llama2_7b_instruct_sft_dpo / train_results.json
yuansui's picture
Model save
f05c27d verified
raw
history blame
216 Bytes
{
"epoch": 7.272,
"total_flos": 0.0,
"train_loss": 0.5793062054685184,
"train_runtime": 6908.77,
"train_samples": 1000,
"train_samples_per_second": 1.158,
"train_steps_per_second": 0.008
}