all_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -3.900073289871216,
    "eval_logits/rejected": -3.9142141342163086,
    "eval_logps/chosen": -293.32354736328125,
    "eval_logps/rejected": -187.59616088867188,
    "eval_loss": 0.6149587035179138,
    "eval_rewards/accuracies": 0.6875,
    "eval_rewards/chosen": -0.1575067937374115,
    "eval_rewards/diff": -2.2334296703338623,
    "eval_rewards/diff_abs": 2.236445188522339,
    "eval_rewards/rejected": -0.33032703399658203,
    "eval_rewards/student_margin": 0.17282025516033173,
    "eval_rewards/teacher_margin": 2.40625,
    "eval_runtime": 11.4547,
    "eval_samples": 1470,
    "eval_samples_per_second": 128.332,
    "eval_steps_per_second": 0.175,
    "train_loss": 0.6488963676806219,
    "train_runtime": 2900.3403,
    "train_samples": 147002,
    "train_samples_per_second": 50.684,
    "train_steps_per_second": 0.132
}