trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.989247311827957,
  "eval_steps": 100,
  "global_step": 46,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 13.15213108549844,
      "learning_rate": 1e-07,
      "logits/chosen": -3.0397768020629883,
      "logits/rejected": -2.909663200378418,
      "logps/chosen": -130.4150848388672,
      "logps/rejected": -304.69915771484375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.22,
      "grad_norm": 12.612483031980211,
      "learning_rate": 4.818756127755237e-07,
      "logits/chosen": -3.014326810836792,
      "logits/rejected": -2.936898708343506,
      "logps/chosen": -144.7748565673828,
      "logps/rejected": -202.1986541748047,
      "loss": 0.6866,
      "rewards/accuracies": 0.6527777910232544,
      "rewards/chosen": 0.0080408351495862,
      "rewards/margins": 0.009001191705465317,
      "rewards/rejected": -0.0009603556245565414,
      "step": 10
    },
    {
      "epoch": 0.43,
      "grad_norm": 12.917969062898747,
      "learning_rate": 3.52267159292835e-07,
      "logits/chosen": -3.0234923362731934,
      "logits/rejected": -2.988145112991333,
      "logps/chosen": -146.7506103515625,
      "logps/rejected": -274.9992980957031,
      "loss": 0.636,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 0.0961742028594017,
      "rewards/margins": 0.11899683624505997,
      "rewards/rejected": -0.022822635248303413,
      "step": 20
    },
    {
      "epoch": 0.65,
      "grad_norm": 13.065587999351779,
      "learning_rate": 1.6549578039787434e-07,
      "logits/chosen": -3.0107874870300293,
      "logits/rejected": -2.984440326690674,
      "logps/chosen": -151.49595642089844,
      "logps/rejected": -286.1717224121094,
      "loss": 0.5771,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 0.16383036971092224,
      "rewards/margins": 0.2549803853034973,
      "rewards/rejected": -0.09115001559257507,
      "step": 30
    },
    {
      "epoch": 0.86,
      "grad_norm": 13.06247410612875,
      "learning_rate": 2.5958610759736126e-08,
      "logits/chosen": -3.0013656616210938,
      "logits/rejected": -2.9986214637756348,
      "logps/chosen": -124.09024810791016,
      "logps/rejected": -292.10986328125,
      "loss": 0.5399,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 0.20066456496715546,
      "rewards/margins": 0.3728640079498291,
      "rewards/rejected": -0.17219944298267365,
      "step": 40
    },
    {
      "epoch": 0.99,
      "step": 46,
      "total_flos": 0.0,
      "train_loss": 0.6001577688300092,
      "train_runtime": 543.7858,
      "train_samples_per_second": 5.467,
      "train_steps_per_second": 0.085
    }
  ],
  "logging_steps": 10,
  "max_steps": 46,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}