{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 7.0,
  "eval_steps": 1000,
  "global_step": 17715,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00039541320680110717,
      "grad_norm": 6.574801445007324,
      "learning_rate": 0.0,
      "loss": 4.3129,
      "step": 1
    },
    {
      "epoch": 0.39541320680110714,
      "grad_norm": 6.2015180587768555,
      "learning_rate": 9.871541501976284e-06,
      "loss": 3.298,
      "step": 1000
    },
    {
      "epoch": 0.39541320680110714,
      "eval_cosine_accuracy": 0.9471027255058289,
      "eval_loss": 0.5520303249359131,
      "eval_runtime": 36.3857,
      "eval_samples_per_second": 261.339,
      "eval_steps_per_second": 1.044,
      "step": 1000
    },
    {
      "epoch": 0.7908264136022143,
      "grad_norm": 8.382867813110352,
      "learning_rate": 1.9752964426877474e-05,
      "loss": 1.9673,
      "step": 2000
    },
    {
      "epoch": 0.7908264136022143,
      "eval_cosine_accuracy": 0.953938364982605,
      "eval_loss": 0.4995275139808655,
      "eval_runtime": 34.9654,
      "eval_samples_per_second": 271.955,
      "eval_steps_per_second": 1.087,
      "step": 2000
    },
    {
      "epoch": 1.1860924535756618,
      "grad_norm": 6.899052619934082,
      "learning_rate": 1.8930140597539545e-05,
      "loss": 1.5646,
      "step": 3000
    },
    {
      "epoch": 1.1860924535756618,
      "eval_cosine_accuracy": 0.9593017101287842,
      "eval_loss": 0.4840952455997467,
      "eval_runtime": 34.4403,
      "eval_samples_per_second": 276.101,
      "eval_steps_per_second": 1.103,
      "step": 3000
    },
    {
      "epoch": 1.581193204267088,
      "grad_norm": 6.6706132888793945,
      "learning_rate": 1.7831722319859405e-05,
      "loss": 1.5836,
      "step": 4000
    },
    {
      "epoch": 1.581193204267088,
      "eval_cosine_accuracy": 0.962666928768158,
      "eval_loss": 0.4643610715866089,
      "eval_runtime": 34.4996,
      "eval_samples_per_second": 275.627,
      "eval_steps_per_second": 1.101,
      "step": 4000
    },
    {
      "epoch": 1.9762939549585146,
      "grad_norm": 6.573458194732666,
      "learning_rate": 1.6733304042179265e-05,
      "loss": 1.4401,
      "step": 5000
    },
    {
      "epoch": 1.9762939549585146,
      "eval_cosine_accuracy": 0.9638237357139587,
      "eval_loss": 0.44956663250923157,
      "eval_runtime": 34.8836,
      "eval_samples_per_second": 272.593,
      "eval_steps_per_second": 1.089,
      "step": 5000
    },
    {
      "epoch": 2.3713947056499407,
      "grad_norm": 5.309902191162109,
      "learning_rate": 1.5634885764499125e-05,
      "loss": 1.2966,
      "step": 6000
    },
    {
      "epoch": 2.3713947056499407,
      "eval_cosine_accuracy": 0.9671889543533325,
      "eval_loss": 0.4553391933441162,
      "eval_runtime": 34.4231,
      "eval_samples_per_second": 276.239,
      "eval_steps_per_second": 1.104,
      "step": 6000
    },
    {
      "epoch": 2.7664954563413673,
      "grad_norm": 5.066314220428467,
      "learning_rate": 1.4538664323374343e-05,
      "loss": 1.2287,
      "step": 7000
    },
    {
      "epoch": 2.7664954563413673,
      "eval_cosine_accuracy": 0.9656115174293518,
      "eval_loss": 0.443572074174881,
      "eval_runtime": 34.6972,
      "eval_samples_per_second": 274.057,
      "eval_steps_per_second": 1.095,
      "step": 7000
    },
    {
      "epoch": 3.1615962070327934,
      "grad_norm": 5.241501331329346,
      "learning_rate": 1.34402460456942e-05,
      "loss": 1.1559,
      "step": 8000
    },
    {
      "epoch": 3.1615962070327934,
      "eval_cosine_accuracy": 0.9663476943969727,
      "eval_loss": 0.44341471791267395,
      "eval_runtime": 34.4306,
      "eval_samples_per_second": 276.179,
      "eval_steps_per_second": 1.104,
      "step": 8000
    },
    {
      "epoch": 3.5566969577242196,
      "grad_norm": 4.690669059753418,
      "learning_rate": 1.234182776801406e-05,
      "loss": 1.1011,
      "step": 9000
    },
    {
      "epoch": 3.5566969577242196,
      "eval_cosine_accuracy": 0.9673992991447449,
      "eval_loss": 0.4327247738838196,
      "eval_runtime": 34.3296,
      "eval_samples_per_second": 276.991,
      "eval_steps_per_second": 1.107,
      "step": 9000
    },
    {
      "epoch": 3.951797708415646,
      "grad_norm": 8.417939186096191,
      "learning_rate": 1.124340949033392e-05,
      "loss": 1.0585,
      "step": 10000
    },
    {
      "epoch": 3.951797708415646,
      "eval_cosine_accuracy": 0.9675044417381287,
      "eval_loss": 0.43402087688446045,
      "eval_runtime": 34.3458,
      "eval_samples_per_second": 276.86,
      "eval_steps_per_second": 1.106,
      "step": 10000
    },
    {
      "epoch": 4.346898459107072,
      "grad_norm": 5.533106327056885,
      "learning_rate": 1.014608963093146e-05,
      "loss": 1.0006,
      "step": 11000
    },
    {
      "epoch": 4.346898459107072,
      "eval_cosine_accuracy": 0.9684509634971619,
      "eval_loss": 0.42998796701431274,
      "eval_runtime": 35.621,
      "eval_samples_per_second": 266.95,
      "eval_steps_per_second": 1.067,
      "step": 11000
    },
    {
      "epoch": 4.741999209798498,
      "grad_norm": 5.348298072814941,
      "learning_rate": 9.047671353251318e-06,
      "loss": 0.9839,
      "step": 12000
    },
    {
      "epoch": 4.741999209798498,
      "eval_cosine_accuracy": 0.9686612486839294,
      "eval_loss": 0.42643481492996216,
      "eval_runtime": 34.294,
      "eval_samples_per_second": 277.279,
      "eval_steps_per_second": 1.108,
      "step": 12000
    },
    {
      "epoch": 5.137099960489925,
      "grad_norm": 5.4071149826049805,
      "learning_rate": 7.950351493848859e-06,
      "loss": 0.9693,
      "step": 13000
    },
    {
      "epoch": 5.137099960489925,
      "eval_cosine_accuracy": 0.9684509634971619,
      "eval_loss": 0.4284929931163788,
      "eval_runtime": 34.4423,
      "eval_samples_per_second": 276.085,
      "eval_steps_per_second": 1.103,
      "step": 13000
    },
    {
      "epoch": 5.532200711181352,
      "grad_norm": 7.151297092437744,
      "learning_rate": 6.853031634446398e-06,
      "loss": 0.9279,
      "step": 14000
    },
    {
      "epoch": 5.532200711181352,
      "eval_cosine_accuracy": 0.9680302739143372,
      "eval_loss": 0.43396368622779846,
      "eval_runtime": 34.0774,
      "eval_samples_per_second": 279.041,
      "eval_steps_per_second": 1.115,
      "step": 14000
    },
    {
      "epoch": 5.927301461872777,
      "grad_norm": 4.7381510734558105,
      "learning_rate": 5.755711775043937e-06,
      "loss": 0.9175,
      "step": 15000
    },
    {
      "epoch": 5.927301461872777,
      "eval_cosine_accuracy": 0.9693974256515503,
      "eval_loss": 0.42655229568481445,
      "eval_runtime": 34.3744,
      "eval_samples_per_second": 276.63,
      "eval_steps_per_second": 1.105,
      "step": 15000
    },
    {
      "epoch": 6.322402212564204,
      "grad_norm": 6.581507205963135,
      "learning_rate": 4.657293497363797e-06,
      "loss": 0.8994,
      "step": 16000
    },
    {
      "epoch": 6.322402212564204,
      "eval_cosine_accuracy": 0.9680302739143372,
      "eval_loss": 0.4257819354534149,
      "eval_runtime": 35.0136,
      "eval_samples_per_second": 271.58,
      "eval_steps_per_second": 1.085,
      "step": 16000
    },
    {
      "epoch": 6.7175029632556305,
      "grad_norm": 4.414986610412598,
      "learning_rate": 3.558875219683656e-06,
      "loss": 0.8934,
      "step": 17000
    },
    {
      "epoch": 6.7175029632556305,
      "eval_cosine_accuracy": 0.9684509634971619,
      "eval_loss": 0.4267849922180176,
      "eval_runtime": 34.3541,
      "eval_samples_per_second": 276.794,
      "eval_steps_per_second": 1.106,
      "step": 17000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 20232,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "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": 256,
  "trial_name": null,
  "trial_params": null
}