tangken333's picture
End of training
1cb1647 verified
{
"epoch": 2.909090909090909,
"total_flos": 3.5981250321711104e+16,
"train_loss": 0.4774101939466264,
"train_runtime": 9083.6349,
"train_samples_per_second": 0.392,
"train_steps_per_second": 0.008
}