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": 26.563100152564367,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": 123.11854553222656,
      "logits/rejected": 97.00198364257812,
      "logps/chosen": -425.18585205078125,
      "logps/rejected": -424.1869201660156,
      "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": 27.293107117678087,
      "learning_rate": 4.907293218369498e-07,
      "logits/chosen": 117.04379272460938,
      "logits/rejected": 135.89993286132812,
      "logps/chosen": -441.55804443359375,
      "logps/rejected": -523.9049682617188,
      "loss": 0.6893,
      "rewards/accuracies": 0.5347222089767456,
      "rewards/chosen": 0.01830328069627285,
      "rewards/margins": 0.016203498467803,
      "rewards/rejected": 0.002099784789606929,
      "step": 10
    },
    {
      "epoch": 0.38,
      "grad_norm": 41.6541290719842,
      "learning_rate": 3.941700805287168e-07,
      "logits/chosen": 119.21928405761719,
      "logits/rejected": 126.63338470458984,
      "logps/chosen": -430.0658264160156,
      "logps/rejected": -516.3239135742188,
      "loss": 0.6084,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": -0.03801226615905762,
      "rewards/margins": 0.25012144446372986,
      "rewards/rejected": -0.2881337106227875,
      "step": 20
    },
    {
      "epoch": 0.57,
      "grad_norm": 30.808060184454277,
      "learning_rate": 2.3293939665883228e-07,
      "logits/chosen": 109.26795959472656,
      "logits/rejected": 112.9657211303711,
      "logps/chosen": -496.3540954589844,
      "logps/rejected": -602.0138549804688,
      "loss": 0.5794,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.41233962774276733,
      "rewards/margins": 0.5356513857841492,
      "rewards/rejected": -0.9479910731315613,
      "step": 30
    },
    {
      "epoch": 0.76,
      "grad_norm": 30.141976904526587,
      "learning_rate": 7.936171419533652e-08,
      "logits/chosen": 107.7169418334961,
      "logits/rejected": 102.60658264160156,
      "logps/chosen": -524.6154174804688,
      "logps/rejected": -586.1959838867188,
      "loss": 0.5711,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.5143404603004456,
      "rewards/margins": 0.4999730587005615,
      "rewards/rejected": -1.0143134593963623,
      "step": 40
    },
    {
      "epoch": 0.95,
      "grad_norm": 26.331594229316682,
      "learning_rate": 2.328513490917311e-09,
      "logits/chosen": 116.38276672363281,
      "logits/rejected": 116.3095703125,
      "logps/chosen": -542.1165771484375,
      "logps/rejected": -607.3255615234375,
      "loss": 0.5531,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.5929387807846069,
      "rewards/margins": 0.5678332448005676,
      "rewards/rejected": -1.1607720851898193,
      "step": 50
    },
    {
      "epoch": 0.99,
      "step": 52,
      "total_flos": 0.0,
      "train_loss": 0.6015224342162793,
      "train_runtime": 574.0149,
      "train_samples_per_second": 11.759,
      "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
}