transformer_en_de10 / train_results.json
Everlyn's picture
model
d95a1fe
raw
history blame contribute delete
197 Bytes
{
"epoch": 4.0,
"train_loss": 2.836493848511954,
"train_runtime": 50326.1237,
"train_samples": 3869033,
"train_samples_per_second": 307.517,
"train_steps_per_second": 0.32
}