distilbert_rand_5_v2_rte / train_results.json
Hartunka's picture
End of training
df71dea verified
{
"epoch": 7.0,
"total_flos": 1154453379287040.0,
"train_loss": 0.5091195225715637,
"train_runtime": 26.6073,
"train_samples": 2490,
"train_samples_per_second": 4679.16,
"train_steps_per_second": 18.792
}