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": 7.772319002734789,
      "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": 7.445029419543799,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.755617618560791,
      "logits/rejected": -2.7461938858032227,
      "logps/chosen": -271.96563720703125,
      "logps/rejected": -260.71453857421875,
      "loss": 0.6914,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.005042961798608303,
      "rewards/margins": 0.0035791893023997545,
      "rewards/rejected": 0.0014637727290391922,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.722931161787328,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.7870638370513916,
      "logits/rejected": -2.7682018280029297,
      "logps/chosen": -261.4487609863281,
      "logps/rejected": -250.8934783935547,
      "loss": 0.6745,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": 0.03232940658926964,
      "rewards/margins": 0.04083934798836708,
      "rewards/rejected": -0.008509937673807144,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.3904115438921405,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.788241386413574,
      "logits/rejected": -2.769221544265747,
      "logps/chosen": -293.80859375,
      "logps/rejected": -254.22073364257812,
      "loss": 0.6486,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.02352513000369072,
      "rewards/margins": 0.12710139155387878,
      "rewards/rejected": -0.150626540184021,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.14675940130506,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.761005401611328,
      "logits/rejected": -2.738503932952881,
      "logps/chosen": -264.15814208984375,
      "logps/rejected": -255.6121826171875,
      "loss": 0.6354,
      "rewards/accuracies": 0.690625011920929,
      "rewards/chosen": -0.11238692700862885,
      "rewards/margins": 0.16062822937965393,
      "rewards/rejected": -0.2730151116847992,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.52526513305357,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7708582878112793,
      "logits/rejected": -2.750849962234497,
      "logps/chosen": -279.1473083496094,
      "logps/rejected": -287.328857421875,
      "loss": 0.6201,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.22138634324073792,
      "rewards/margins": 0.13677750527858734,
      "rewards/rejected": -0.35816383361816406,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.649035688173973,
      "train_runtime": 1642.5856,
      "train_samples_per_second": 9.304,
      "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
}