all_results.json

{
    "epoch": 0.9221902017291066,
    "eval_logits/chosen": -0.025785785168409348,
    "eval_logits/rejected": 0.1337740421295166,
    "eval_logps/chosen": -129.20240783691406,
    "eval_logps/ref_chosen": -127.95065307617188,
    "eval_logps/ref_rejected": -164.48851013183594,
    "eval_logps/rejected": -165.79054260253906,
    "eval_loss": 0.5649449825286865,
    "eval_rewards/accuracies": 0.520370364189148,
    "eval_rewards/chosen": -0.012517303228378296,
    "eval_rewards/margins": 0.0005030279280617833,
    "eval_rewards/rejected": -0.013020330108702183,
    "eval_runtime": 144.4308,
    "eval_samples": 1080,
    "eval_samples_per_second": 7.478,
    "eval_sft_loss": 0.4369891583919525,
    "eval_steps_per_second": 1.869,
    "total_flos": 0.0,
    "train_loss": 0.5759013295173645,
    "train_runtime": 619.8336,
    "train_samples": 1387,
    "train_samples_per_second": 2.238,
    "train_steps_per_second": 0.016
}