sdgroeve's picture
End of training
545747d
raw
history blame
210 Bytes
{
"epoch": 2.99,
"total_flos": 3.3497451642252165e+18,
"train_loss": 0.5413832122217669,
"train_runtime": 1364.9822,
"train_samples_per_second": 98.902,
"train_steps_per_second": 0.771
}