djbp's picture
End of training
f8702d3 verified
raw
history blame
208 Bytes
{
"epoch": 7.0,
"total_flos": 5.277497331917003e+18,
"train_loss": 0.41070275199144407,
"train_runtime": 6937.9689,
"train_samples_per_second": 9.709,
"train_steps_per_second": 0.019
}