ruDialoGPT-medium / train_results.json
d.tsimerman
squash
cc2664b
{
"epoch": 1.0,
"train_loss": 3.4927627832876236,
"train_runtime": 361845.2402,
"train_samples": 8570756,
"train_samples_per_second": 23.686,
"train_steps_per_second": 1.974
}