{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 50,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 3523.8,
      "epoch": 0.1,
      "grad_norm": 0.43054767584991477,
      "kl": 6.384849548339844e-05,
      "learning_rate": 2e-05,
      "loss": 0.0,
      "reward": 1.388770055770874,
      "reward_std": 0.3910748928785324,
      "rewards/accuracy_reward": 0.35,
      "rewards/cosine_scaled_reward": 0.08877005875110626,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.950000011920929,
      "step": 5
    },
    {
      "completion_length": 3160.3,
      "epoch": 0.2,
      "grad_norm": 0.4762921847387886,
      "kl": 0.0013164520263671876,
      "learning_rate": 1.9396926207859085e-05,
      "loss": 0.0001,
      "reward": 1.651026475429535,
      "reward_std": 0.4035664960741997,
      "rewards/accuracy_reward": 0.55,
      "rewards/cosine_scaled_reward": 0.26769318282604215,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.8333333492279053,
      "step": 10
    },
    {
      "completion_length": 3849.3,
      "epoch": 0.3,
      "grad_norm": 0.4129094492269806,
      "kl": 0.0094390869140625,
      "learning_rate": 1.766044443118978e-05,
      "loss": 0.0004,
      "reward": 1.010682585835457,
      "reward_std": 0.4277887806296349,
      "rewards/accuracy_reward": 0.2,
      "rewards/cosine_scaled_reward": -0.08931739777326583,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9000000119209289,
      "step": 15
    },
    {
      "completion_length": 3270.9,
      "epoch": 0.4,
      "grad_norm": 0.37998680595283263,
      "kl": 0.018731689453125,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.0007,
      "reward": 1.2987079173326492,
      "reward_std": 0.4039229401387274,
      "rewards/accuracy_reward": 0.3,
      "rewards/cosine_scaled_reward": 0.08204126060009002,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9166666686534881,
      "step": 20
    },
    {
      "completion_length": 1608.7,
      "epoch": 0.5,
      "grad_norm": 1.2730991490794583,
      "kl": 0.046636962890625,
      "learning_rate": 1.1736481776669307e-05,
      "loss": 0.0019,
      "reward": 1.6013766705989838,
      "reward_std": 0.27621791209094226,
      "rewards/accuracy_reward": 0.55,
      "rewards/cosine_scaled_reward": 0.2347100019454956,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.8166666746139526,
      "step": 25
    },
    {
      "completion_length": 1354.2,
      "epoch": 0.6,
      "grad_norm": 1.255486410378316,
      "kl": 0.07315673828125,
      "learning_rate": 8.263518223330698e-06,
      "loss": 0.0029,
      "reward": 1.628249180316925,
      "reward_std": 0.390990880690515,
      "rewards/accuracy_reward": 0.6,
      "rewards/cosine_scaled_reward": 0.24491585344076156,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.7833333477377892,
      "step": 30
    },
    {
      "completion_length": 3099.05,
      "epoch": 0.7,
      "grad_norm": 0.5106247243030179,
      "kl": 0.10113525390625,
      "learning_rate": 5.000000000000003e-06,
      "loss": 0.0041,
      "reward": 1.235433614253998,
      "reward_std": 0.2736087866127491,
      "rewards/accuracy_reward": 0.3,
      "rewards/cosine_scaled_reward": 0.06876694485545158,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.8666666716337204,
      "step": 35
    },
    {
      "completion_length": 3219.95,
      "epoch": 0.8,
      "grad_norm": 0.4492671357745314,
      "kl": 0.022119140625,
      "learning_rate": 2.339555568810221e-06,
      "loss": 0.0009,
      "reward": 1.7884446397423743,
      "reward_std": 0.4801254738122225,
      "rewards/accuracy_reward": 0.6,
      "rewards/cosine_scaled_reward": 0.3884446397423744,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.8000000149011612,
      "step": 40
    },
    {
      "completion_length": 3702.2,
      "epoch": 0.9,
      "grad_norm": 0.5504575964230932,
      "kl": 0.022021484375,
      "learning_rate": 6.030737921409169e-07,
      "loss": 0.0009,
      "reward": 1.189654390513897,
      "reward_std": 0.5811847668141127,
      "rewards/accuracy_reward": 0.4,
      "rewards/cosine_scaled_reward": 0.12298769373446702,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.6666666775941849,
      "step": 45
    },
    {
      "completion_length": 3407.35,
      "epoch": 1.0,
      "grad_norm": 0.4356952229419395,
      "kl": 0.022021484375,
      "learning_rate": 0.0,
      "loss": 0.0009,
      "reward": 1.7177351847290994,
      "reward_std": 0.5703799482434988,
      "rewards/accuracy_reward": 0.55,
      "rewards/cosine_scaled_reward": 0.3177351653575897,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.8500000059604644,
      "step": 50
    },
    {
      "epoch": 1.0,
      "eval_completion_length": 3372.135,
      "eval_kl": 0.02354248046875,
      "eval_loss": 0.0009422407019883394,
      "eval_reward": 1.3059012717008591,
      "eval_reward_std": 0.5387985078245401,
      "eval_rewards/accuracy_reward": 0.4,
      "eval_rewards/cosine_scaled_reward": 0.1492345985211432,
      "eval_rewards/format_reward": 0.0,
      "eval_rewards/reasoning_steps_reward": 0.7566666758060455,
      "eval_runtime": 1376.7645,
      "eval_samples_per_second": 0.072,
      "eval_steps_per_second": 0.018,
      "step": 50
    },
    {
      "epoch": 1.0,
      "step": 50,
      "total_flos": 0.0,
      "train_loss": 0.0012682228069752456,
      "train_runtime": 3550.1257,
      "train_samples_per_second": 0.028,
      "train_steps_per_second": 0.014
    }
  ],
  "logging_steps": 5,
  "max_steps": 50,
  "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": 2,
  "trial_name": null,
  "trial_params": null
}