runtime": 301.1428,
      "eval_samples_per_second": 28.189,
      "eval_steps_per_second": 3.527,
      "step": 1500
    },
    {
      "epoch": 1.2984293193717278,
      "grad_norm": 2.543201446533203,
      "learning_rate": 7.740000000000001e-05,
      "loss": 0.4364,
      "step": 1550
    },
    {
      "epoch": 1.3403141361256545,
      "grad_norm": 1.5807560682296753,
      "learning_rate": 7.99e-05,
      "loss": 0.433,
      "step": 1600
    },
    {
      "epoch": 1.3821989528795813,
      "grad_norm": 1.587385892868042,
      "learning_rate": 8.24e-05,
      "loss": 0.4285,
      "step": 1650
    },
    {
      "epoch": 1.4240837696335078,
      "grad_norm": 2.1177687644958496,
      "learning_rate": 8.49e-05,
      "loss": 0.4209,
      "step": 1700
    },
    {
      "epoch": 1.4659685863874345,
      "grad_norm": 1.6078702211380005,
      "learning_rate": 8.740000000000001e-05,
      "loss": 0.426,
      "step": 1750
    },
    {
      "epoch": 1.5078534031413613,
      "grad_norm": 2.10492205619812,
      "learning_rate": 8.99e-05,
      "loss": 0.4191,
      "step": 1800
    },
    {
      "epoch": 1.5497382198952878,
      "grad_norm": 2.169309616088867,
      "learning_rate": 9.240000000000001e-05,
      "loss": 0.4178,
      "step": 1850
    },
    {
      "epoch": 1.5916230366492146,
      "grad_norm": 1.559006929397583,
      "learning_rate": 9.49e-05,
      "loss": 0.4133,
      "step": 1900
    },
    {
      "epoch": 1.6335078534031413,
      "grad_norm": 2.0436346530914307,
      "learning_rate": 9.74e-05,
      "loss": 0.4234,
      "step": 1950
    },
    {
      "epoch": 1.675392670157068,
      "grad_norm": 2.089486837387085,
      "learning_rate": 9.99e-05,
      "loss": 0.4092,
      "step": 2000
    },
    {
      "epoch": 1.675392670157068,
      "eval_loss": 0.37686464190483093,
      "eval_runtime": 310.3113,
      "eval_samples_per_second": 27.356,
      "eval_steps_per_second": 3.422,
      "step": 2000
    },
    {
      "epoch": 1.7179057591623037,
      "grad_norm": 2.684563398361206,
      "learning_rate": 9.952e-05,
      "loss": 0.4099,
      "step": 2050
    },
    {
      "epoch": 1.7597905759162304,
      "grad_norm": 1.7815254926681519,
      "learning_rate": 9.902e-05,
      "loss": 0.4093,
      "step": 2100
    },
    {
      "epoch": 1.8016753926701572,
      "grad_norm": 1.9900892972946167,
      "learning_rate": 9.852e-05,
      "loss": 0.4116,
      "step": 2150
    },
    {
      "epoch": 1.8435602094240837,
      "grad_norm": 1.8637460470199585,
      "learning_rate": 9.802e-05,
      "loss": 0.4089,
      "step": 2200
    },
    {
      "epoch": 1.8854450261780105,
      "grad_norm": 1.58171546459198,
      "learning_rate": 9.752e-05,
      "loss": 0.4053,
      "step": 2250
    },
    {
      "epoch": 1.9273298429319372,
      "grad_norm": 1.4995630979537964,
      "learning_rate": 9.702e-05,
      "loss": 0.4049,
      "step": 2300
    },
    {
      "epoch": 1.9692146596858637,
      "grad_norm": 2.447049379348755,
      "learning_rate": 9.652e-05,
      "loss": 0.4029,
      "step": 2350
    },
    {
      "epoch": 2.0110994764397905,
      "grad_norm": 1.8776074647903442,
      "learning_rate": 9.602e-05,
      "loss": 0.3988,
      "step": 2400
    },
    {
      "epoch": 2.0529842931937172,
      "grad_norm": 3.1785056591033936,
      "learning_rate": 9.552000000000001e-05,
      "loss": 0.4023,
      "step": 2450
    },
    {
      "epoch": 2.094869109947644,
      "grad_norm": 1.4416707754135132,
      "learning_rate": 9.502000000000001e-05,
      "loss": 0.4008,
      "step": 2500
    },
    {
      "epoch": 2.094869109947644,
      "eval_loss": 0.35975897312164307,
      "eval_runtime": 275.2855,
      "eval_samples_per_second": 30.837,
      "eval_steps_per_second": 3.858,
      "step": 2500
    },
    {
      "epoch": 2.1367539267015707,
      "grad_norm": 1.3332245349884033,
      "learning_rate": 9.452000000000001e-05,
      "loss": 0.3985,
      "step": 2550
    },
    {
      "epoch": 2.1786387434554975,
      "grad_norm": 1.291338324546814,
      "learning_rate": 9.402000000000001e-05,
      "loss": 0.4021,
      "step": 2600
    },
    {
      "epoch": 2.220523560209424,
      "grad_norm": 1.6172423362731934,
      "learning_rate": 9.352000000000001e-05,
      "loss": 0.4006,
      "step": 2650
    },
    {
      "epoch": 2.262408376963351,
      "grad_norm": 1.8729867935180664,
      "learning_rate": 9.302e-05,
      "loss": 0.3966,
      "step": 2700
    },
    {
      "epoch": 2.3042931937172773,
      "grad_norm": 1.623901605606079,
      "learning_rate": 9.252e-05,
      "loss": 0.3951,
      "step": 2750
    },
    {
      "epoch": 2.346178010471204,
      "grad_norm": 1.032320261001587,
      "learning_rate": 9.202e-05,
      "loss": 0.3904,
      "step": 2800
    },
    {
      "epoch": 2.3880628272251307,
      "grad_norm": 2.0027778148651123,
      "learning_rate": 9.152e-05,
      "loss": 0.3933,
      "step": 2850
    },
    {
      "epoch": 2.4299476439790575,
      "grad_norm": 1.5681087970733643,
      "learning_rate": 9.102e-05,
      "loss": 0.4024,
      "step": 2900
    },
    {
      "epoch": 2.4718324607329842,
      "grad_norm": 1.2643622159957886,
      "learning_rate": 9.052000000000001e-05,
      "loss": 0.3939,
      "step": 2950
    },
    {
      "epoch": 2.513717277486911,
      "grad_norm": 1.4029465913772583,
      "learning_rate": 9.002000000000001e-05,
      "loss": 0.3941,
      "step": 3000
    },
    {
      "epoch": 2.513717277486911,
      "eval_loss": 0.3533318042755127,
      "eval_runtime": 271.9646,
      "eval_samples_per_second": 31.214,
      "eval_steps_per_second": 3.905,
      "step": 3000
    },
    {
      "epoch": 2.5556020942408377,
      "grad_norm": 1.6669938564300537,
      "learning_rate": 8.952000000000001e-05,
      "loss": 0.3908,
      "step": 3050
    },
    {
      "epoch": 2.5974869109947645,
      "grad_norm": 1.6020203828811646,
      "learning_rate": 8.902e-05,
      "loss": 0.3905,
      "step": 3100
    },
    {
      "epoch": 2.639371727748691,
      "grad_norm": 2.2321176528930664,
      "learning_rate": 8.852e-05,
      "loss": 0.389,
      "step": 3150
    },
    {
      "epoch": 2.681256544502618,
      "grad_norm": 1.0864609479904175,
      "learning_rate": 8.802e-05,
      "loss": 0.3906,
      "step": 3200
    },
    {
      "epoch": 2.7231413612565447,
      "grad_norm": 2.0679004192352295,
      "learning_rate": 8.752e-05,
      "loss": 0.3869,
      "step": 3250
    },
    {
      "epoch": 2.765026178010471,
      "grad_norm": 2.6636288166046143,
      "learning_rate": 8.702e-05,
      "loss": 0.3841,
      "step": 3300
    },
    {
      "epoch": 2.8069109947643978,
      "grad_norm": 1.435549259185791,
      "learning_rate": 8.652e-05,
      "loss": 0.3841,
      "step": 3350
    },
    {
      "epoch": 2.8487958115183245,
      "grad_norm": 2.208448648452759,
      "learning_rate": 8.602e-05,
      "loss": 0.3828,
      "step": 3400
    },
    {
      "epoch": 2.8906806282722513,
      "grad_norm": 1.731292963027954,
      "learning_rate": 8.552e-05,
      "loss": 0.3842,
      "step": 3450
    },
    {
      "epoch": 2.932565445026178,
      "grad_norm": 1.7065523862838745,
      "learning_rate": 8.502e-05,
      "loss": 0.3834,
      "step": 3500
    },
    {
      "epoch": 2.932565445026178,
      "eval_loss": 0.3439213037490845,
      "eval_runtime": 274.3567,
      "eval_samples_per_second": 30.941,
      "eval_steps_per_second": 3.871,
      "step": 3500
    }
  ],
  "logging_steps": 50,
  "max_steps": 12000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 11,
  "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": 3.1412972145703104e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}