trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6019110676397562,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 8.000000000000001e-06,
      "logits/chosen": -58.747344970703125,
      "logits/rejected": -59.84019470214844,
      "logps/chosen": -150.3143768310547,
      "logps/rejected": -179.38966369628906,
      "loss": 0.5314,
      "rewards/accuracies": 0.7318750023841858,
      "rewards/chosen": -0.8851571679115295,
      "rewards/margins": 0.6559739708900452,
      "rewards/rejected": -1.5411310195922852,
      "step": 200
    },
    {
      "epoch": 0.24,
      "learning_rate": 9.68421052631579e-06,
      "logits/chosen": -64.05355072021484,
      "logits/rejected": -65.13407897949219,
      "logps/chosen": -155.71795654296875,
      "logps/rejected": -190.02166748046875,
      "loss": 0.4485,
      "rewards/accuracies": 0.7715625166893005,
      "rewards/chosen": -1.3824467658996582,
      "rewards/margins": 1.1924123764038086,
      "rewards/rejected": -2.5748589038848877,
      "step": 400
    },
    {
      "epoch": 0.36,
      "learning_rate": 9.263157894736842e-06,
      "logits/chosen": -64.59612274169922,
      "logits/rejected": -65.59567260742188,
      "logps/chosen": -158.51266479492188,
      "logps/rejected": -191.44497680664062,
      "loss": 0.4208,
      "rewards/accuracies": 0.7871875166893005,
      "rewards/chosen": -1.4413859844207764,
      "rewards/margins": 1.3812304735183716,
      "rewards/rejected": -2.8226163387298584,
      "step": 600
    },
    {
      "epoch": 0.48,
      "learning_rate": 8.842105263157895e-06,
      "logits/chosen": -65.70255279541016,
      "logits/rejected": -66.5833969116211,
      "logps/chosen": -156.6477508544922,
      "logps/rejected": -195.2394561767578,
      "loss": 0.4062,
      "rewards/accuracies": 0.7973437309265137,
      "rewards/chosen": -1.4948838949203491,
      "rewards/margins": 1.513500690460205,
      "rewards/rejected": -3.0083847045898438,
      "step": 800
    },
    {
      "epoch": 0.6,
      "learning_rate": 8.421052631578948e-06,
      "logits/chosen": -65.27395629882812,
      "logits/rejected": -66.06521606445312,
      "logps/chosen": -156.45945739746094,
      "logps/rejected": -196.63790893554688,
      "loss": 0.3898,
      "rewards/accuracies": 0.7978125214576721,
      "rewards/chosen": -1.6718982458114624,
      "rewards/margins": 1.6537814140319824,
      "rewards/rejected": -3.3256795406341553,
      "step": 1000
    }
  ],
  "logging_steps": 200,
  "max_steps": 5000,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 0.0,
  "trial_name": null,
  "trial_params": null
}