trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9927007299270073,
  "eval_steps": 500,
  "global_step": 68,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.291970802919708,
      "grad_norm": 3.5247514247894287,
      "learning_rate": 7.136479226306842e-06,
      "log_odds_chosen": 6.104650020599365,
      "log_odds_ratio": -0.028288718312978745,
      "logits/chosen": -1.0308363437652588,
      "logits/rejected": -1.0280412435531616,
      "logps/chosen": -0.8738629221916199,
      "logps/rejected": -6.279038429260254,
      "loss": 0.6595,
      "nll_loss": 0.6566511392593384,
      "rewards/accuracies": 0.9906250238418579,
      "rewards/chosen": -0.08738629519939423,
      "rewards/margins": 0.5405175685882568,
      "rewards/rejected": -0.6279038190841675,
      "step": 20
    },
    {
      "epoch": 0.583941605839416,
      "grad_norm": 3.725033760070801,
      "learning_rate": 3.4864065794137955e-06,
      "log_odds_chosen": 6.303352355957031,
      "log_odds_ratio": -0.01220815908163786,
      "logits/chosen": -0.9482170939445496,
      "logits/rejected": -0.9314813613891602,
      "logps/chosen": -0.7882482409477234,
      "logps/rejected": -6.333585262298584,
      "loss": 0.6424,
      "nll_loss": 0.6412121653556824,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -0.07882483303546906,
      "rewards/margins": 0.5545337200164795,
      "rewards/rejected": -0.6333585381507874,
      "step": 40
    },
    {
      "epoch": 0.8759124087591241,
      "grad_norm": 3.5018422603607178,
      "learning_rate": 3.347323820519814e-07,
      "log_odds_chosen": 6.304269790649414,
      "log_odds_ratio": -0.02359706163406372,
      "logits/chosen": -0.9824663996696472,
      "logits/rejected": -0.9669343829154968,
      "logps/chosen": -0.7988778352737427,
      "logps/rejected": -6.3369140625,
      "loss": 0.6456,
      "nll_loss": 0.6432429552078247,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -0.07988778501749039,
      "rewards/margins": 0.553803563117981,
      "rewards/rejected": -0.6336914300918579,
      "step": 60
    }
  ],
  "logging_steps": 20,
  "max_steps": 68,
  "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": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}