trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.985781990521327,
  "eval_steps": 100,
  "global_step": 52,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 979.150375266187,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": 124.78954315185547,
      "logits/rejected": 100.39772033691406,
      "logps/chosen": -796.0274658203125,
      "logps/rejected": -794.6148071289062,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.19,
      "grad_norm": 1217.072788634617,
      "learning_rate": 4.907293218369498e-07,
      "logits/chosen": 113.76150512695312,
      "logits/rejected": 135.77798461914062,
      "logps/chosen": -777.50439453125,
      "logps/rejected": -873.83935546875,
      "loss": 1.2776,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 2.8628532886505127,
      "rewards/margins": 0.8314082622528076,
      "rewards/rejected": 2.031445026397705,
      "step": 10
    },
    {
      "epoch": 0.38,
      "grad_norm": 743.1615361693152,
      "learning_rate": 3.941700805287168e-07,
      "logits/chosen": 126.83177185058594,
      "logits/rejected": 138.27816772460938,
      "logps/chosen": -776.5384521484375,
      "logps/rejected": -864.4357299804688,
      "loss": 1.1766,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -1.2370330095291138,
      "rewards/margins": 2.700045108795166,
      "rewards/rejected": -3.9370779991149902,
      "step": 20
    },
    {
      "epoch": 0.57,
      "grad_norm": 828.8687529436529,
      "learning_rate": 2.3293939665883228e-07,
      "logits/chosen": 125.66447448730469,
      "logits/rejected": 129.14175415039062,
      "logps/chosen": -809.836669921875,
      "logps/rejected": -870.1324462890625,
      "loss": 1.2,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.1635781079530716,
      "rewards/margins": 3.582213878631592,
      "rewards/rejected": -3.7457919120788574,
      "step": 30
    },
    {
      "epoch": 0.76,
      "grad_norm": 738.6354009412435,
      "learning_rate": 7.936171419533652e-08,
      "logits/chosen": 126.02899169921875,
      "logits/rejected": 121.68770599365234,
      "logps/chosen": -817.9525756835938,
      "logps/rejected": -828.2701416015625,
      "loss": 1.1216,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 0.8308059573173523,
      "rewards/margins": 2.3288683891296387,
      "rewards/rejected": -1.4980621337890625,
      "step": 40
    },
    {
      "epoch": 0.95,
      "grad_norm": 713.4723758027491,
      "learning_rate": 2.328513490917311e-09,
      "logits/chosen": 136.25003051757812,
      "logits/rejected": 136.1801300048828,
      "logps/chosen": -824.2185668945312,
      "logps/rejected": -853.8348388671875,
      "loss": 0.985,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": 2.142530918121338,
      "rewards/margins": 3.103809356689453,
      "rewards/rejected": -0.9612787365913391,
      "step": 50
    },
    {
      "epoch": 0.99,
      "step": 52,
      "total_flos": 0.0,
      "train_loss": 1.1432483746455266,
      "train_runtime": 572.1827,
      "train_samples_per_second": 11.797,
      "train_steps_per_second": 0.091
    }
  ],
  "logging_steps": 10,
  "max_steps": 52,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}