alberti-bert-base-multilingual-cased / evals /eval_results_mbert.json
versae's picture
Adding evaluations on val set
c081a91
raw
history blame
122 Bytes
{
"eval_accuracy": 0.2531256913599916,
"eval_loss": 4.986973319130462,
"eval_perplexity": 146.49236618254264
}