trivia_albert_xxl_finetuned / train_results.json
shuheng's picture
End of training
6cb1913 verified
{
"epoch": 6.0,
"total_flos": 3.780810569313792e+16,
"train_loss": 0.7023615477220068,
"train_runtime": 12579.5465,
"train_samples": 13545,
"train_samples_per_second": 6.46,
"train_steps_per_second": 0.202
}