checkpoint-100/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5191434133679429,
  "eval_steps": 25,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12978585334198572,
      "grad_norm": 0.0001288286439375952,
      "learning_rate": 9.985564567238237e-05,
      "logits/chosen": -2.892515182495117,
      "logits/rejected": -2.8839669227600098,
      "logps/chosen": -65.34242248535156,
      "logps/rejected": -120.23748779296875,
      "loss": 0.2047,
      "rewards/accuracies": 0.7599999904632568,
      "rewards/chosen": -2.3054304122924805,
      "rewards/margins": 7.21456241607666,
      "rewards/rejected": -9.519991874694824,
      "step": 25
    },
    {
      "epoch": 0.12978585334198572,
      "eval_logits/chosen": -3.1952435970306396,
      "eval_logits/rejected": -3.1321091651916504,
      "eval_logps/chosen": -126.43878936767578,
      "eval_logps/rejected": -265.6844177246094,
      "eval_loss": 4.6241984819062054e-07,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -8.37295913696289,
      "eval_rewards/margins": 15.654376983642578,
      "eval_rewards/rejected": -24.02733612060547,
      "eval_runtime": 140.1846,
      "eval_samples_per_second": 0.585,
      "eval_steps_per_second": 0.585,
      "step": 25
    },
    {
      "epoch": 0.25957170668397145,
      "grad_norm": 1.2395206795190461e-05,
      "learning_rate": 9.795296239506012e-05,
      "logits/chosen": -2.2458553314208984,
      "logits/rejected": -2.1915533542633057,
      "logps/chosen": -99.26245880126953,
      "logps/rejected": -254.2440643310547,
      "loss": 0.0,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -5.69963264465332,
      "rewards/margins": 17.1381778717041,
      "rewards/rejected": -22.837810516357422,
      "step": 50
    },
    {
      "epoch": 0.25957170668397145,
      "eval_logits/chosen": -1.7933237552642822,
      "eval_logits/rejected": -1.7323336601257324,
      "eval_logps/chosen": -89.24785614013672,
      "eval_logps/rejected": -249.6978302001953,
      "eval_loss": 4.8683162390261714e-08,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -4.653865337371826,
      "eval_rewards/margins": 17.77481460571289,
      "eval_rewards/rejected": -22.428680419921875,
      "eval_runtime": 140.0068,
      "eval_samples_per_second": 0.586,
      "eval_steps_per_second": 0.586,
      "step": 50
    },
    {
      "epoch": 0.3893575600259572,
      "grad_norm": 3.164160079904832e-05,
      "learning_rate": 9.392069159975199e-05,
      "logits/chosen": -1.7290905714035034,
      "logits/rejected": -1.6691365242004395,
      "logps/chosen": -87.64579010009766,
      "logps/rejected": -252.55418395996094,
      "loss": 0.0,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -4.562257766723633,
      "rewards/margins": 18.078048706054688,
      "rewards/rejected": -22.640310287475586,
      "step": 75
    },
    {
      "epoch": 0.3893575600259572,
      "eval_logits/chosen": -1.7083513736724854,
      "eval_logits/rejected": -1.6462993621826172,
      "eval_logps/chosen": -86.74995422363281,
      "eval_logps/rejected": -249.47621154785156,
      "eval_loss": 3.741008924862399e-08,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -4.404075622558594,
      "eval_rewards/margins": 18.002437591552734,
      "eval_rewards/rejected": -22.406513214111328,
      "eval_runtime": 140.046,
      "eval_samples_per_second": 0.586,
      "eval_steps_per_second": 0.586,
      "step": 75
    },
    {
      "epoch": 0.5191434133679429,
      "grad_norm": 9.557259545545094e-06,
      "learning_rate": 8.793790613463955e-05,
      "logits/chosen": -1.706992506980896,
      "logits/rejected": -1.6317412853240967,
      "logps/chosen": -85.46414184570312,
      "logps/rejected": -248.94418334960938,
      "loss": 0.0,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -4.233969211578369,
      "rewards/margins": 18.118690490722656,
      "rewards/rejected": -22.3526611328125,
      "step": 100
    },
    {
      "epoch": 0.5191434133679429,
      "eval_logits/chosen": -1.701735019683838,
      "eval_logits/rejected": -1.6400282382965088,
      "eval_logps/chosen": -86.22808837890625,
      "eval_logps/rejected": -250.09986877441406,
      "eval_loss": 3.298516659810957e-08,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -4.3518877029418945,
      "eval_rewards/margins": 18.11699104309082,
      "eval_rewards/rejected": -22.468881607055664,
      "eval_runtime": 140.2981,
      "eval_samples_per_second": 0.584,
      "eval_steps_per_second": 0.584,
      "step": 100
    }
  ],
  "logging_steps": 25,
  "max_steps": 384,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}