zkdeng's picture
End of training
38f98de
raw
history blame contribute delete
210 Bytes
{
"epoch": 3.99,
"total_flos": 1.8332979581807493e+18,
"train_loss": 1.1345372907485736,
"train_runtime": 1112.1938,
"train_samples_per_second": 77.684,
"train_steps_per_second": 1.212
}