nash_dpo_merge_iter_real_plus_3 / train_results.json
YYYYYYibo's picture
Model save
2cb6382 verified
raw
history blame contribute delete
No virus
192 Bytes
{
"epoch": 1.0,
"train_loss": 0.665724610671019,
"train_runtime": 1696.1298,
"train_samples": 5000,
"train_samples_per_second": 2.948,
"train_steps_per_second": 0.023
}