{
  "best_global_step": 25,
  "best_metric": 1.090248929358495e-06,
  "best_model_checkpoint": "outputs/microsoft/Phi-3.5-mini-instruct/aime_n89_mcl_256_pretrained/checkpoint-25",
  "epoch": 2.98876404494382,
  "eval_steps": 5,
  "global_step": 66,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2247191011235955,
      "grad_norm": 0.0,
      "learning_rate": 9.393939393939395e-07,
      "loss": 0.0,
      "step": 5
    },
    {
      "epoch": 0.2247191011235955,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 254.57142857142858,
      "eval_kl": 2.9193503516060965e-05,
      "eval_loss": 0.009287942200899124,
      "eval_num_tokens": 38036.0,
      "eval_reward": 0.017857142857142856,
      "eval_reward_std": 0.03571428571428571,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 215.8763,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 5
    },
    {
      "epoch": 0.449438202247191,
      "grad_norm": 0.0,
      "learning_rate": 8.636363636363636e-07,
      "loss": 0.0,
      "step": 10
    },
    {
      "epoch": 0.449438202247191,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.80357142857142,
      "eval_kl": 2.65155519757952e-05,
      "eval_loss": 0.001207446912303567,
      "eval_num_tokens": 76032.0,
      "eval_reward": 0.03571428571428571,
      "eval_reward_std": 0.07142857142857142,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 215.5058,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 10
    },
    {
      "epoch": 0.6741573033707865,
      "grad_norm": 0.023744648322463036,
      "learning_rate": 7.878787878787878e-07,
      "loss": 0.0017,
      "step": 15
    },
    {
      "epoch": 0.6741573033707865,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.35714285714286,
      "eval_kl": 2.9138156345912387e-05,
      "eval_loss": 0.004041091073304415,
      "eval_num_tokens": 112150.0,
      "eval_reward": 0.05357142857142857,
      "eval_reward_std": 0.10714285714285714,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.03571428571428571,
      "eval_runtime": 215.0271,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 15
    },
    {
      "epoch": 0.898876404494382,
      "grad_norm": 0.0,
      "learning_rate": 7.121212121212121e-07,
      "loss": 0.0,
      "step": 20
    },
    {
      "epoch": 0.898876404494382,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.67857142857142,
      "eval_kl": 2.76097229548863e-05,
      "eval_loss": 0.0019887436646968126,
      "eval_num_tokens": 154187.0,
      "eval_reward": 0.017857142857142856,
      "eval_reward_std": 0.03571428571428571,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 214.6133,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 20
    },
    {
      "epoch": 1.1348314606741572,
      "grad_norm": 0.0,
      "learning_rate": 6.363636363636363e-07,
      "loss": 0.0059,
      "step": 25
    },
    {
      "epoch": 1.1348314606741572,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 256.0,
      "eval_kl": 2.7860913957868305e-05,
      "eval_loss": 1.090248929358495e-06,
      "eval_num_tokens": 193516.0,
      "eval_reward": 0.0,
      "eval_reward_std": 0.0,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_rewards/returns_int_reward_func": 0.0,
      "eval_runtime": 214.372,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 25
    },
    {
      "epoch": 1.3595505617977528,
      "grad_norm": 0.0,
      "learning_rate": 5.606060606060605e-07,
      "loss": 0.0076,
      "step": 30
    },
    {
      "epoch": 1.3595505617977528,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.875,
      "eval_kl": 2.8052500316074918e-05,
      "eval_loss": 0.0006666756235063076,
      "eval_num_tokens": 231687.0,
      "eval_reward": 0.05357142857142857,
      "eval_reward_std": 0.0683876531464713,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.03571428571428571,
      "eval_runtime": 214.4867,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 30
    },
    {
      "epoch": 1.5842696629213484,
      "grad_norm": 0.0,
      "learning_rate": 4.848484848484849e-07,
      "loss": 0.0,
      "step": 35
    },
    {
      "epoch": 1.5842696629213484,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.10714285714286,
      "eval_kl": 2.9695885522024973e-05,
      "eval_loss": 0.005562338512390852,
      "eval_num_tokens": 270259.0,
      "eval_reward": 0.05357142857142857,
      "eval_reward_std": 0.10714285714285714,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.03571428571428571,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 214.5264,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 35
    },
    {
      "epoch": 1.8089887640449438,
      "grad_norm": 0.0,
      "learning_rate": 4.090909090909091e-07,
      "loss": 0.0,
      "step": 40
    },
    {
      "epoch": 1.8089887640449438,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.08928571428572,
      "eval_kl": 2.831646374293736e-05,
      "eval_loss": 0.005824462044984102,
      "eval_num_tokens": 308803.0,
      "eval_reward": 0.03571428571428571,
      "eval_reward_std": 0.07142857142857142,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 214.9441,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 40
    },
    {
      "epoch": 2.044943820224719,
      "grad_norm": 1.279886532756791e-06,
      "learning_rate": 3.333333333333333e-07,
      "loss": 0.0,
      "step": 45
    },
    {
      "epoch": 2.044943820224719,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 256.0,
      "eval_kl": 2.8508050101143973e-05,
      "eval_loss": 1.099406858884322e-06,
      "eval_num_tokens": 347075.0,
      "eval_reward": 0.0,
      "eval_reward_std": 0.0,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_rewards/returns_int_reward_func": 0.0,
      "eval_runtime": 214.5574,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 45
    },
    {
      "epoch": 2.2696629213483144,
      "grad_norm": 0.02009560726583004,
      "learning_rate": 2.5757575757575754e-07,
      "loss": 0.0,
      "step": 50
    },
    {
      "epoch": 2.2696629213483144,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 254.96428571428572,
      "eval_kl": 3.0040740966796875e-05,
      "eval_loss": 0.0017646737396717072,
      "eval_num_tokens": 384887.0,
      "eval_reward": 0.03571428571428571,
      "eval_reward_std": 0.07142857142857142,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 214.6312,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 50
    },
    {
      "epoch": 2.49438202247191,
      "grad_norm": 0.0,
      "learning_rate": 1.818181818181818e-07,
      "loss": 0.0043,
      "step": 55
    },
    {
      "epoch": 2.49438202247191,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 256.0,
      "eval_kl": 2.7247837611607144e-05,
      "eval_loss": 1.1022784747183323e-06,
      "eval_num_tokens": 422399.0,
      "eval_reward": 0.0,
      "eval_reward_std": 0.0,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_rewards/returns_int_reward_func": 0.0,
      "eval_runtime": 214.4993,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 55
    },
    {
      "epoch": 2.7191011235955056,
      "grad_norm": 0.018734661862254143,
      "learning_rate": 1.0606060606060605e-07,
      "loss": 0.0,
      "step": 60
    },
    {
      "epoch": 2.7191011235955056,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 254.64285714285714,
      "eval_kl": 2.9112611498151508e-05,
      "eval_loss": 0.008906994014978409,
      "eval_num_tokens": 461252.0,
      "eval_reward": 0.03571428571428571,
      "eval_reward_std": 0.07142857142857142,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 214.653,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 60
    },
    {
      "epoch": 2.943820224719101,
      "grad_norm": 0.019450360909104347,
      "learning_rate": 3.0303030303030305e-08,
      "loss": 0.0047,
      "step": 65
    },
    {
      "epoch": 2.943820224719101,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 255.73214285714286,
      "eval_kl": 3.141164779663086e-05,
      "eval_loss": 0.0016528593841940165,
      "eval_num_tokens": 500984.0,
      "eval_reward": 0.03571428571428571,
      "eval_reward_std": 0.07142857142857142,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.017857142857142856,
      "eval_rewards/returns_int_reward_func": 0.017857142857142856,
      "eval_runtime": 214.5471,
      "eval_samples_per_second": 0.065,
      "eval_steps_per_second": 0.019,
      "step": 65
    }
  ],
  "logging_steps": 5,
  "max_steps": 66,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 5,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}