deepseek-8b-dpo-full2 / all_results.json
zfz1's picture
Model save
4ba1d38 verified
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.26106402876377105,
"train_runtime": 17697.4864,
"train_samples": 20000,
"train_samples_per_second": 1.13,
"train_steps_per_second": 0.141
}