zkdeng's picture
End of training
a325aa2
{
"epoch": 4.0,
"total_flos": 6.8284774551813e+19,
"train_loss": 1.060067500640039,
"train_runtime": 12677.4822,
"train_samples_per_second": 67.943,
"train_steps_per_second": 1.061
}