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": 7.771931028477214,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.807276487350464,
      "logits/rejected": -2.7759768962860107,
      "logps/chosen": -315.42626953125,
      "logps/rejected": -227.5915985107422,
      "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": 7.451444947473661,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.755561113357544,
      "logits/rejected": -2.7461612224578857,
      "logps/chosen": -271.9822998046875,
      "logps/rejected": -260.6978759765625,
      "loss": 0.6915,
      "rewards/accuracies": 0.5069444179534912,
      "rewards/chosen": 0.004876165185123682,
      "rewards/margins": 0.003245496889576316,
      "rewards/rejected": 0.0016306680627167225,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.715407871729835,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.7868247032165527,
      "logits/rejected": -2.7679924964904785,
      "logps/chosen": -261.4691467285156,
      "logps/rejected": -250.8953094482422,
      "loss": 0.6746,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.0321255698800087,
      "rewards/margins": 0.04065405949950218,
      "rewards/rejected": -0.008528487756848335,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.412395635719943,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7875523567199707,
      "logits/rejected": -2.7685694694519043,
      "logps/chosen": -293.82598876953125,
      "logps/rejected": -254.20108032226562,
      "loss": 0.6487,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.02369900979101658,
      "rewards/margins": 0.12673112750053406,
      "rewards/rejected": -0.15043014287948608,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.170584817408116,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.7602181434631348,
      "logits/rejected": -2.7377209663391113,
      "logps/chosen": -264.1241149902344,
      "logps/rejected": -255.56826782226562,
      "loss": 0.6353,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.11204648017883301,
      "rewards/margins": 0.1605294644832611,
      "rewards/rejected": -0.27257591485977173,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.574236292000831,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7699923515319824,
      "logits/rejected": -2.7500431537628174,
      "logps/chosen": -279.1482238769531,
      "logps/rejected": -287.313720703125,
      "loss": 0.6202,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.22139573097229004,
      "rewards/margins": 0.13661642372608185,
      "rewards/rejected": -0.3580121695995331,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6490718308141676,
      "train_runtime": 1645.2518,
      "train_samples_per_second": 9.289,
      "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
}