{
  "best_metric": 0.395076185464859,
  "best_model_checkpoint": "mikhail-panzo/ceb_b128_le4_s4000/checkpoint-1500",
  "epoch": 316.83168316831683,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 3.9603960396039604,
      "grad_norm": 2.952680826187134,
      "learning_rate": 2.5e-06,
      "loss": 0.7707,
      "step": 50
    },
    {
      "epoch": 7.920792079207921,
      "grad_norm": 1.126000165939331,
      "learning_rate": 5e-06,
      "loss": 0.6934,
      "step": 100
    },
    {
      "epoch": 11.881188118811881,
      "grad_norm": 2.129329204559326,
      "learning_rate": 7.5e-06,
      "loss": 0.5982,
      "step": 150
    },
    {
      "epoch": 15.841584158415841,
      "grad_norm": 1.1155015230178833,
      "learning_rate": 1e-05,
      "loss": 0.5093,
      "step": 200
    },
    {
      "epoch": 19.801980198019802,
      "grad_norm": 1.3643360137939453,
      "learning_rate": 1.25e-05,
      "loss": 0.4912,
      "step": 250
    },
    {
      "epoch": 23.762376237623762,
      "grad_norm": 0.9774149060249329,
      "learning_rate": 1.5e-05,
      "loss": 0.4764,
      "step": 300
    },
    {
      "epoch": 27.722772277227723,
      "grad_norm": 0.8793525695800781,
      "learning_rate": 1.75e-05,
      "loss": 0.4648,
      "step": 350
    },
    {
      "epoch": 31.683168316831683,
      "grad_norm": 0.7882798910140991,
      "learning_rate": 2e-05,
      "loss": 0.459,
      "step": 400
    },
    {
      "epoch": 35.64356435643565,
      "grad_norm": 1.2033296823501587,
      "learning_rate": 2.25e-05,
      "loss": 0.4519,
      "step": 450
    },
    {
      "epoch": 39.603960396039604,
      "grad_norm": 0.8620955348014832,
      "learning_rate": 2.5e-05,
      "loss": 0.4433,
      "step": 500
    },
    {
      "epoch": 39.603960396039604,
      "eval_loss": 0.4056467115879059,
      "eval_runtime": 7.4199,
      "eval_samples_per_second": 24.259,
      "eval_steps_per_second": 3.1,
      "step": 500
    },
    {
      "epoch": 43.56435643564357,
      "grad_norm": 1.266213297843933,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.444,
      "step": 550
    },
    {
      "epoch": 47.524752475247524,
      "grad_norm": 1.174134373664856,
      "learning_rate": 3e-05,
      "loss": 0.437,
      "step": 600
    },
    {
      "epoch": 51.48514851485149,
      "grad_norm": 0.8188551664352417,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.4318,
      "step": 650
    },
    {
      "epoch": 55.445544554455445,
      "grad_norm": 1.3083854913711548,
      "learning_rate": 3.5e-05,
      "loss": 0.4297,
      "step": 700
    },
    {
      "epoch": 59.40594059405941,
      "grad_norm": 0.8094576001167297,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.4238,
      "step": 750
    },
    {
      "epoch": 63.366336633663366,
      "grad_norm": 1.2506986856460571,
      "learning_rate": 4e-05,
      "loss": 0.4233,
      "step": 800
    },
    {
      "epoch": 67.32673267326733,
      "grad_norm": 1.7902193069458008,
      "learning_rate": 4.25e-05,
      "loss": 0.4208,
      "step": 850
    },
    {
      "epoch": 71.2871287128713,
      "grad_norm": 2.209643602371216,
      "learning_rate": 4.5e-05,
      "loss": 0.4195,
      "step": 900
    },
    {
      "epoch": 75.24752475247524,
      "grad_norm": 2.256385087966919,
      "learning_rate": 4.75e-05,
      "loss": 0.4202,
      "step": 950
    },
    {
      "epoch": 79.20792079207921,
      "grad_norm": 0.9167351126670837,
      "learning_rate": 5e-05,
      "loss": 0.4142,
      "step": 1000
    },
    {
      "epoch": 79.20792079207921,
      "eval_loss": 0.39529716968536377,
      "eval_runtime": 6.6198,
      "eval_samples_per_second": 27.191,
      "eval_steps_per_second": 3.474,
      "step": 1000
    },
    {
      "epoch": 83.16831683168317,
      "grad_norm": 0.9256124496459961,
      "learning_rate": 5.25e-05,
      "loss": 0.4137,
      "step": 1050
    },
    {
      "epoch": 87.12871287128714,
      "grad_norm": 1.9655492305755615,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.4132,
      "step": 1100
    },
    {
      "epoch": 91.08910891089108,
      "grad_norm": 1.08187997341156,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.4078,
      "step": 1150
    },
    {
      "epoch": 95.04950495049505,
      "grad_norm": 1.8732513189315796,
      "learning_rate": 6e-05,
      "loss": 0.4061,
      "step": 1200
    },
    {
      "epoch": 99.00990099009901,
      "grad_norm": 1.3401920795440674,
      "learning_rate": 6.25e-05,
      "loss": 0.4054,
      "step": 1250
    },
    {
      "epoch": 102.97029702970298,
      "grad_norm": 0.8879286646842957,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.4046,
      "step": 1300
    },
    {
      "epoch": 106.93069306930693,
      "grad_norm": 1.2213553190231323,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.3999,
      "step": 1350
    },
    {
      "epoch": 110.89108910891089,
      "grad_norm": 2.05886173248291,
      "learning_rate": 7e-05,
      "loss": 0.3983,
      "step": 1400
    },
    {
      "epoch": 114.85148514851485,
      "grad_norm": 1.10814368724823,
      "learning_rate": 7.25e-05,
      "loss": 0.3964,
      "step": 1450
    },
    {
      "epoch": 118.81188118811882,
      "grad_norm": 1.0220105648040771,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.3972,
      "step": 1500
    },
    {
      "epoch": 118.81188118811882,
      "eval_loss": 0.395076185464859,
      "eval_runtime": 7.1134,
      "eval_samples_per_second": 25.304,
      "eval_steps_per_second": 3.233,
      "step": 1500
    },
    {
      "epoch": 122.77227722772277,
      "grad_norm": 1.1671476364135742,
      "learning_rate": 7.75e-05,
      "loss": 0.3909,
      "step": 1550
    },
    {
      "epoch": 126.73267326732673,
      "grad_norm": 0.8167665600776672,
      "learning_rate": 8e-05,
      "loss": 0.3913,
      "step": 1600
    },
    {
      "epoch": 130.69306930693068,
      "grad_norm": 1.9696727991104126,
      "learning_rate": 8.25e-05,
      "loss": 0.3888,
      "step": 1650
    },
    {
      "epoch": 134.65346534653466,
      "grad_norm": 1.1445192098617554,
      "learning_rate": 8.5e-05,
      "loss": 0.3885,
      "step": 1700
    },
    {
      "epoch": 138.6138613861386,
      "grad_norm": 1.908463954925537,
      "learning_rate": 8.75e-05,
      "loss": 0.3888,
      "step": 1750
    },
    {
      "epoch": 142.5742574257426,
      "grad_norm": 1.6358473300933838,
      "learning_rate": 9e-05,
      "loss": 0.3878,
      "step": 1800
    },
    {
      "epoch": 146.53465346534654,
      "grad_norm": 1.7407046556472778,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.3868,
      "step": 1850
    },
    {
      "epoch": 150.4950495049505,
      "grad_norm": 3.984604597091675,
      "learning_rate": 9.5e-05,
      "loss": 0.3877,
      "step": 1900
    },
    {
      "epoch": 154.45544554455446,
      "grad_norm": 1.5218504667282104,
      "learning_rate": 9.75e-05,
      "loss": 0.3828,
      "step": 1950
    },
    {
      "epoch": 158.41584158415841,
      "grad_norm": 1.2710033655166626,
      "learning_rate": 0.0001,
      "loss": 0.3806,
      "step": 2000
    },
    {
      "epoch": 158.41584158415841,
      "eval_loss": 0.39701494574546814,
      "eval_runtime": 7.067,
      "eval_samples_per_second": 25.47,
      "eval_steps_per_second": 3.255,
      "step": 2000
    },
    {
      "epoch": 162.37623762376236,
      "grad_norm": 1.905275583267212,
      "learning_rate": 9.75e-05,
      "loss": 0.3817,
      "step": 2050
    },
    {
      "epoch": 166.33663366336634,
      "grad_norm": 2.0633037090301514,
      "learning_rate": 9.5e-05,
      "loss": 0.3805,
      "step": 2100
    },
    {
      "epoch": 170.2970297029703,
      "grad_norm": 1.110232949256897,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.378,
      "step": 2150
    },
    {
      "epoch": 174.25742574257427,
      "grad_norm": 2.0630099773406982,
      "learning_rate": 9e-05,
      "loss": 0.3742,
      "step": 2200
    },
    {
      "epoch": 178.21782178217822,
      "grad_norm": 1.2983876466751099,
      "learning_rate": 8.75e-05,
      "loss": 0.3745,
      "step": 2250
    },
    {
      "epoch": 182.17821782178217,
      "grad_norm": 0.6691053509712219,
      "learning_rate": 8.5e-05,
      "loss": 0.3764,
      "step": 2300
    },
    {
      "epoch": 186.13861386138615,
      "grad_norm": 1.2052526473999023,
      "learning_rate": 8.25e-05,
      "loss": 0.3719,
      "step": 2350
    },
    {
      "epoch": 190.0990099009901,
      "grad_norm": 1.1967781782150269,
      "learning_rate": 8e-05,
      "loss": 0.3711,
      "step": 2400
    },
    {
      "epoch": 194.05940594059405,
      "grad_norm": 0.8571879267692566,
      "learning_rate": 7.75e-05,
      "loss": 0.3715,
      "step": 2450
    },
    {
      "epoch": 198.01980198019803,
      "grad_norm": 1.2208068370819092,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.3653,
      "step": 2500
    },
    {
      "epoch": 198.01980198019803,
      "eval_loss": 0.3959733843803406,
      "eval_runtime": 6.5933,
      "eval_samples_per_second": 27.3,
      "eval_steps_per_second": 3.488,
      "step": 2500
    },
    {
      "epoch": 201.98019801980197,
      "grad_norm": 0.9527010917663574,
      "learning_rate": 7.25e-05,
      "loss": 0.3657,
      "step": 2550
    },
    {
      "epoch": 205.94059405940595,
      "grad_norm": 0.8948647975921631,
      "learning_rate": 7e-05,
      "loss": 0.3641,
      "step": 2600
    },
    {
      "epoch": 209.9009900990099,
      "grad_norm": 0.8640767335891724,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.3613,
      "step": 2650
    },
    {
      "epoch": 213.86138613861385,
      "grad_norm": 0.6822031736373901,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.3635,
      "step": 2700
    },
    {
      "epoch": 217.82178217821783,
      "grad_norm": 0.6595421433448792,
      "learning_rate": 6.25e-05,
      "loss": 0.3609,
      "step": 2750
    },
    {
      "epoch": 221.78217821782178,
      "grad_norm": 1.1080352067947388,
      "learning_rate": 6e-05,
      "loss": 0.3589,
      "step": 2800
    },
    {
      "epoch": 225.74257425742573,
      "grad_norm": 0.8837751746177673,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.3588,
      "step": 2850
    },
    {
      "epoch": 229.7029702970297,
      "grad_norm": 0.6309406757354736,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.3591,
      "step": 2900
    },
    {
      "epoch": 233.66336633663366,
      "grad_norm": 1.04404616355896,
      "learning_rate": 5.25e-05,
      "loss": 0.3591,
      "step": 2950
    },
    {
      "epoch": 237.62376237623764,
      "grad_norm": 1.1506808996200562,
      "learning_rate": 5e-05,
      "loss": 0.3566,
      "step": 3000
    },
    {
      "epoch": 237.62376237623764,
      "eval_loss": 0.4034203588962555,
      "eval_runtime": 7.3045,
      "eval_samples_per_second": 24.642,
      "eval_steps_per_second": 3.149,
      "step": 3000
    },
    {
      "epoch": 241.58415841584159,
      "grad_norm": 0.6666073203086853,
      "learning_rate": 4.75e-05,
      "loss": 0.3545,
      "step": 3050
    },
    {
      "epoch": 245.54455445544554,
      "grad_norm": 0.7490206360816956,
      "learning_rate": 4.5e-05,
      "loss": 0.3564,
      "step": 3100
    },
    {
      "epoch": 249.5049504950495,
      "grad_norm": 0.6986690163612366,
      "learning_rate": 4.25e-05,
      "loss": 0.355,
      "step": 3150
    },
    {
      "epoch": 253.46534653465346,
      "grad_norm": 0.45787835121154785,
      "learning_rate": 4e-05,
      "loss": 0.3508,
      "step": 3200
    },
    {
      "epoch": 257.4257425742574,
      "grad_norm": 0.7505810260772705,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.355,
      "step": 3250
    },
    {
      "epoch": 261.38613861386136,
      "grad_norm": 0.6505373120307922,
      "learning_rate": 3.5e-05,
      "loss": 0.3547,
      "step": 3300
    },
    {
      "epoch": 265.34653465346537,
      "grad_norm": 0.5026346445083618,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.353,
      "step": 3350
    },
    {
      "epoch": 269.3069306930693,
      "grad_norm": 0.7670180797576904,
      "learning_rate": 3e-05,
      "loss": 0.3503,
      "step": 3400
    },
    {
      "epoch": 273.26732673267327,
      "grad_norm": 0.5188130736351013,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.3503,
      "step": 3450
    },
    {
      "epoch": 277.2277227722772,
      "grad_norm": 0.5008130073547363,
      "learning_rate": 2.5e-05,
      "loss": 0.349,
      "step": 3500
    },
    {
      "epoch": 277.2277227722772,
      "eval_loss": 0.3990880846977234,
      "eval_runtime": 7.7645,
      "eval_samples_per_second": 23.183,
      "eval_steps_per_second": 2.962,
      "step": 3500
    },
    {
      "epoch": 281.18811881188117,
      "grad_norm": 0.5325412154197693,
      "learning_rate": 2.25e-05,
      "loss": 0.3471,
      "step": 3550
    },
    {
      "epoch": 285.1485148514852,
      "grad_norm": 0.5069642066955566,
      "learning_rate": 2e-05,
      "loss": 0.3475,
      "step": 3600
    },
    {
      "epoch": 289.1089108910891,
      "grad_norm": 0.6686875224113464,
      "learning_rate": 1.75e-05,
      "loss": 0.3472,
      "step": 3650
    },
    {
      "epoch": 293.0693069306931,
      "grad_norm": 0.49365779757499695,
      "learning_rate": 1.5e-05,
      "loss": 0.3471,
      "step": 3700
    },
    {
      "epoch": 297.029702970297,
      "grad_norm": 0.4248306155204773,
      "learning_rate": 1.25e-05,
      "loss": 0.351,
      "step": 3750
    },
    {
      "epoch": 300.990099009901,
      "grad_norm": 0.5032713413238525,
      "learning_rate": 1e-05,
      "loss": 0.3496,
      "step": 3800
    },
    {
      "epoch": 304.9504950495049,
      "grad_norm": 0.43658721446990967,
      "learning_rate": 7.5e-06,
      "loss": 0.3479,
      "step": 3850
    },
    {
      "epoch": 308.91089108910893,
      "grad_norm": 0.3654322326183319,
      "learning_rate": 5e-06,
      "loss": 0.3483,
      "step": 3900
    },
    {
      "epoch": 312.8712871287129,
      "grad_norm": 0.4196314811706543,
      "learning_rate": 2.5e-06,
      "loss": 0.3448,
      "step": 3950
    },
    {
      "epoch": 316.83168316831683,
      "grad_norm": 0.39204898476600647,
      "learning_rate": 0.0,
      "loss": 0.3466,
      "step": 4000
    },
    {
      "epoch": 316.83168316831683,
      "eval_loss": 0.39805811643600464,
      "eval_runtime": 6.6963,
      "eval_samples_per_second": 26.881,
      "eval_steps_per_second": 3.435,
      "step": 4000
    }
  ],
  "logging_steps": 50,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 334,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.643923525044128e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}