{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9893390191897655,
  "eval_steps": 39,
  "global_step": 116,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_completion_length": 580.933360529403,
      "eval_kl": 0.0,
      "eval_loss": 1.623958087293431e-05,
      "eval_reward": 0.5275844605966878,
      "eval_reward_std": 0.35651993932434545,
      "eval_rewards/accuracy_reward": 0.5264149032080897,
      "eval_rewards/format_reward": 0.0011695573350396782,
      "eval_runtime": 4528.6778,
      "eval_samples_per_second": 1.104,
      "eval_steps_per_second": 0.01,
      "step": 0
    },
    {
      "completion_length": 604.259627532959,
      "epoch": 0.08528784648187633,
      "grad_norm": 1.9525558948516846,
      "kl": 0.00012731552124023438,
      "learning_rate": 1.25e-06,
      "loss": 0.0,
      "reward": 0.6582589574158192,
      "reward_std": 0.3278909016400576,
      "rewards/accuracy_reward": 0.6573660992085933,
      "rewards/format_reward": 0.0008928571827709675,
      "step": 5
    },
    {
      "completion_length": 616.6596267700195,
      "epoch": 0.17057569296375266,
      "grad_norm": 1.6776472330093384,
      "kl": 0.0016300678253173828,
      "learning_rate": 2.5e-06,
      "loss": 0.0001,
      "reward": 0.6910714641213417,
      "reward_std": 0.3150623720139265,
      "rewards/accuracy_reward": 0.6908482506871223,
      "rewards/format_reward": 0.00022321429569274187,
      "step": 10
    },
    {
      "completion_length": 640.9160995483398,
      "epoch": 0.255863539445629,
      "grad_norm": 0.13862861692905426,
      "kl": 0.004425430297851562,
      "learning_rate": 2.9938448364256362e-06,
      "loss": 0.0002,
      "reward": 0.7236607506871223,
      "reward_std": 0.26740851644426583,
      "rewards/accuracy_reward": 0.7236607506871223,
      "rewards/format_reward": 0.0,
      "step": 15
    },
    {
      "completion_length": 631.7879745483399,
      "epoch": 0.3411513859275053,
      "grad_norm": 0.13688388466835022,
      "kl": 0.007097053527832031,
      "learning_rate": 2.956412726139078e-06,
      "loss": 0.0003,
      "reward": 0.7348214671015739,
      "reward_std": 0.22995020002126693,
      "rewards/accuracy_reward": 0.7348214671015739,
      "rewards/format_reward": 0.0,
      "step": 20
    },
    {
      "completion_length": 619.988638305664,
      "epoch": 0.42643923240938164,
      "grad_norm": 0.8423397541046143,
      "kl": 0.004146957397460937,
      "learning_rate": 2.88581929876693e-06,
      "loss": 0.0002,
      "reward": 0.7430803924798965,
      "reward_std": 0.22932490669190883,
      "rewards/accuracy_reward": 0.7430803924798965,
      "rewards/format_reward": 0.0,
      "step": 25
    },
    {
      "completion_length": 606.4966804504395,
      "epoch": 0.511727078891258,
      "grad_norm": 0.11532779783010483,
      "kl": 0.00457305908203125,
      "learning_rate": 2.7836719084521715e-06,
      "loss": 0.0002,
      "reward": 0.752455385029316,
      "reward_std": 0.206120328605175,
      "rewards/accuracy_reward": 0.752455385029316,
      "rewards/format_reward": 0.0,
      "step": 30
    },
    {
      "completion_length": 589.5859703063965,
      "epoch": 0.5970149253731343,
      "grad_norm": 0.10156127065420151,
      "kl": 0.004618453979492188,
      "learning_rate": 2.652296367060421e-06,
      "loss": 0.0002,
      "reward": 0.783928607404232,
      "reward_std": 0.19030314590781927,
      "rewards/accuracy_reward": 0.783928607404232,
      "rewards/format_reward": 0.0,
      "step": 35
    },
    {
      "epoch": 0.6652452025586354,
      "eval_completion_length": 582.3961823836617,
      "eval_kl": 0.005152472896852355,
      "eval_loss": 0.00021490754443220794,
      "eval_reward": 0.6949793272260306,
      "eval_reward_std": 0.20541082589306695,
      "eval_rewards/accuracy_reward": 0.6949793272260306,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 4509.8983,
      "eval_samples_per_second": 1.109,
      "eval_steps_per_second": 0.01,
      "step": 39
    },
    {
      "completion_length": 591.9185485839844,
      "epoch": 0.6823027718550106,
      "grad_norm": 0.12734617292881012,
      "kl": 0.005767822265625,
      "learning_rate": 2.4946839873611927e-06,
      "loss": 0.0002,
      "reward": 0.7678571790456772,
      "reward_std": 0.2127251597121358,
      "rewards/accuracy_reward": 0.7678571790456772,
      "rewards/format_reward": 0.0,
      "step": 40
    },
    {
      "completion_length": 592.6504753112793,
      "epoch": 0.767590618336887,
      "grad_norm": 0.08878223598003387,
      "kl": 0.00388336181640625,
      "learning_rate": 2.314423473302218e-06,
      "loss": 0.0002,
      "reward": 0.7566964656114579,
      "reward_std": 0.1911870390176773,
      "rewards/accuracy_reward": 0.7566964656114579,
      "rewards/format_reward": 0.0,
      "step": 45
    },
    {
      "completion_length": 579.0308303833008,
      "epoch": 0.8528784648187633,
      "grad_norm": 0.10042494535446167,
      "kl": 0.006903076171875,
      "learning_rate": 2.1156192081791355e-06,
      "loss": 0.0003,
      "reward": 0.7850446835160255,
      "reward_std": 0.17857401072978973,
      "rewards/accuracy_reward": 0.7850446835160255,
      "rewards/format_reward": 0.0,
      "step": 50
    },
    {
      "completion_length": 574.0143119812012,
      "epoch": 0.9381663113006397,
      "grad_norm": 0.10724026709794998,
      "kl": 0.00340118408203125,
      "learning_rate": 1.9027978012115653e-06,
      "loss": 0.0001,
      "reward": 0.7665178909897804,
      "reward_std": 0.16513762380927802,
      "rewards/accuracy_reward": 0.7665178909897804,
      "rewards/format_reward": 0.0,
      "step": 55
    },
    {
      "completion_length": 575.9379244717685,
      "epoch": 1.0341151385927505,
      "grad_norm": 0.07216066867113113,
      "kl": 0.0033409812233664775,
      "learning_rate": 1.6808050203829845e-06,
      "loss": 0.0001,
      "reward": 0.7869318506934426,
      "reward_std": 0.18544116904112426,
      "rewards/accuracy_reward": 0.7869318506934426,
      "rewards/format_reward": 0.0,
      "step": 60
    },
    {
      "completion_length": 565.3899772644043,
      "epoch": 1.1194029850746268,
      "grad_norm": 0.09482983499765396,
      "kl": 0.0034759521484375,
      "learning_rate": 1.454695458298667e-06,
      "loss": 0.0001,
      "reward": 0.7991071745753289,
      "reward_std": 0.17083449736237527,
      "rewards/accuracy_reward": 0.7991071745753289,
      "rewards/format_reward": 0.0,
      "step": 65
    },
    {
      "completion_length": 571.7268089294433,
      "epoch": 1.2046908315565032,
      "grad_norm": 0.08009294420480728,
      "kl": 0.00347137451171875,
      "learning_rate": 1.2296174432791415e-06,
      "loss": 0.0001,
      "reward": 0.7839286103844643,
      "reward_std": 0.17667411090806126,
      "rewards/accuracy_reward": 0.7837053954601287,
      "rewards/format_reward": 0.00022321429569274187,
      "step": 70
    },
    {
      "completion_length": 565.4739059448242,
      "epoch": 1.2899786780383795,
      "grad_norm": 0.06680043786764145,
      "kl": 0.00329742431640625,
      "learning_rate": 1.0106958161686963e-06,
      "loss": 0.0001,
      "reward": 0.7879464671015739,
      "reward_std": 0.1701756412163377,
      "rewards/accuracy_reward": 0.7879464671015739,
      "rewards/format_reward": 0.0,
      "step": 75
    },
    {
      "epoch": 1.3411513859275053,
      "eval_completion_length": 560.9655020060101,
      "eval_kl": 0.004444744891158197,
      "eval_loss": 0.00021425970771815628,
      "eval_reward": 0.6994489503012921,
      "eval_reward_std": 0.208146112796903,
      "eval_rewards/accuracy_reward": 0.6994489503012921,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 4435.898,
      "eval_samples_per_second": 1.127,
      "eval_steps_per_second": 0.01,
      "step": 78
    },
    {
      "completion_length": 551.9654197692871,
      "epoch": 1.375266524520256,
      "grad_norm": 0.1283424198627472,
      "kl": 0.0043430328369140625,
      "learning_rate": 8.029152419343472e-07,
      "loss": 0.0002,
      "reward": 0.7924107536673546,
      "reward_std": 0.17128037009388208,
      "rewards/accuracy_reward": 0.7924107536673546,
      "rewards/format_reward": 0.0,
      "step": 80
    },
    {
      "completion_length": 541.5716796875,
      "epoch": 1.4605543710021323,
      "grad_norm": 0.3723917603492737,
      "kl": 0.003989791870117188,
      "learning_rate": 6.11006712953975e-07,
      "loss": 0.0002,
      "reward": 0.8220982536673546,
      "reward_std": 0.15211040936410428,
      "rewards/accuracy_reward": 0.8220982536673546,
      "rewards/format_reward": 0.0,
      "step": 85
    },
    {
      "completion_length": 563.2312767028809,
      "epoch": 1.5458422174840085,
      "grad_norm": 0.09019841253757477,
      "kl": 0.0033016204833984375,
      "learning_rate": 4.3933982822017883e-07,
      "loss": 0.0001,
      "reward": 0.7973214641213417,
      "reward_std": 0.17157677421346307,
      "rewards/accuracy_reward": 0.7973214641213417,
      "rewards/format_reward": 0.0,
      "step": 90
    },
    {
      "completion_length": 554.155606842041,
      "epoch": 1.6311300639658848,
      "grad_norm": 0.06605476140975952,
      "kl": 0.0035243988037109374,
      "learning_rate": 2.9182330117358096e-07,
      "loss": 0.0001,
      "reward": 0.799107177555561,
      "reward_std": 0.15023288973607124,
      "rewards/accuracy_reward": 0.799107177555561,
      "rewards/format_reward": 0.0,
      "step": 95
    },
    {
      "completion_length": 562.7194396972657,
      "epoch": 1.716417910447761,
      "grad_norm": 0.09438558667898178,
      "kl": 0.003635406494140625,
      "learning_rate": 1.718159615201853e-07,
      "loss": 0.0001,
      "reward": 0.7908482491970062,
      "reward_std": 0.1763416165485978,
      "rewards/accuracy_reward": 0.7908482491970062,
      "rewards/format_reward": 0.0,
      "step": 100
    },
    {
      "completion_length": 581.4640892028808,
      "epoch": 1.8017057569296375,
      "grad_norm": 0.08443931490182877,
      "kl": 0.003319549560546875,
      "learning_rate": 8.20502774480395e-08,
      "loss": 0.0001,
      "reward": 0.7691964648663998,
      "reward_std": 0.178659126162529,
      "rewards/accuracy_reward": 0.7691964648663998,
      "rewards/format_reward": 0.0,
      "step": 105
    },
    {
      "completion_length": 578.9288177490234,
      "epoch": 1.886993603411514,
      "grad_norm": 0.09037923067808151,
      "kl": 0.0033283233642578125,
      "learning_rate": 2.4570139579284723e-08,
      "loss": 0.0001,
      "reward": 0.7825893223285675,
      "reward_std": 0.16389768896624446,
      "rewards/accuracy_reward": 0.7825893223285675,
      "rewards/format_reward": 0.0,
      "step": 110
    },
    {
      "completion_length": 573.8774856567383,
      "epoch": 1.9722814498933903,
      "grad_norm": 0.06668173521757126,
      "kl": 0.0031665802001953126,
      "learning_rate": 6.843232656998933e-10,
      "loss": 0.0001,
      "reward": 0.7763393208384514,
      "reward_std": 0.15611171433702112,
      "rewards/accuracy_reward": 0.7763393208384514,
      "rewards/format_reward": 0.0,
      "step": 115
    },
    {
      "completion_length": 572.2143058776855,
      "epoch": 1.9893390191897655,
      "kl": 0.00342559814453125,
      "reward": 0.7700893357396126,
      "reward_std": 0.18100517243146896,
      "rewards/accuracy_reward": 0.7700893357396126,
      "rewards/format_reward": 0.0,
      "step": 116,
      "total_flos": 0.0,
      "train_loss": 0.0002975721653738533,
      "train_runtime": 28855.0053,
      "train_samples_per_second": 0.52,
      "train_steps_per_second": 0.004
    }
  ],
  "logging_steps": 5,
  "max_steps": 116,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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
}