tangken333's picture
End of training
5113f05 verified
{
"epoch": 3.0,
"total_flos": 3.952374498459648e+16,
"train_loss": 0.17466499442341685,
"train_runtime": 4967.3247,
"train_samples_per_second": 0.84,
"train_steps_per_second": 0.018
}