djbp's picture
End of training
1129da7 verified
raw
history blame contribute delete
222 Bytes
{
"epoch": 9.836065573770492,
"total_flos": 1.9063087726729052e+18,
"train_loss": 0.3150376001993815,
"train_runtime": 8083.4568,
"train_samples_per_second": 9.644,
"train_steps_per_second": 0.037
}