{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 480,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "grad_norm": 4.709561347961426,
      "learning_rate": 4.994647308096509e-05,
      "loss": 8.9503,
      "step": 10
    },
    {
      "epoch": 0.12,
      "grad_norm": 2.5953965187072754,
      "learning_rate": 4.9786121534345265e-05,
      "loss": 8.0851,
      "step": 20
    },
    {
      "epoch": 0.19,
      "grad_norm": 2.367886543273926,
      "learning_rate": 4.951963201008076e-05,
      "loss": 7.5574,
      "step": 30
    },
    {
      "epoch": 0.25,
      "grad_norm": 3.426809549331665,
      "learning_rate": 4.914814565722671e-05,
      "loss": 6.9461,
      "step": 40
    },
    {
      "epoch": 0.31,
      "grad_norm": 4.523465633392334,
      "learning_rate": 4.867325323737765e-05,
      "loss": 6.1526,
      "step": 50
    },
    {
      "epoch": 0.38,
      "grad_norm": 1.7781277894973755,
      "learning_rate": 4.8096988312782174e-05,
      "loss": 5.4116,
      "step": 60
    },
    {
      "epoch": 0.44,
      "grad_norm": 1.3401896953582764,
      "learning_rate": 4.742181853831721e-05,
      "loss": 5.0721,
      "step": 70
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.9671522378921509,
      "learning_rate": 4.665063509461097e-05,
      "loss": 4.7695,
      "step": 80
    },
    {
      "epoch": 0.56,
      "grad_norm": 1.055595874786377,
      "learning_rate": 4.5786740307563636e-05,
      "loss": 4.6179,
      "step": 90
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.7612900733947754,
      "learning_rate": 4.4833833507280884e-05,
      "loss": 4.4848,
      "step": 100
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.7046266198158264,
      "learning_rate": 4.379599518697444e-05,
      "loss": 4.3627,
      "step": 110
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.9548586010932922,
      "learning_rate": 4.267766952966369e-05,
      "loss": 4.3319,
      "step": 120
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.6869709491729736,
      "learning_rate": 4.148364537750172e-05,
      "loss": 4.269,
      "step": 130
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.6282512545585632,
      "learning_rate": 4.021903572521802e-05,
      "loss": 4.2557,
      "step": 140
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.5287710428237915,
      "learning_rate": 3.888925582549006e-05,
      "loss": 4.1534,
      "step": 150
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.7542024850845337,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 4.2017,
      "step": 160
    },
    {
      "epoch": 1.06,
      "grad_norm": 0.49876415729522705,
      "learning_rate": 3.6057217255475034e-05,
      "loss": 4.1732,
      "step": 170
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.5626935958862305,
      "learning_rate": 3.456708580912725e-05,
      "loss": 4.1358,
      "step": 180
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.493310809135437,
      "learning_rate": 3.303598663257904e-05,
      "loss": 4.1514,
      "step": 190
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.4654150605201721,
      "learning_rate": 3.147047612756302e-05,
      "loss": 4.0925,
      "step": 200
    },
    {
      "epoch": 1.31,
      "grad_norm": 0.592623770236969,
      "learning_rate": 2.9877258050403212e-05,
      "loss": 4.0496,
      "step": 210
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.5564578771591187,
      "learning_rate": 2.8263154805501297e-05,
      "loss": 4.0853,
      "step": 220
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.6952773332595825,
      "learning_rate": 2.663507823075358e-05,
      "loss": 4.1283,
      "step": 230
    },
    {
      "epoch": 1.5,
      "grad_norm": 0.5385617613792419,
      "learning_rate": 2.5e-05,
      "loss": 4.0567,
      "step": 240
    },
    {
      "epoch": 1.56,
      "grad_norm": 0.5663427114486694,
      "learning_rate": 2.3364921769246423e-05,
      "loss": 4.0141,
      "step": 250
    },
    {
      "epoch": 1.62,
      "grad_norm": 0.5520788431167603,
      "learning_rate": 2.173684519449872e-05,
      "loss": 4.0356,
      "step": 260
    },
    {
      "epoch": 1.69,
      "grad_norm": 0.5162128806114197,
      "learning_rate": 2.0122741949596797e-05,
      "loss": 4.082,
      "step": 270
    },
    {
      "epoch": 1.75,
      "grad_norm": 0.5291630625724792,
      "learning_rate": 1.852952387243698e-05,
      "loss": 4.0767,
      "step": 280
    },
    {
      "epoch": 1.81,
      "grad_norm": 0.6226648092269897,
      "learning_rate": 1.6964013367420966e-05,
      "loss": 3.9817,
      "step": 290
    },
    {
      "epoch": 1.88,
      "grad_norm": 0.5460664629936218,
      "learning_rate": 1.5432914190872757e-05,
      "loss": 4.0618,
      "step": 300
    },
    {
      "epoch": 1.94,
      "grad_norm": 0.7545162439346313,
      "learning_rate": 1.3942782744524973e-05,
      "loss": 4.0106,
      "step": 310
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6207989454269409,
      "learning_rate": 1.2500000000000006e-05,
      "loss": 4.0549,
      "step": 320
    },
    {
      "epoch": 2.06,
      "grad_norm": 0.5338532328605652,
      "learning_rate": 1.1110744174509952e-05,
      "loss": 3.919,
      "step": 330
    },
    {
      "epoch": 2.12,
      "grad_norm": 0.5484297275543213,
      "learning_rate": 9.780964274781984e-06,
      "loss": 4.053,
      "step": 340
    },
    {
      "epoch": 2.19,
      "grad_norm": 0.6356564164161682,
      "learning_rate": 8.51635462249828e-06,
      "loss": 4.0032,
      "step": 350
    },
    {
      "epoch": 2.25,
      "grad_norm": 0.5518457889556885,
      "learning_rate": 7.3223304703363135e-06,
      "loss": 3.991,
      "step": 360
    },
    {
      "epoch": 2.31,
      "grad_norm": 0.5176472067832947,
      "learning_rate": 6.204004813025568e-06,
      "loss": 3.9941,
      "step": 370
    },
    {
      "epoch": 2.38,
      "grad_norm": 0.5543831586837769,
      "learning_rate": 5.166166492719124e-06,
      "loss": 4.0388,
      "step": 380
    },
    {
      "epoch": 2.44,
      "grad_norm": 0.5504453182220459,
      "learning_rate": 4.213259692436367e-06,
      "loss": 4.0767,
      "step": 390
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.5619158148765564,
      "learning_rate": 3.3493649053890326e-06,
      "loss": 3.9709,
      "step": 400
    },
    {
      "epoch": 2.56,
      "grad_norm": 0.5513697266578674,
      "learning_rate": 2.578181461682794e-06,
      "loss": 3.9714,
      "step": 410
    },
    {
      "epoch": 2.62,
      "grad_norm": 0.590857744216919,
      "learning_rate": 1.9030116872178316e-06,
      "loss": 4.0366,
      "step": 420
    },
    {
      "epoch": 2.69,
      "grad_norm": 0.5728959441184998,
      "learning_rate": 1.3267467626223606e-06,
      "loss": 3.9307,
      "step": 430
    },
    {
      "epoch": 2.75,
      "grad_norm": 0.5163474082946777,
      "learning_rate": 8.51854342773295e-07,
      "loss": 4.0751,
      "step": 440
    },
    {
      "epoch": 2.81,
      "grad_norm": 0.5247732996940613,
      "learning_rate": 4.803679899192392e-07,
      "loss": 4.0141,
      "step": 450
    },
    {
      "epoch": 2.88,
      "grad_norm": 0.5492649674415588,
      "learning_rate": 2.1387846565474045e-07,
      "loss": 3.9676,
      "step": 460
    },
    {
      "epoch": 2.94,
      "grad_norm": 0.5566267371177673,
      "learning_rate": 5.352691903491303e-08,
      "loss": 4.0041,
      "step": 470
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6199953556060791,
      "learning_rate": 0.0,
      "loss": 3.909,
      "step": 480
    },
    {
      "epoch": 3.0,
      "step": 480,
      "total_flos": 3.260181978788659e+17,
      "train_loss": 4.515984590848287,
      "train_runtime": 5513.8168,
      "train_samples_per_second": 0.696,
      "train_steps_per_second": 0.087
    }
  ],
  "logging_steps": 10,
  "max_steps": 480,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 1000,
  "total_flos": 3.260181978788659e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}