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": 7.7730547112344635,
      "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": 7.459164668463289,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.7555456161499023,
      "logits/rejected": -2.7461628913879395,
      "logps/chosen": -271.9652099609375,
      "logps/rejected": -260.6877136230469,
      "loss": 0.6915,
      "rewards/accuracies": 0.4965277910232544,
      "rewards/chosen": 0.005047248676419258,
      "rewards/margins": 0.0033153239637613297,
      "rewards/rejected": 0.0017319255275651813,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.712012764608614,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.787020206451416,
      "logits/rejected": -2.768185615539551,
      "logps/chosen": -261.4574890136719,
      "logps/rejected": -250.9368896484375,
      "loss": 0.6745,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.0322418287396431,
      "rewards/margins": 0.041185710579156876,
      "rewards/rejected": -0.008943881839513779,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.374097934015646,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7877423763275146,
      "logits/rejected": -2.7687172889709473,
      "logps/chosen": -293.801025390625,
      "logps/rejected": -254.2049102783203,
      "loss": 0.6486,
      "rewards/accuracies": 0.6781250238418579,
      "rewards/chosen": -0.023449189960956573,
      "rewards/margins": 0.12701918184757233,
      "rewards/rejected": -0.1504683643579483,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.143971290047324,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.760338544845581,
      "logits/rejected": -2.737778663635254,
      "logps/chosen": -264.13775634765625,
      "logps/rejected": -255.5776824951172,
      "loss": 0.6353,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.11218307912349701,
      "rewards/margins": 0.1604871302843094,
      "rewards/rejected": -0.2726702094078064,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.540975248834071,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7702414989471436,
      "logits/rejected": -2.750225305557251,
      "logps/chosen": -279.04254150390625,
      "logps/rejected": -287.20135498046875,
      "loss": 0.6202,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.22033901512622833,
      "rewards/margins": 0.13654986023902893,
      "rewards/rejected": -0.35688892006874084,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6490067789110087,
      "train_runtime": 1651.119,
      "train_samples_per_second": 9.256,
      "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
}