mistral_gsm8k_dpo_0 / train_results.json
weijie210's picture
Model save
a49977d verified
{
"epoch": 2.0,
"train_loss": 0.06552632141962543,
"train_runtime": 3687.7187,
"train_samples": 4570,
"train_samples_per_second": 2.478,
"train_steps_per_second": 0.078
}