1SV64 / train_results.json
gotzmann's picture
..
f822829
{
"epoch": 1.0,
"total_flos": 2.81863078907845e+18,
"train_loss": 1.6374119961558387,
"train_runtime": 72812.9252,
"train_samples_per_second": 0.402,
"train_steps_per_second": 0.05
}