nash_dpo_doff_no_golden_iter_3 / all_results.json
YYYYYYibo's picture
Model save
8d6f77f verified
{
"epoch": 1.0,
"train_loss": 0.5280830058050744,
"train_runtime": 8024.3189,
"train_samples": 20735,
"train_samples_per_second": 2.584,
"train_steps_per_second": 0.02
}