all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": -13.25,
    "eval_logits/rejected": -12.9375,
    "eval_logps/chosen": -700.0,
    "eval_logps/rejected": -692.0,
    "eval_loss": 1.0602344274520874,
    "eval_rewards/accuracies": 0.5,
    "eval_rewards/chosen": -3.796875,
    "eval_rewards/margins": 0.21484375,
    "eval_rewards/rejected": -4.0,
    "eval_runtime": 46.548,
    "eval_samples": 2000,
    "eval_samples_per_second": 42.966,
    "eval_steps_per_second": 0.687,
    "total_flos": 0.0,
    "train_loss": 0.06547752297897848,
    "train_runtime": 12417.9968,
    "train_samples": 61118,
    "train_samples_per_second": 14.765,
    "train_steps_per_second": 0.231
}