{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9893390191897654,
  "eval_steps": 100,
  "global_step": 58,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 602.0948867797852,
      "epoch": 0.017057569296375266,
      "grad_norm": 0.5694698691368103,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0295,
      "reward": 0.628348246216774,
      "reward_std": 0.33944267593324184,
      "rewards/accuracy_reward": 0.628348246216774,
      "rewards/format_reward": 0.0,
      "step": 1
    },
    {
      "completion_length": 608.0460691452026,
      "epoch": 0.08528784648187633,
      "grad_norm": 0.44655460119247437,
      "kl": 0.00025597214698791504,
      "learning_rate": 2.5e-06,
      "loss": 0.0209,
      "reward": 0.6188616305589676,
      "reward_std": 0.3364576445892453,
      "rewards/accuracy_reward": 0.6185826119035482,
      "rewards/format_reward": 0.00027901786961592734,
      "step": 5
    },
    {
      "completion_length": 611.7455612182617,
      "epoch": 0.17057569296375266,
      "grad_norm": 0.33932358026504517,
      "kl": 0.007887935638427735,
      "learning_rate": 2.956412726139078e-06,
      "loss": 0.0649,
      "reward": 0.6808035992085933,
      "reward_std": 0.3017056807875633,
      "rewards/accuracy_reward": 0.6808035992085933,
      "rewards/format_reward": 0.0,
      "step": 10
    },
    {
      "completion_length": 599.1960098266602,
      "epoch": 0.255863539445629,
      "grad_norm": 0.2163078784942627,
      "kl": 0.0038982391357421874,
      "learning_rate": 2.7836719084521715e-06,
      "loss": 0.0672,
      "reward": 0.763839316368103,
      "reward_std": 0.22549265380948783,
      "rewards/accuracy_reward": 0.763839316368103,
      "rewards/format_reward": 0.0,
      "step": 15
    },
    {
      "completion_length": 586.0535995483399,
      "epoch": 0.3411513859275053,
      "grad_norm": 0.19350580871105194,
      "kl": 0.17370662689208985,
      "learning_rate": 2.4946839873611927e-06,
      "loss": 0.0694,
      "reward": 0.7767857506871223,
      "reward_std": 0.18400336392223834,
      "rewards/accuracy_reward": 0.7767857506871223,
      "rewards/format_reward": 0.0,
      "step": 20
    },
    {
      "completion_length": 593.2033752441406,
      "epoch": 0.42643923240938164,
      "grad_norm": 0.10564573109149933,
      "kl": 0.005166435241699218,
      "learning_rate": 2.1156192081791355e-06,
      "loss": 0.049,
      "reward": 0.7665178924798965,
      "reward_std": 0.17804672606289387,
      "rewards/accuracy_reward": 0.7665178924798965,
      "rewards/format_reward": 0.0,
      "step": 25
    },
    {
      "completion_length": 594.5585067749023,
      "epoch": 0.511727078891258,
      "grad_norm": 0.12090254575014114,
      "kl": 0.014031219482421874,
      "learning_rate": 1.6808050203829845e-06,
      "loss": 0.0376,
      "reward": 0.7640625357627868,
      "reward_std": 0.18270524106919767,
      "rewards/accuracy_reward": 0.7640625357627868,
      "rewards/format_reward": 0.0,
      "step": 30
    },
    {
      "completion_length": 593.5230171203614,
      "epoch": 0.5970149253731343,
      "grad_norm": 0.09557133167982101,
      "kl": 0.0069671630859375,
      "learning_rate": 1.2296174432791415e-06,
      "loss": 0.0324,
      "reward": 0.7504464641213417,
      "reward_std": 0.1842412568628788,
      "rewards/accuracy_reward": 0.7504464641213417,
      "rewards/format_reward": 0.0,
      "step": 35
    },
    {
      "completion_length": 579.4169898986817,
      "epoch": 0.6823027718550106,
      "grad_norm": 0.14584919810295105,
      "kl": 0.007229232788085937,
      "learning_rate": 8.029152419343472e-07,
      "loss": 0.0363,
      "reward": 0.7756696790456772,
      "reward_std": 0.16937738354317844,
      "rewards/accuracy_reward": 0.7756696790456772,
      "rewards/format_reward": 0.0,
      "step": 40
    },
    {
      "completion_length": 594.1415420532227,
      "epoch": 0.767590618336887,
      "grad_norm": 0.1907661259174347,
      "kl": 0.006542587280273437,
      "learning_rate": 4.3933982822017883e-07,
      "loss": 0.0253,
      "reward": 0.7526786029338837,
      "reward_std": 0.18670037733390926,
      "rewards/accuracy_reward": 0.7526786029338837,
      "rewards/format_reward": 0.0,
      "step": 45
    },
    {
      "completion_length": 598.6035972595215,
      "epoch": 0.8528784648187633,
      "grad_norm": 0.11912515759468079,
      "kl": 0.007036590576171875,
      "learning_rate": 1.718159615201853e-07,
      "loss": 0.0423,
      "reward": 0.7354911103844642,
      "reward_std": 0.19358062762767075,
      "rewards/accuracy_reward": 0.7354911103844642,
      "rewards/format_reward": 0.0,
      "step": 50
    },
    {
      "completion_length": 585.3272605895996,
      "epoch": 0.9381663113006397,
      "grad_norm": 0.13946112990379333,
      "kl": 0.013706207275390625,
      "learning_rate": 2.4570139579284723e-08,
      "loss": 0.0355,
      "reward": 0.7765625357627869,
      "reward_std": 0.18904975447803735,
      "rewards/accuracy_reward": 0.7765625357627869,
      "rewards/format_reward": 0.0,
      "step": 55
    },
    {
      "completion_length": 584.9951883951823,
      "epoch": 0.9893390191897654,
      "kl": 0.004435221354166667,
      "reward": 0.768973246216774,
      "reward_std": 0.17140772442022958,
      "rewards/accuracy_reward": 0.768973246216774,
      "rewards/format_reward": 0.0,
      "step": 58,
      "total_flos": 0.0,
      "train_loss": 0.04314287324789269,
      "train_runtime": 6503.1999,
      "train_samples_per_second": 1.153,
      "train_steps_per_second": 0.009
    }
  ],
  "logging_steps": 5,
  "max_steps": 58,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}