qwen2-0.5b-hh-sft / all_results.json
yangzhao02's picture
End of training
b3a2f58 verified
raw
history blame contribute delete
388 Bytes
{
"epoch": 1.0,
"eval_loss": 2.4862771034240723,
"eval_runtime": 0.4251,
"eval_samples": 100,
"eval_samples_per_second": 32.932,
"eval_steps_per_second": 2.352,
"total_flos": 27695775744.0,
"train_loss": 1.5830808877944946,
"train_runtime": 12.0584,
"train_samples": 500,
"train_samples_per_second": 6.551,
"train_steps_per_second": 0.083
}