{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.993680884676145,
  "eval_steps": 100,
  "global_step": 632,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.13,
      "grad_norm": 0.5669772624969482,
      "learning_rate": 4.9888580822471086e-05,
      "loss": 1.4429,
      "step": 20
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.5426832437515259,
      "learning_rate": 4.953167838259285e-05,
      "loss": 1.142,
      "step": 40
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.5566983819007874,
      "learning_rate": 4.893250847214369e-05,
      "loss": 1.0441,
      "step": 60
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.5931661128997803,
      "learning_rate": 4.8096988312782174e-05,
      "loss": 1.0139,
      "step": 80
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.7013124227523804,
      "learning_rate": 4.703336925007311e-05,
      "loss": 0.9813,
      "step": 100
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.9671486616134644,
      "eval_runtime": 276.2617,
      "eval_samples_per_second": 10.03,
      "eval_steps_per_second": 1.256,
      "step": 100
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.6544580459594727,
      "learning_rate": 4.575215526568278e-05,
      "loss": 0.9546,
      "step": 120
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.6307847499847412,
      "learning_rate": 4.426599924321815e-05,
      "loss": 0.9498,
      "step": 140
    },
    {
      "epoch": 1.01,
      "grad_norm": 0.6634028553962708,
      "learning_rate": 4.2589578012157426e-05,
      "loss": 0.9497,
      "step": 160
    },
    {
      "epoch": 1.14,
      "grad_norm": 0.6686026453971863,
      "learning_rate": 4.073944740390061e-05,
      "loss": 0.9203,
      "step": 180
    },
    {
      "epoch": 1.26,
      "grad_norm": 0.7216710448265076,
      "learning_rate": 3.873387875136252e-05,
      "loss": 0.9108,
      "step": 200
    },
    {
      "epoch": 1.26,
      "eval_loss": 0.9249987602233887,
      "eval_runtime": 276.209,
      "eval_samples_per_second": 10.032,
      "eval_steps_per_second": 1.256,
      "step": 200
    },
    {
      "epoch": 1.39,
      "grad_norm": 0.6702993512153625,
      "learning_rate": 3.6592678446789516e-05,
      "loss": 0.908,
      "step": 220
    },
    {
      "epoch": 1.52,
      "grad_norm": 0.7004181146621704,
      "learning_rate": 3.433699233979222e-05,
      "loss": 0.9077,
      "step": 240
    },
    {
      "epoch": 1.64,
      "grad_norm": 0.6908763647079468,
      "learning_rate": 3.198909690730063e-05,
      "loss": 0.903,
      "step": 260
    },
    {
      "epoch": 1.77,
      "grad_norm": 0.6947716474533081,
      "learning_rate": 2.9572179257784215e-05,
      "loss": 0.8984,
      "step": 280
    },
    {
      "epoch": 1.9,
      "grad_norm": 0.6954981088638306,
      "learning_rate": 2.711010814234896e-05,
      "loss": 0.8976,
      "step": 300
    },
    {
      "epoch": 1.9,
      "eval_loss": 0.9090741276741028,
      "eval_runtime": 276.2477,
      "eval_samples_per_second": 10.031,
      "eval_steps_per_second": 1.256,
      "step": 300
    },
    {
      "epoch": 2.02,
      "grad_norm": 0.6500376462936401,
      "learning_rate": 2.462719823413707e-05,
      "loss": 0.8959,
      "step": 320
    },
    {
      "epoch": 2.15,
      "grad_norm": 0.7603716850280762,
      "learning_rate": 2.214797000393479e-05,
      "loss": 0.8837,
      "step": 340
    },
    {
      "epoch": 2.27,
      "grad_norm": 0.820634126663208,
      "learning_rate": 1.9696907563384687e-05,
      "loss": 0.8655,
      "step": 360
    },
    {
      "epoch": 2.4,
      "grad_norm": 0.7582141757011414,
      "learning_rate": 1.7298216867269906e-05,
      "loss": 0.8695,
      "step": 380
    },
    {
      "epoch": 2.53,
      "grad_norm": 0.8109495043754578,
      "learning_rate": 1.4975586662791783e-05,
      "loss": 0.8687,
      "step": 400
    },
    {
      "epoch": 2.53,
      "eval_loss": 0.9004740118980408,
      "eval_runtime": 276.3126,
      "eval_samples_per_second": 10.028,
      "eval_steps_per_second": 1.256,
      "step": 400
    },
    {
      "epoch": 2.65,
      "grad_norm": 0.7401316165924072,
      "learning_rate": 1.2751954546633871e-05,
      "loss": 0.8688,
      "step": 420
    },
    {
      "epoch": 2.78,
      "grad_norm": 0.7914406657218933,
      "learning_rate": 1.0649280440162326e-05,
      "loss": 0.8565,
      "step": 440
    },
    {
      "epoch": 2.91,
      "grad_norm": 0.7221879363059998,
      "learning_rate": 8.68832971985347e-06,
      "loss": 0.8672,
      "step": 460
    },
    {
      "epoch": 3.03,
      "grad_norm": 0.8058634400367737,
      "learning_rate": 6.8884681446869105e-06,
      "loss": 0.8644,
      "step": 480
    },
    {
      "epoch": 3.16,
      "grad_norm": 0.8008745908737183,
      "learning_rate": 5.267470605739952e-06,
      "loss": 0.8548,
      "step": 500
    },
    {
      "epoch": 3.16,
      "eval_loss": 0.8958437442779541,
      "eval_runtime": 276.5086,
      "eval_samples_per_second": 10.021,
      "eval_steps_per_second": 1.255,
      "step": 500
    },
    {
      "epoch": 3.29,
      "grad_norm": 0.8349476456642151,
      "learning_rate": 3.841345586714251e-06,
      "loss": 0.8522,
      "step": 520
    },
    {
      "epoch": 3.41,
      "grad_norm": 0.8091734051704407,
      "learning_rate": 2.624177068970124e-06,
      "loss": 0.8547,
      "step": 540
    },
    {
      "epoch": 3.54,
      "grad_norm": 0.7861022353172302,
      "learning_rate": 1.6279854423664697e-06,
      "loss": 0.8387,
      "step": 560
    },
    {
      "epoch": 3.67,
      "grad_norm": 0.7953125834465027,
      "learning_rate": 8.62608795509276e-07,
      "loss": 0.8471,
      "step": 580
    },
    {
      "epoch": 3.79,
      "grad_norm": 0.8621749877929688,
      "learning_rate": 3.3560575775019864e-07,
      "loss": 0.8468,
      "step": 600
    },
    {
      "epoch": 3.79,
      "eval_loss": 0.895173192024231,
      "eval_runtime": 276.743,
      "eval_samples_per_second": 10.013,
      "eval_steps_per_second": 1.254,
      "step": 600
    },
    {
      "epoch": 3.92,
      "grad_norm": 0.7905401587486267,
      "learning_rate": 5.218085243859638e-08,
      "loss": 0.8597,
      "step": 620
    },
    {
      "epoch": 3.99,
      "step": 632,
      "total_flos": 1.7866046476474122e+18,
      "train_loss": 0.9217187543458576,
      "train_runtime": 18676.6384,
      "train_samples_per_second": 2.169,
      "train_steps_per_second": 0.034
    }
  ],
  "logging_steps": 20,
  "max_steps": 632,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 1.7866046476474122e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}