{
  "best_global_step": 40,
  "best_metric": 1.3157120974938152e-06,
  "best_model_checkpoint": "outputs/microsoft/Phi-3.5-mini-instruct/countdown_n100_mcl_256_quantized/checkpoint-30",
  "epoch": 3.0,
  "eval_steps": 5,
  "global_step": 75,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 0.02508719451725483,
      "learning_rate": 9.466666666666666e-07,
      "loss": 0.0128,
      "step": 5
    },
    {
      "epoch": 0.2,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 253.2625,
      "eval_kl": 3.578364849090576e-05,
      "eval_loss": 0.006681546568870544,
      "eval_num_tokens": 30708.0,
      "eval_reward": 0.0375,
      "eval_reward_std": 0.075,
      "eval_rewards/equation_reward_func": 0.0125,
      "eval_rewards/format_reward_func": 0.025,
      "eval_runtime": 650.0244,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 5
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.019141826778650284,
      "learning_rate": 8.799999999999999e-07,
      "loss": 0.0401,
      "step": 10
    },
    {
      "epoch": 0.4,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 249.975,
      "eval_kl": 3.7050247192382814e-05,
      "eval_loss": 0.00558798061683774,
      "eval_num_tokens": 61561.0,
      "eval_reward": 0.0375,
      "eval_reward_std": 0.075,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0375,
      "eval_runtime": 649.5002,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 10
    },
    {
      "epoch": 0.6,
      "grad_norm": 7.132045993785141e-07,
      "learning_rate": 8.133333333333333e-07,
      "loss": 0.0043,
      "step": 15
    },
    {
      "epoch": 0.6,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 244.3375,
      "eval_kl": 3.70025634765625e-05,
      "eval_loss": 0.012430070899426937,
      "eval_num_tokens": 93125.0,
      "eval_reward": 0.0625,
      "eval_reward_std": 0.08273502588272094,
      "eval_rewards/equation_reward_func": 0.05,
      "eval_rewards/format_reward_func": 0.0125,
      "eval_runtime": 649.4356,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 15
    },
    {
      "epoch": 0.8,
      "grad_norm": 1.0400003702670801e-05,
      "learning_rate": 7.466666666666667e-07,
      "loss": 0.0438,
      "step": 20
    },
    {
      "epoch": 0.8,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 252.0375,
      "eval_kl": 3.63469123840332e-05,
      "eval_loss": 0.011388374492526054,
      "eval_num_tokens": 123758.0,
      "eval_reward": 0.0125,
      "eval_reward_std": 0.025,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0125,
      "eval_runtime": 649.3474,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 20
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.02121024951338768,
      "learning_rate": 6.800000000000001e-07,
      "loss": 0.0122,
      "step": 25
    },
    {
      "epoch": 1.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 250.45,
      "eval_kl": 3.706812858581543e-05,
      "eval_loss": 0.012571494095027447,
      "eval_num_tokens": 154976.0,
      "eval_reward": 0.025,
      "eval_reward_std": 0.05,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.025,
      "eval_runtime": 650.3118,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 25
    },
    {
      "epoch": 1.2,
      "grad_norm": 6.681543254671851e-06,
      "learning_rate": 6.133333333333332e-07,
      "loss": 0.0094,
      "step": 30
    },
    {
      "epoch": 1.2,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 249.7875,
      "eval_kl": 3.7503242492675784e-05,
      "eval_loss": 1.499523023085203e-06,
      "eval_num_tokens": 186515.0,
      "eval_reward": 0.0,
      "eval_reward_std": 0.0,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 649.356,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 30
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.021921534091234207,
      "learning_rate": 5.466666666666666e-07,
      "loss": 0.0028,
      "step": 35
    },
    {
      "epoch": 1.4,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 254.8625,
      "eval_kl": 3.6659836769104e-05,
      "eval_loss": 0.005745898000895977,
      "eval_num_tokens": 217639.0,
      "eval_reward": 0.025,
      "eval_reward_std": 0.05,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.025,
      "eval_runtime": 649.0248,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 35
    },
    {
      "epoch": 1.6,
      "grad_norm": 0.023085055872797966,
      "learning_rate": 4.8e-07,
      "loss": 0.022,
      "step": 40
    },
    {
      "epoch": 1.6,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 244.175,
      "eval_kl": 3.291070461273193e-05,
      "eval_loss": 1.3157120974938152e-06,
      "eval_num_tokens": 248128.0,
      "eval_reward": 0.0,
      "eval_reward_std": 0.0,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 650.8854,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 40
    },
    {
      "epoch": 1.8,
      "grad_norm": 6.143287691884325e-07,
      "learning_rate": 4.1333333333333333e-07,
      "loss": 0.0119,
      "step": 45
    },
    {
      "epoch": 1.8,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 246.9375,
      "eval_kl": 3.790557384490967e-05,
      "eval_loss": 0.00949961505830288,
      "eval_num_tokens": 279151.0,
      "eval_reward": 0.025,
      "eval_reward_std": 0.05,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.025,
      "eval_runtime": 650.0193,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 45
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.366271693674207e-07,
      "learning_rate": 3.4666666666666665e-07,
      "loss": 0.0195,
      "step": 50
    },
    {
      "epoch": 2.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 248.4875,
      "eval_kl": 4.260838031768799e-05,
      "eval_loss": 1.7038606756614172e-06,
      "eval_num_tokens": 310444.0,
      "eval_reward": 0.0,
      "eval_reward_std": 0.0,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 649.8376,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 50
    },
    {
      "epoch": 2.2,
      "grad_norm": 1.6701025060683605e-06,
      "learning_rate": 2.8e-07,
      "loss": 0.0,
      "step": 55
    },
    {
      "epoch": 2.2,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 246.7875,
      "eval_kl": 3.412365913391113e-05,
      "eval_loss": 0.002006751950830221,
      "eval_num_tokens": 341912.0,
      "eval_reward": 0.0375,
      "eval_reward_std": 0.075,
      "eval_rewards/equation_reward_func": 0.0375,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 648.9136,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 55
    },
    {
      "epoch": 2.4,
      "grad_norm": 0.024540750309824944,
      "learning_rate": 2.1333333333333334e-07,
      "loss": 0.0154,
      "step": 60
    },
    {
      "epoch": 2.4,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 245.5125,
      "eval_kl": 3.3614039421081546e-05,
      "eval_loss": 0.003208440961316228,
      "eval_num_tokens": 372741.0,
      "eval_reward": 0.0125,
      "eval_reward_std": 0.025,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.0125,
      "eval_runtime": 649.682,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 60
    },
    {
      "epoch": 2.6,
      "grad_norm": 0.021638961508870125,
      "learning_rate": 1.4666666666666666e-07,
      "loss": 0.0327,
      "step": 65
    },
    {
      "epoch": 2.6,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 248.375,
      "eval_kl": 3.7464499473571775e-05,
      "eval_loss": 0.008837858214974403,
      "eval_num_tokens": 403457.0,
      "eval_reward": 0.025,
      "eval_reward_std": 0.05,
      "eval_rewards/equation_reward_func": 0.0,
      "eval_rewards/format_reward_func": 0.025,
      "eval_runtime": 650.2835,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 65
    },
    {
      "epoch": 2.8,
      "grad_norm": 9.145921353592712e-07,
      "learning_rate": 8e-08,
      "loss": 0.0034,
      "step": 70
    },
    {
      "epoch": 2.8,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 245.85,
      "eval_kl": 4.0978193283081055e-05,
      "eval_loss": 0.020375821739435196,
      "eval_num_tokens": 434973.0,
      "eval_reward": 0.05,
      "eval_reward_std": 0.1,
      "eval_rewards/equation_reward_func": 0.025,
      "eval_rewards/format_reward_func": 0.025,
      "eval_runtime": 648.4803,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 70
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.0,
      "learning_rate": 1.3333333333333334e-08,
      "loss": 0.0,
      "step": 75
    },
    {
      "epoch": 3.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 247.1125,
      "eval_kl": 3.719329833984375e-05,
      "eval_loss": 0.008141107857227325,
      "eval_num_tokens": 466413.0,
      "eval_reward": 0.025,
      "eval_reward_std": 0.05,
      "eval_rewards/equation_reward_func": 0.025,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 649.5897,
      "eval_samples_per_second": 0.031,
      "eval_steps_per_second": 0.008,
      "step": 75
    }
  ],
  "logging_steps": 5,
  "max_steps": 75,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 15,
  "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
}