{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.427584953661639,
  "eval_steps": 1000,
  "global_step": 11765,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00018171906232963838,
      "grad_norm": 7.298370361328125,
      "learning_rate": 0.0,
      "loss": 3.0984,
      "step": 1
    },
    {
      "epoch": 0.1817190623296384,
      "grad_norm": 6.053193092346191,
      "learning_rate": 3.6307468653461755e-06,
      "loss": 2.7134,
      "step": 1000
    },
    {
      "epoch": 0.1817190623296384,
      "eval_cosine_accuracy": 0.9391103386878967,
      "eval_loss": 1.3783869743347168,
      "eval_runtime": 18.26,
      "eval_samples_per_second": 520.755,
      "eval_steps_per_second": 4.107,
      "step": 1000
    },
    {
      "epoch": 0.3634381246592768,
      "grad_norm": 5.73389196395874,
      "learning_rate": 7.265128111938943e-06,
      "loss": 2.1597,
      "step": 2000
    },
    {
      "epoch": 0.3634381246592768,
      "eval_cosine_accuracy": 0.9412136077880859,
      "eval_loss": 1.2863198518753052,
      "eval_runtime": 18.0791,
      "eval_samples_per_second": 525.968,
      "eval_steps_per_second": 4.148,
      "step": 2000
    },
    {
      "epoch": 1.117572233327276,
      "grad_norm": 5.810453414916992,
      "learning_rate": 1.089950935853171e-05,
      "loss": 1.8694,
      "step": 3000
    },
    {
      "epoch": 1.117572233327276,
      "eval_cosine_accuracy": 0.9422652125358582,
      "eval_loss": 1.2363770008087158,
      "eval_runtime": 18.2075,
      "eval_samples_per_second": 522.257,
      "eval_steps_per_second": 4.119,
      "step": 3000
    },
    {
      "epoch": 1.2992912956569145,
      "grad_norm": 5.974574089050293,
      "learning_rate": 1.4533890605124479e-05,
      "loss": 1.6564,
      "step": 4000
    },
    {
      "epoch": 1.2992912956569145,
      "eval_cosine_accuracy": 0.9448943138122559,
      "eval_loss": 1.18900728225708,
      "eval_runtime": 18.0459,
      "eval_samples_per_second": 526.935,
      "eval_steps_per_second": 4.156,
      "step": 4000
    },
    {
      "epoch": 2.053425404324914,
      "grad_norm": 6.072054386138916,
      "learning_rate": 1.8164637470470654e-05,
      "loss": 1.4993,
      "step": 5000
    },
    {
      "epoch": 2.053425404324914,
      "eval_cosine_accuracy": 0.9467872381210327,
      "eval_loss": 1.1734814643859863,
      "eval_runtime": 18.1595,
      "eval_samples_per_second": 523.638,
      "eval_steps_per_second": 4.13,
      "step": 5000
    },
    {
      "epoch": 2.235144466654552,
      "grad_norm": 5.4134297370910645,
      "learning_rate": 1.9550245320734146e-05,
      "loss": 1.3577,
      "step": 6000
    },
    {
      "epoch": 2.235144466654552,
      "eval_cosine_accuracy": 0.9507834911346436,
      "eval_loss": 1.1353392601013184,
      "eval_runtime": 19.0833,
      "eval_samples_per_second": 498.29,
      "eval_steps_per_second": 3.93,
      "step": 6000
    },
    {
      "epoch": 2.4168635289841904,
      "grad_norm": 5.546666145324707,
      "learning_rate": 1.8642558604397603e-05,
      "loss": 1.2577,
      "step": 7000
    },
    {
      "epoch": 2.4168635289841904,
      "eval_cosine_accuracy": 0.953517735004425,
      "eval_loss": 1.1202903985977173,
      "eval_runtime": 18.369,
      "eval_samples_per_second": 517.667,
      "eval_steps_per_second": 4.083,
      "step": 7000
    },
    {
      "epoch": 3.1709976376521896,
      "grad_norm": 5.281265735626221,
      "learning_rate": 1.773396329274941e-05,
      "loss": 1.1667,
      "step": 8000
    },
    {
      "epoch": 3.1709976376521896,
      "eval_cosine_accuracy": 0.9548848271369934,
      "eval_loss": 1.1058506965637207,
      "eval_runtime": 18.194,
      "eval_samples_per_second": 522.645,
      "eval_steps_per_second": 4.122,
      "step": 8000
    },
    {
      "epoch": 3.352716699981828,
      "grad_norm": 6.0141777992248535,
      "learning_rate": 1.682627657641287e-05,
      "loss": 1.1052,
      "step": 9000
    },
    {
      "epoch": 3.352716699981828,
      "eval_cosine_accuracy": 0.9570932984352112,
      "eval_loss": 1.104658603668213,
      "eval_runtime": 18.2305,
      "eval_samples_per_second": 521.599,
      "eval_steps_per_second": 4.114,
      "step": 9000
    },
    {
      "epoch": 4.106850808649828,
      "grad_norm": 5.310253620147705,
      "learning_rate": 1.5917681264764674e-05,
      "loss": 1.0559,
      "step": 10000
    },
    {
      "epoch": 4.106850808649828,
      "eval_cosine_accuracy": 0.9553055167198181,
      "eval_loss": 1.1141564846038818,
      "eval_runtime": 18.2072,
      "eval_samples_per_second": 522.266,
      "eval_steps_per_second": 4.119,
      "step": 10000
    },
    {
      "epoch": 4.288569870979465,
      "grad_norm": 5.053646087646484,
      "learning_rate": 1.500999454842813e-05,
      "loss": 1.0006,
      "step": 11000
    },
    {
      "epoch": 4.288569870979465,
      "eval_cosine_accuracy": 0.9562519788742065,
      "eval_loss": 1.1013787984848022,
      "eval_runtime": 18.9532,
      "eval_samples_per_second": 501.71,
      "eval_steps_per_second": 3.957,
      "step": 11000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 27515,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 0.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": null
}