TheBguy87's picture
first GPT2 model with ~2M words
5b9aa08
{
"epoch": 3.0,
"train_loss": 5.383951053103885,
"train_runtime": 1331.1447,
"train_samples": 2962,
"train_samples_per_second": 6.675,
"train_steps_per_second": 1.668
}