T5Lae-Large-WeightedLoss / train_results.json
hrezaei's picture
End of training
bd4356c
{
"total_flos": 4.621498060193661e+18,
"train_loss": 0.2757485848851502,
"train_runtime": 24028.7714,
"train_samples": 2000000,
"train_samples_per_second": 87.277,
"train_steps_per_second": 21.819
}