all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": -2.612835168838501,
    "eval_logits/rejected": -2.5376927852630615,
    "eval_logps/chosen": -343.6803283691406,
    "eval_logps/rejected": -345.713134765625,
    "eval_loss": 0.6980592012405396,
    "eval_rewards/accuracies": 0.79296875,
    "eval_rewards/chosen": -5.03585958480835,
    "eval_rewards/margins": 3.8046231269836426,
    "eval_rewards/rejected": -8.840482711791992,
    "eval_runtime": 317.955,
    "eval_samples": 2000,
    "eval_samples_per_second": 6.29,
    "eval_steps_per_second": 0.101,
    "train_loss": 0.20528750838680607,
    "train_runtime": 83798.7665,
    "train_samples": 61966,
    "train_samples_per_second": 2.218,
    "train_steps_per_second": 0.069
}