distilled-mt5-small-0.2-0.25 / all_results.json
Lvxue's picture
End of training
809983d
raw
history blame
No virus
411 Bytes
{
"epoch": 5.0,
"eval_bleu": 5.3011,
"eval_gen_len": 42.6653,
"eval_loss": 3.161741256713867,
"eval_runtime": 608.4882,
"eval_samples": 1999,
"eval_samples_per_second": 3.285,
"eval_steps_per_second": 0.411,
"train_loss": 7.9021027734375,
"train_runtime": 2649.083,
"train_samples": 10000,
"train_samples_per_second": 18.874,
"train_steps_per_second": 4.719
}