trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.4,
  "eval_steps": 500,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 71.375,
      "epoch": 0.8,
      "grad_norm": 4.6866044998168945,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 1.289890080690384,
      "reward_std": 0.017315300239715725,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 1.0,
      "rewards/question_recreation_reward_func": 0.16489006020128727,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 2
    },
    {
      "completion_length": 58.0,
      "epoch": 1.4,
      "grad_norm": 13.606494903564453,
      "kl": 0.0015446399817543959,
      "learning_rate": 4.415111107797445e-07,
      "loss": 0.0,
      "reward": 1.874943772951762,
      "reward_std": 0.04934494259456793,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 1.3333333333333333,
      "rewards/question_recreation_reward_func": 0.4166103998819987,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 4
    },
    {
      "completion_length": 96.83333333333333,
      "epoch": 2.0,
      "grad_norm": 6.199663039296865e-05,
      "kl": 0.002855731061572442,
      "learning_rate": 2.934120444167326e-07,
      "loss": 0.0,
      "reward": 1.6040082772572835,
      "reward_std": 0.014235751082499823,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 1.3333333333333333,
      "rewards/question_recreation_reward_func": 0.14567496379216513,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 6
    },
    {
      "completion_length": 75.375,
      "epoch": 2.8,
      "grad_norm": 4.742353916168213,
      "kl": 0.016857809572684346,
      "learning_rate": 1.2500000000000005e-07,
      "loss": 0.0,
      "reward": 1.8766698390245438,
      "reward_std": 0.034587837872095406,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 1.5,
      "rewards/question_recreation_reward_func": 0.2516698334366083,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 8
    },
    {
      "completion_length": 96.66666666666667,
      "epoch": 3.4,
      "grad_norm": 5.535033226013184,
      "kl": 0.010410198903021714,
      "learning_rate": 1.507684480352292e-08,
      "loss": 0.0,
      "reward": 1.0010255078474681,
      "reward_std": 0.06484196474775672,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.6666666666666666,
      "rewards/question_recreation_reward_func": 0.20935882876316705,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 10
    },
    {
      "epoch": 3.4,
      "step": 10,
      "total_flos": 0.0,
      "train_loss": 9.378274330629211e-06,
      "train_runtime": 526.1011,
      "train_samples_per_second": 0.076,
      "train_steps_per_second": 0.019
    }
  ],
  "logging_steps": 2,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 10,
  "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": 2,
  "trial_name": null,
  "trial_params": null
}