WeTextProcessing
WeTextProcessing copied to clipboard
关于中英文TN的速度问题
都是20个字/20个单词左右的长度 在12700K的处理器上 中文tn大概6-10ms 但是英文tn大概要45-50ms 速度差了将近10倍,请问下速度上还有提升空间吗?
开始我以为是whitelist的原因,但是把whitelist去掉后速度没有明显变化.....
btw, 提个小的base case:
res = normalizer.normalize(
"My No. is 0,013,225, and your balance is still $21341.34, which is due in 2 months and 14 days")
里面的
My number is oh zero one threetwo two five, and your balance is still twenty one thousand three hundred and forty one point three four dollars, which is due in two months and fourteen days
three和two连在一起了...