prm_version2_subsample_hf / eval_results.json
DongfuJiang's picture
End of training
86a5737 verified
raw
history blame contribute delete
177 Bytes
{
"epoch": 0.9998370964063468,
"eval_loss": 0.13488060235977173,
"eval_runtime": 55.2744,
"eval_samples_per_second": 44.885,
"eval_steps_per_second": 5.626
}