{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.5,
  "eval_steps": 17500,
  "global_step": 35000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 1.1427639722824097,
      "learning_rate": 1.9712463768115945e-05,
      "loss": 0.8488,
      "step": 1000
    },
    {
      "epoch": 0.2,
      "grad_norm": 1.295353651046753,
      "learning_rate": 1.9132753623188406e-05,
      "loss": 0.7965,
      "step": 2000
    },
    {
      "epoch": 0.3,
      "grad_norm": 1.2221827507019043,
      "learning_rate": 1.855304347826087e-05,
      "loss": 0.7858,
      "step": 3000
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.9316444993019104,
      "learning_rate": 1.7973333333333333e-05,
      "loss": 0.7841,
      "step": 4000
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.7418642044067383,
      "learning_rate": 1.7393623188405798e-05,
      "loss": 0.7773,
      "step": 5000
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8065320253372192,
      "learning_rate": 1.6813913043478263e-05,
      "loss": 0.769,
      "step": 6000
    },
    {
      "epoch": 0.7,
      "grad_norm": 1.1335772275924683,
      "learning_rate": 1.6234202898550724e-05,
      "loss": 0.7678,
      "step": 7000
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.7617415189743042,
      "learning_rate": 1.565449275362319e-05,
      "loss": 0.7622,
      "step": 8000
    },
    {
      "epoch": 0.9,
      "grad_norm": 2.318528413772583,
      "learning_rate": 1.5074782608695653e-05,
      "loss": 0.7638,
      "step": 9000
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.8935675621032715,
      "learning_rate": 1.4495072463768116e-05,
      "loss": 0.7555,
      "step": 10000
    },
    {
      "epoch": 1.1,
      "grad_norm": 1.0269213914871216,
      "learning_rate": 1.3915362318840582e-05,
      "loss": 0.6674,
      "step": 11000
    },
    {
      "epoch": 1.2,
      "grad_norm": 1.0384411811828613,
      "learning_rate": 1.3335652173913046e-05,
      "loss": 0.6693,
      "step": 12000
    },
    {
      "epoch": 1.3,
      "grad_norm": 0.9560378193855286,
      "learning_rate": 1.2755942028985509e-05,
      "loss": 0.6716,
      "step": 13000
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.8821602463722229,
      "learning_rate": 1.2176231884057972e-05,
      "loss": 0.6674,
      "step": 14000
    },
    {
      "epoch": 1.5,
      "grad_norm": 1.1022319793701172,
      "learning_rate": 1.1597101449275363e-05,
      "loss": 0.6644,
      "step": 15000
    },
    {
      "epoch": 1.6,
      "grad_norm": 1.0467220544815063,
      "learning_rate": 1.1017391304347826e-05,
      "loss": 0.6624,
      "step": 16000
    },
    {
      "epoch": 1.7,
      "grad_norm": 1.3060600757598877,
      "learning_rate": 1.043768115942029e-05,
      "loss": 0.6653,
      "step": 17000
    },
    {
      "epoch": 1.75,
      "eval_loss": 0.7719004154205322,
      "eval_runtime": 101.1437,
      "eval_samples_per_second": 19.774,
      "eval_steps_per_second": 2.472,
      "step": 17500
    },
    {
      "epoch": 1.8,
      "grad_norm": 0.9809356927871704,
      "learning_rate": 9.857971014492754e-06,
      "loss": 0.6609,
      "step": 18000
    },
    {
      "epoch": 1.9,
      "grad_norm": 1.2537463903427124,
      "learning_rate": 9.279420289855074e-06,
      "loss": 0.6646,
      "step": 19000
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2699357271194458,
      "learning_rate": 8.699710144927537e-06,
      "loss": 0.6597,
      "step": 20000
    },
    {
      "epoch": 2.1,
      "grad_norm": 0.926004946231842,
      "learning_rate": 8.120000000000002e-06,
      "loss": 0.5656,
      "step": 21000
    },
    {
      "epoch": 2.2,
      "grad_norm": 1.199065923690796,
      "learning_rate": 7.540289855072464e-06,
      "loss": 0.5765,
      "step": 22000
    },
    {
      "epoch": 2.3,
      "grad_norm": 1.3075120449066162,
      "learning_rate": 6.961159420289856e-06,
      "loss": 0.5737,
      "step": 23000
    },
    {
      "epoch": 2.4,
      "grad_norm": 1.2956862449645996,
      "learning_rate": 6.382028985507247e-06,
      "loss": 0.5668,
      "step": 24000
    },
    {
      "epoch": 2.5,
      "grad_norm": 1.5514229536056519,
      "learning_rate": 5.80231884057971e-06,
      "loss": 0.5674,
      "step": 25000
    },
    {
      "epoch": 2.6,
      "grad_norm": 1.0971230268478394,
      "learning_rate": 5.222608695652175e-06,
      "loss": 0.5725,
      "step": 26000
    },
    {
      "epoch": 2.7,
      "grad_norm": 1.1359177827835083,
      "learning_rate": 4.643478260869566e-06,
      "loss": 0.5693,
      "step": 27000
    },
    {
      "epoch": 2.8,
      "grad_norm": 1.0388602018356323,
      "learning_rate": 4.06376811594203e-06,
      "loss": 0.5583,
      "step": 28000
    },
    {
      "epoch": 2.9,
      "grad_norm": 1.0011905431747437,
      "learning_rate": 3.4846376811594208e-06,
      "loss": 0.5643,
      "step": 29000
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0876168012619019,
      "learning_rate": 2.904927536231884e-06,
      "loss": 0.5647,
      "step": 30000
    },
    {
      "epoch": 3.1,
      "grad_norm": 1.2214146852493286,
      "learning_rate": 2.325217391304348e-06,
      "loss": 0.4977,
      "step": 31000
    },
    {
      "epoch": 3.2,
      "grad_norm": 1.4341120719909668,
      "learning_rate": 1.7455072463768116e-06,
      "loss": 0.4908,
      "step": 32000
    },
    {
      "epoch": 3.3,
      "grad_norm": 1.2273516654968262,
      "learning_rate": 1.166376811594203e-06,
      "loss": 0.4934,
      "step": 33000
    },
    {
      "epoch": 3.4,
      "grad_norm": 1.446751356124878,
      "learning_rate": 5.866666666666667e-07,
      "loss": 0.4886,
      "step": 34000
    },
    {
      "epoch": 3.5,
      "grad_norm": 0.9594036936759949,
      "learning_rate": 6.956521739130436e-09,
      "loss": 0.4948,
      "step": 35000
    },
    {
      "epoch": 3.5,
      "eval_loss": 0.8087719082832336,
      "eval_runtime": 99.7391,
      "eval_samples_per_second": 20.052,
      "eval_steps_per_second": 2.507,
      "step": 35000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 35000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 17500,
  "total_flos": 4.40543553257472e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}