{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.998300283286119,
  "eval_steps": 1000,
  "global_step": 14112,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002834467120181406,
      "grad_norm": 5.449254512786865,
      "learning_rate": 0.0,
      "loss": 2.6239,
      "step": 1
    },
    {
      "epoch": 0.2834467120181406,
      "grad_norm": 3.8293514251708984,
      "learning_rate": 2.1225212464589238e-05,
      "loss": 2.239,
      "step": 1000
    },
    {
      "epoch": 0.2834467120181406,
      "eval_cosine_accuracy": 0.9631311893463135,
      "eval_loss": 1.070603609085083,
      "eval_runtime": 24.3925,
      "eval_samples_per_second": 388.071,
      "eval_steps_per_second": 1.517,
      "step": 1000
    },
    {
      "epoch": 0.5668934240362812,
      "grad_norm": 3.8862180709838867,
      "learning_rate": 2.8613385826771657e-05,
      "loss": 1.6315,
      "step": 2000
    },
    {
      "epoch": 0.5668934240362812,
      "eval_cosine_accuracy": 0.9705260992050171,
      "eval_loss": 0.9979932308197021,
      "eval_runtime": 23.6004,
      "eval_samples_per_second": 401.094,
      "eval_steps_per_second": 1.568,
      "step": 2000
    },
    {
      "epoch": 0.8503401360544217,
      "grad_norm": 3.756399631500244,
      "learning_rate": 2.6251181102362205e-05,
      "loss": 1.4885,
      "step": 3000
    },
    {
      "epoch": 0.8503401360544217,
      "eval_cosine_accuracy": 0.9735897183418274,
      "eval_loss": 0.8850170373916626,
      "eval_runtime": 23.076,
      "eval_samples_per_second": 410.209,
      "eval_steps_per_second": 1.603,
      "step": 3000
    },
    {
      "epoch": 1.1337110481586403,
      "grad_norm": 4.797242164611816,
      "learning_rate": 2.3888976377952758e-05,
      "loss": 1.8836,
      "step": 4000
    },
    {
      "epoch": 1.1337110481586403,
      "eval_cosine_accuracy": 0.9749630093574524,
      "eval_loss": 0.8781672716140747,
      "eval_runtime": 23.5925,
      "eval_samples_per_second": 401.23,
      "eval_steps_per_second": 1.568,
      "step": 4000
    },
    {
      "epoch": 1.4169971671388102,
      "grad_norm": 3.78368878364563,
      "learning_rate": 2.152913385826772e-05,
      "loss": 1.3749,
      "step": 5000
    },
    {
      "epoch": 1.4169971671388102,
      "eval_cosine_accuracy": 0.9753856062889099,
      "eval_loss": 0.8662684559822083,
      "eval_runtime": 23.3141,
      "eval_samples_per_second": 406.021,
      "eval_steps_per_second": 1.587,
      "step": 5000
    },
    {
      "epoch": 1.7002832861189803,
      "grad_norm": 4.068625450134277,
      "learning_rate": 1.916692913385827e-05,
      "loss": 1.3631,
      "step": 6000
    },
    {
      "epoch": 1.7002832861189803,
      "eval_cosine_accuracy": 0.9768645763397217,
      "eval_loss": 0.8464893102645874,
      "eval_runtime": 23.2294,
      "eval_samples_per_second": 407.502,
      "eval_steps_per_second": 1.593,
      "step": 6000
    },
    {
      "epoch": 1.9835694050991501,
      "grad_norm": 3.349905252456665,
      "learning_rate": 1.6807086614173228e-05,
      "loss": 1.3027,
      "step": 7000
    },
    {
      "epoch": 1.9835694050991501,
      "eval_cosine_accuracy": 0.9771814942359924,
      "eval_loss": 0.8421413898468018,
      "eval_runtime": 23.2999,
      "eval_samples_per_second": 406.268,
      "eval_steps_per_second": 1.588,
      "step": 7000
    },
    {
      "epoch": 2.26685552407932,
      "grad_norm": 3.673818826675415,
      "learning_rate": 1.444488188976378e-05,
      "loss": 1.2536,
      "step": 8000
    },
    {
      "epoch": 2.26685552407932,
      "eval_cosine_accuracy": 0.9765476584434509,
      "eval_loss": 0.8478096723556519,
      "eval_runtime": 25.0357,
      "eval_samples_per_second": 378.1,
      "eval_steps_per_second": 1.478,
      "step": 8000
    },
    {
      "epoch": 2.55014164305949,
      "grad_norm": 3.332353115081787,
      "learning_rate": 1.208503937007874e-05,
      "loss": 1.2155,
      "step": 9000
    },
    {
      "epoch": 2.55014164305949,
      "eval_cosine_accuracy": 0.9766532778739929,
      "eval_loss": 0.8358168005943298,
      "eval_runtime": 23.5231,
      "eval_samples_per_second": 402.412,
      "eval_steps_per_second": 1.573,
      "step": 9000
    },
    {
      "epoch": 2.83342776203966,
      "grad_norm": 3.4396846294403076,
      "learning_rate": 9.722834645669293e-06,
      "loss": 1.1847,
      "step": 10000
    },
    {
      "epoch": 2.83342776203966,
      "eval_cosine_accuracy": 0.9782379269599915,
      "eval_loss": 0.8319438695907593,
      "eval_runtime": 23.4855,
      "eval_samples_per_second": 403.058,
      "eval_steps_per_second": 1.575,
      "step": 10000
    },
    {
      "epoch": 3.11671388101983,
      "grad_norm": 4.006954193115234,
      "learning_rate": 7.365354330708662e-06,
      "loss": 1.1857,
      "step": 11000
    },
    {
      "epoch": 3.11671388101983,
      "eval_cosine_accuracy": 0.9778153300285339,
      "eval_loss": 0.834368109703064,
      "eval_runtime": 23.4546,
      "eval_samples_per_second": 403.588,
      "eval_steps_per_second": 1.578,
      "step": 11000
    },
    {
      "epoch": 3.4,
      "grad_norm": 3.616560697555542,
      "learning_rate": 5.003149606299213e-06,
      "loss": 1.1544,
      "step": 12000
    },
    {
      "epoch": 3.4,
      "eval_cosine_accuracy": 0.9772871136665344,
      "eval_loss": 0.8377622365951538,
      "eval_runtime": 23.5128,
      "eval_samples_per_second": 402.59,
      "eval_steps_per_second": 1.574,
      "step": 12000
    },
    {
      "epoch": 3.68328611898017,
      "grad_norm": 3.396005630493164,
      "learning_rate": 2.6409448818897637e-06,
      "loss": 1.1787,
      "step": 13000
    },
    {
      "epoch": 3.68328611898017,
      "eval_cosine_accuracy": 0.9770758748054504,
      "eval_loss": 0.831461489200592,
      "eval_runtime": 23.4599,
      "eval_samples_per_second": 403.497,
      "eval_steps_per_second": 1.577,
      "step": 13000
    },
    {
      "epoch": 3.96657223796034,
      "grad_norm": 3.424636125564575,
      "learning_rate": 2.7874015748031496e-07,
      "loss": 1.1452,
      "step": 14000
    },
    {
      "epoch": 3.96657223796034,
      "eval_cosine_accuracy": 0.9772871136665344,
      "eval_loss": 0.8320337533950806,
      "eval_runtime": 23.4637,
      "eval_samples_per_second": 403.431,
      "eval_steps_per_second": 1.577,
      "step": 14000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 14112,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}