train_sst2_101112_1760638077 / train_results.json
rbelanec's picture
End of training
d2001a6 verified
raw
history blame contribute delete
252 Bytes
{
"epoch": 20.0,
"num_input_tokens_seen": 67752208,
"total_flos": 3.0568189449075425e+18,
"train_loss": 0.020170351656610876,
"train_runtime": 56363.7675,
"train_samples_per_second": 21.508,
"train_steps_per_second": 5.377
}