trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9874476987447699,
  "eval_steps": 500,
  "global_step": 59,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016736401673640166,
      "grad_norm": 38.17963514831367,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.289388656616211,
      "logits/rejected": -2.266869068145752,
      "logps/chosen": -490.1561584472656,
      "logps/rejected": -577.6099853515625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16736401673640167,
      "grad_norm": 39.09010964537688,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.251622438430786,
      "logits/rejected": -2.1955909729003906,
      "logps/chosen": -519.5778198242188,
      "logps/rejected": -626.5214233398438,
      "loss": 0.6765,
      "rewards/accuracies": 0.5208333134651184,
      "rewards/chosen": -0.12538284063339233,
      "rewards/margins": 0.08053015172481537,
      "rewards/rejected": -0.2059129923582077,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 37.575241185625224,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.1704227924346924,
      "logits/rejected": -2.113105297088623,
      "logps/chosen": -574.7396240234375,
      "logps/rejected": -690.5613403320312,
      "loss": 0.6614,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.6970238089561462,
      "rewards/margins": 0.36311453580856323,
      "rewards/rejected": -1.0601383447647095,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 40.16374886117498,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.12542462348938,
      "logits/rejected": -2.090336561203003,
      "logps/chosen": -564.1171264648438,
      "logps/rejected": -661.2869873046875,
      "loss": 0.6174,
      "rewards/accuracies": 0.640625,
      "rewards/chosen": -0.5616307854652405,
      "rewards/margins": 0.25331956148147583,
      "rewards/rejected": -0.8149503469467163,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 53.94348043999881,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.1451804637908936,
      "logits/rejected": -2.0768136978149414,
      "logps/chosen": -549.3674926757812,
      "logps/rejected": -696.8201293945312,
      "loss": 0.5933,
      "rewards/accuracies": 0.703125,
      "rewards/chosen": -0.43148988485336304,
      "rewards/margins": 0.4352260231971741,
      "rewards/rejected": -0.8667157888412476,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 48.7316594517855,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.0674972534179688,
      "logits/rejected": -2.0488924980163574,
      "logps/chosen": -601.2271118164062,
      "logps/rejected": -747.360595703125,
      "loss": 0.5873,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.7331333756446838,
      "rewards/margins": 0.4694816470146179,
      "rewards/rejected": -1.2026150226593018,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6273458044407731,
      "train_runtime": 1643.7461,
      "train_samples_per_second": 9.298,
      "train_steps_per_second": 0.036
    }
  ],
  "logging_steps": 10,
  "max_steps": 59,
  "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
}