Could you please check the BLEU error metric again? The scores that the system produce are very low. When the same solution is tested localy, the BLEU score on a held-out dataset defined as a subset of the training dataset you provided to us is much higher than what we get here.
'%3e%3cpath fill='%23001489' d='M0 0v6h9V0z'/%3e%3cpath fill='%23e03c31' d='M0 0v3h9V0z'/%3e%3cg stroke='%23fff' stroke-width='2'%3e%3cpath id='d' d='m0 0 4.5 3L0 6m4.5-3H9'/%3e%3cuse xlink:href='%23b' stroke='%23ffb81c' clip-path='url(%23c)'/%3e%3c/g%3e%3cuse xlink:href='%23d' fill='none' stroke='%23007749' stroke-width='1.2'/%3e%3c/g%3e%3c/svg%3e)
Could you please check the BLEU error metric again? The scores that the system produce are very low. When the same solution is tested localy, the BLEU score on a held-out dataset defined as a subset of the training dataset you provided to us is much higher than what we get here.
Is it possible to request a file from personal submission? Seems like I have one file from the personal submission that might have a better score