vanilla_doff_dpo_iter_2 / train_results.json
YYYYYYibo's picture
Model save
4325609 verified
{
"epoch": 1.0,
"train_loss": 0.6705375785141988,
"train_runtime": 6538.0645,
"train_samples": 19600,
"train_samples_per_second": 2.998,
"train_steps_per_second": 0.023
}