{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 1000,
  "global_step": 10464,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00019116803670426305,
      "grad_norm": 4.807868003845215,
      "learning_rate": 0.0,
      "loss": 2.5242,
      "step": 1
    },
    {
      "epoch": 0.19116803670426305,
      "grad_norm": 3.5757293701171875,
      "learning_rate": 1.4319159101767798e-05,
      "loss": 1.9554,
      "step": 1000
    },
    {
      "epoch": 0.19116803670426305,
      "eval_cosine_accuracy": 0.9553137421607971,
      "eval_loss": 1.039905071258545,
      "eval_runtime": 24.1175,
      "eval_samples_per_second": 392.496,
      "eval_steps_per_second": 1.534,
      "step": 1000
    },
    {
      "epoch": 0.3823360734085261,
      "grad_norm": 3.9157936573028564,
      "learning_rate": 2.8652651696129955e-05,
      "loss": 1.4719,
      "step": 2000
    },
    {
      "epoch": 0.3823360734085261,
      "eval_cosine_accuracy": 0.9632368683815002,
      "eval_loss": 0.9348638653755188,
      "eval_runtime": 29.104,
      "eval_samples_per_second": 325.248,
      "eval_steps_per_second": 1.271,
      "step": 2000
    },
    {
      "epoch": 0.5735041101127891,
      "grad_norm": 3.248283863067627,
      "learning_rate": 2.8556635335351283e-05,
      "loss": 1.2229,
      "step": 3000
    },
    {
      "epoch": 0.5735041101127891,
      "eval_cosine_accuracy": 0.9708430171012878,
      "eval_loss": 0.8906580209732056,
      "eval_runtime": 23.3307,
      "eval_samples_per_second": 405.731,
      "eval_steps_per_second": 1.586,
      "step": 3000
    },
    {
      "epoch": 0.7646721468170522,
      "grad_norm": 3.514542579650879,
      "learning_rate": 2.696351760395093e-05,
      "loss": 1.1579,
      "step": 4000
    },
    {
      "epoch": 0.7646721468170522,
      "eval_cosine_accuracy": 0.9724276065826416,
      "eval_loss": 0.8390598297119141,
      "eval_runtime": 23.2953,
      "eval_samples_per_second": 406.348,
      "eval_steps_per_second": 1.588,
      "step": 4000
    },
    {
      "epoch": 0.9558401835213153,
      "grad_norm": 3.2615766525268555,
      "learning_rate": 2.5371992990281982e-05,
      "loss": 1.6785,
      "step": 5000
    },
    {
      "epoch": 0.9558401835213153,
      "eval_cosine_accuracy": 0.9715825319290161,
      "eval_loss": 0.8355301022529602,
      "eval_runtime": 23.317,
      "eval_samples_per_second": 405.97,
      "eval_steps_per_second": 1.587,
      "step": 5000
    },
    {
      "epoch": 1.1469520351614753,
      "grad_norm": 3.2241148948669434,
      "learning_rate": 2.3778875258881633e-05,
      "loss": 1.5581,
      "step": 6000
    },
    {
      "epoch": 1.1469520351614753,
      "eval_cosine_accuracy": 0.9743291735649109,
      "eval_loss": 0.8016239404678345,
      "eval_runtime": 23.5346,
      "eval_samples_per_second": 402.216,
      "eval_steps_per_second": 1.572,
      "step": 6000
    },
    {
      "epoch": 1.3380470093636538,
      "grad_norm": 3.1033902168273926,
      "learning_rate": 2.218735064521268e-05,
      "loss": 1.1703,
      "step": 7000
    },
    {
      "epoch": 1.3380470093636538,
      "eval_cosine_accuracy": 0.9749630093574524,
      "eval_loss": 0.7846724390983582,
      "eval_runtime": 23.495,
      "eval_samples_per_second": 402.895,
      "eval_steps_per_second": 1.575,
      "step": 7000
    },
    {
      "epoch": 1.5291419835658322,
      "grad_norm": 2.888745069503784,
      "learning_rate": 2.059423291381233e-05,
      "loss": 1.1501,
      "step": 8000
    },
    {
      "epoch": 1.5291419835658322,
      "eval_cosine_accuracy": 0.9768645763397217,
      "eval_loss": 0.7773452401161194,
      "eval_runtime": 23.6912,
      "eval_samples_per_second": 399.558,
      "eval_steps_per_second": 1.562,
      "step": 8000
    },
    {
      "epoch": 1.7202369577680106,
      "grad_norm": 3.1285414695739746,
      "learning_rate": 1.900270830014338e-05,
      "loss": 1.1301,
      "step": 9000
    },
    {
      "epoch": 1.7202369577680106,
      "eval_cosine_accuracy": 0.9761250615119934,
      "eval_loss": 0.7771484851837158,
      "eval_runtime": 23.6285,
      "eval_samples_per_second": 400.618,
      "eval_steps_per_second": 1.566,
      "step": 9000
    },
    {
      "epoch": 1.9113319319701891,
      "grad_norm": 7.133597373962402,
      "learning_rate": 1.740959056874303e-05,
      "loss": 1.1102,
      "step": 10000
    },
    {
      "epoch": 1.9113319319701891,
      "eval_cosine_accuracy": 0.9766532778739929,
      "eval_loss": 0.7554724216461182,
      "eval_runtime": 23.6719,
      "eval_samples_per_second": 399.884,
      "eval_steps_per_second": 1.563,
      "step": 10000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 20924,
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}