{
  "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": 608.9531555175781,
      "epoch": 0.017057569296375266,
      "grad_norm": 2.6430623531341553,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 0.647321455180645,
      "reward_std": 0.32679445296525955,
      "rewards/accuracy_reward": 0.6462053880095482,
      "rewards/format_reward": 0.0011160714784637094,
      "step": 1
    },
    {
      "completion_length": 617.5315542221069,
      "epoch": 0.08528784648187633,
      "grad_norm": 2.203550338745117,
      "kl": 0.00031819939613342285,
      "learning_rate": 2.5e-06,
      "loss": 0.0,
      "reward": 0.6328125279396772,
      "reward_std": 0.3299180152826011,
      "rewards/accuracy_reward": 0.6325335092842579,
      "rewards/format_reward": 0.00027901786961592734,
      "step": 5
    },
    {
      "completion_length": 614.0942283630371,
      "epoch": 0.17057569296375266,
      "grad_norm": 23014.91015625,
      "kl": 1.7154921531677245,
      "learning_rate": 2.956412726139078e-06,
      "loss": 0.0684,
      "reward": 0.7095982447266579,
      "reward_std": 0.2823256587609649,
      "rewards/accuracy_reward": 0.7089286014437676,
      "rewards/format_reward": 0.0006696428870782256,
      "step": 10
    },
    {
      "completion_length": 627.2895385742188,
      "epoch": 0.255863539445629,
      "grad_norm": 17940.5390625,
      "kl": 2.3679866790771484,
      "learning_rate": 2.7836719084521715e-06,
      "loss": 0.0948,
      "reward": 0.7515625342726707,
      "reward_std": 0.24004750959575177,
      "rewards/accuracy_reward": 0.7515625342726707,
      "rewards/format_reward": 0.0,
      "step": 15
    },
    {
      "completion_length": 611.387979888916,
      "epoch": 0.3411513859275053,
      "grad_norm": 10021.2578125,
      "kl": 3.927571487426758,
      "learning_rate": 2.4946839873611927e-06,
      "loss": 0.1568,
      "reward": 0.7631696775555611,
      "reward_std": 0.2211073562502861,
      "rewards/accuracy_reward": 0.7631696775555611,
      "rewards/format_reward": 0.0,
      "step": 20
    },
    {
      "completion_length": 619.4598495483399,
      "epoch": 0.42643923240938164,
      "grad_norm": 0.10919574648141861,
      "kl": 0.0038677215576171874,
      "learning_rate": 2.1156192081791355e-06,
      "loss": 0.0002,
      "reward": 0.7560268223285675,
      "reward_std": 0.20058796666562556,
      "rewards/accuracy_reward": 0.7560268223285675,
      "rewards/format_reward": 0.0,
      "step": 25
    },
    {
      "completion_length": 620.3272598266601,
      "epoch": 0.511727078891258,
      "grad_norm": 0.1221357136964798,
      "kl": 0.0031772613525390624,
      "learning_rate": 1.6808050203829845e-06,
      "loss": 0.0001,
      "reward": 0.7466518178582191,
      "reward_std": 0.2014338149689138,
      "rewards/accuracy_reward": 0.7466518178582191,
      "rewards/format_reward": 0.0,
      "step": 30
    },
    {
      "completion_length": 605.8370811462403,
      "epoch": 0.5970149253731343,
      "grad_norm": 0.12291086465120316,
      "kl": 0.0034273147583007814,
      "learning_rate": 1.2296174432791415e-06,
      "loss": 0.0001,
      "reward": 0.7455357447266578,
      "reward_std": 0.1848256574012339,
      "rewards/accuracy_reward": 0.7455357447266578,
      "rewards/format_reward": 0.0,
      "step": 35
    },
    {
      "completion_length": 599.0658744812011,
      "epoch": 0.6823027718550106,
      "grad_norm": 0.41419172286987305,
      "kl": 0.003479766845703125,
      "learning_rate": 8.029152419343472e-07,
      "loss": 0.0001,
      "reward": 0.758482177555561,
      "reward_std": 0.19037142880260943,
      "rewards/accuracy_reward": 0.758482177555561,
      "rewards/format_reward": 0.0,
      "step": 40
    },
    {
      "completion_length": 612.0359664916992,
      "epoch": 0.767590618336887,
      "grad_norm": 0.51832515001297,
      "kl": 0.003022003173828125,
      "learning_rate": 4.3933982822017883e-07,
      "loss": 0.0001,
      "reward": 0.7497768223285675,
      "reward_std": 0.1870642237365246,
      "rewards/accuracy_reward": 0.7497768223285675,
      "rewards/format_reward": 0.0,
      "step": 45
    },
    {
      "completion_length": 614.3917701721191,
      "epoch": 0.8528784648187633,
      "grad_norm": 0.08213239908218384,
      "kl": 0.03495864868164063,
      "learning_rate": 1.718159615201853e-07,
      "loss": 0.0014,
      "reward": 0.7412946790456771,
      "reward_std": 0.19202891541644931,
      "rewards/accuracy_reward": 0.7412946790456771,
      "rewards/format_reward": 0.0,
      "step": 50
    },
    {
      "completion_length": 604.3663208007813,
      "epoch": 0.9381663113006397,
      "grad_norm": 0.5027450323104858,
      "kl": 0.0035785675048828126,
      "learning_rate": 2.4570139579284723e-08,
      "loss": 0.0001,
      "reward": 0.772098246216774,
      "reward_std": 0.18952333889901637,
      "rewards/accuracy_reward": 0.772098246216774,
      "rewards/format_reward": 0.0,
      "step": 55
    },
    {
      "completion_length": 594.7775586446127,
      "epoch": 0.9893390191897654,
      "kl": 0.0032895406087239585,
      "reward": 0.772693489988645,
      "reward_std": 0.17002321127802134,
      "rewards/accuracy_reward": 0.772693489988645,
      "rewards/format_reward": 0.0,
      "step": 58,
      "total_flos": 0.0,
      "train_loss": 0.027603623356527666,
      "train_runtime": 13992.0401,
      "train_samples_per_second": 0.536,
      "train_steps_per_second": 0.004
    }
  ],
  "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
}