trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9915966386554622,
  "eval_steps": 100,
  "global_step": 59,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 6.891863950600918,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -1.1214768886566162,
      "logits/rejected": -1.0666239261627197,
      "logps/chosen": -773.5914306640625,
      "logps/rejected": -765.6082763671875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.17,
      "grad_norm": 6.6410695205888075,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -1.2806458473205566,
      "logits/rejected": -0.9247087836265564,
      "logps/chosen": -503.0013122558594,
      "logps/rejected": -899.8956298828125,
      "loss": 0.6907,
      "rewards/accuracies": 0.5069444179534912,
      "rewards/chosen": -0.0002252649428555742,
      "rewards/margins": 0.004411212634295225,
      "rewards/rejected": -0.004636477679014206,
      "step": 10
    },
    {
      "epoch": 0.34,
      "grad_norm": 7.3762614729181015,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -1.2530821561813354,
      "logits/rejected": -0.9845311045646667,
      "logps/chosen": -543.0537109375,
      "logps/rejected": -933.8472900390625,
      "loss": 0.6527,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": -0.019722815603017807,
      "rewards/margins": 0.08204329013824463,
      "rewards/rejected": -0.10176609456539154,
      "step": 20
    },
    {
      "epoch": 0.5,
      "grad_norm": 8.929333715687283,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -1.1901506185531616,
      "logits/rejected": -0.9701334238052368,
      "logps/chosen": -547.3389892578125,
      "logps/rejected": -972.1591796875,
      "loss": 0.5279,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.08818452060222626,
      "rewards/margins": 0.44495606422424316,
      "rewards/rejected": -0.5331405997276306,
      "step": 30
    },
    {
      "epoch": 0.67,
      "grad_norm": 7.370675587057317,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -1.121147632598877,
      "logits/rejected": -0.9864951372146606,
      "logps/chosen": -544.5201416015625,
      "logps/rejected": -1050.002197265625,
      "loss": 0.4145,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": -0.19094732403755188,
      "rewards/margins": 0.9460043907165527,
      "rewards/rejected": -1.1369515657424927,
      "step": 40
    },
    {
      "epoch": 0.84,
      "grad_norm": 6.641314363386866,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -1.0754799842834473,
      "logits/rejected": -1.0215175151824951,
      "logps/chosen": -568.5784912109375,
      "logps/rejected": -1094.7982177734375,
      "loss": 0.3982,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.3303142786026001,
      "rewards/margins": 1.3637864589691162,
      "rewards/rejected": -1.6941007375717163,
      "step": 50
    },
    {
      "epoch": 0.99,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5082056158680027,
      "train_runtime": 904.9767,
      "train_samples_per_second": 4.188,
      "train_steps_per_second": 0.065
    }
  ],
  "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
}