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": 51.625,
      "epoch": 0.8,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": -0.0001,
      "reward": 1.7738320790231228,
      "reward_std": 0.002120281533279922,
      "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": 0.14883213862776756,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 2
    },
    {
      "completion_length": 53.166666666666664,
      "epoch": 1.4,
      "grad_norm": 6.040588855743408,
      "kl": 0.006199638561687,
      "learning_rate": 4.415111107797445e-07,
      "loss": 0.0,
      "reward": 2.596501588821411,
      "reward_std": 0.01238746444384257,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 2.0,
      "rewards/question_recreation_reward_func": 0.47150157888730365,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 4
    },
    {
      "completion_length": 66.5,
      "epoch": 2.0,
      "grad_norm": 0.00015045171312522143,
      "kl": 0.00036060751396386576,
      "learning_rate": 2.934120444167326e-07,
      "loss": -0.0,
      "reward": 1.5683003862698872,
      "reward_std": 0.010879872056345144,
      "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": 0.10996711999177933,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 6
    },
    {
      "completion_length": 71.75,
      "epoch": 2.8,
      "grad_norm": 6.939717769622803,
      "kl": 0.01342857114536855,
      "learning_rate": 1.2500000000000005e-07,
      "loss": 0.0,
      "reward": 1.8515289053320885,
      "reward_std": 0.01689862343482673,
      "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": 0.2265289444476366,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.125,
      "step": 8
    },
    {
      "completion_length": 53.166666666666664,
      "epoch": 3.4,
      "grad_norm": 10.517523765563965,
      "kl": 0.021501405475040276,
      "learning_rate": 1.507684480352292e-08,
      "loss": 0.0001,
      "reward": 1.6254764199256897,
      "reward_std": 0.008517678147957971,
      "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": 0.16714302822947502,
      "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": 2.5437328076804986e-06,
      "train_runtime": 1032.5939,
      "train_samples_per_second": 0.039,
      "train_steps_per_second": 0.01
    }
  ],
  "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
}