Model: jackf857/llama-3-8b-base-ipo-ultrafeedback-4xh200-batch-128-rerun-2-runpod Source: Original Platform
16 lines
591 B
JSON
16 lines
591 B
JSON
{
|
|
"epoch": 0.9989528795811519,
|
|
"eval_logits/chosen": -0.40516504645347595,
|
|
"eval_logits/rejected": -0.40924227237701416,
|
|
"eval_logps/chosen": -4.689553737640381,
|
|
"eval_logps/rejected": -7.291287422180176,
|
|
"eval_loss": 2328.408447265625,
|
|
"eval_rewards/accuracies": 0.6840000152587891,
|
|
"eval_rewards/chosen": -0.03572991117835045,
|
|
"eval_rewards/margins": 0.024208687245845795,
|
|
"eval_rewards/rejected": -0.05993859842419624,
|
|
"eval_runtime": 78.645,
|
|
"eval_samples": 2000,
|
|
"eval_samples_per_second": 25.431,
|
|
"eval_steps_per_second": 1.589
|
|
} |