ALMA-13B-Pretrain-Cy-1 / train_results.json
a-r-j's picture
initial upload
760ce79
{
"epoch": 3.0,
"train_loss": 0.9383809857774829,
"train_runtime": 151857.9043,
"train_samples": 210373,
"train_samples_per_second": 4.156,
"train_steps_per_second": 0.03
}