djbp's picture
End of training
3531c26 verified
raw
history blame
223 Bytes
{
"epoch": 6.885245901639344,
"total_flos": 4.1785312376666235e+18,
"train_loss": 0.38806929134187246,
"train_runtime": 12116.6555,
"train_samples_per_second": 4.47,
"train_steps_per_second": 0.009
}