MuggleMath_70B / all_results.json
路卡
init
2e554e9
raw
history blame contribute delete
168 Bytes
{
"epoch": 3.0,
"train_loss": 0.28055200490389104,
"train_runtime": 252630.326,
"train_samples_per_second": 5.482,
"train_steps_per_second": 0.043
}