train_stsb_101112_1760638041 / train_results.json
rbelanec's picture
End of training
9525b57 verified
{
"epoch": 20.0,
"num_input_tokens_seen": 8712528,
"total_flos": 3.9233147577237504e+17,
"train_loss": 0.6318969237610951,
"train_runtime": 3960.2188,
"train_samples_per_second": 26.13,
"train_steps_per_second": 6.535
}