Orra 0.3.1

Faster cleanup, and numbers in digits when you have a vocabulary.

• With Clean Up Text on, a dictation with no filler and no spoken correction is pasted at once, without waiting for the model.
• When the model does run, it copies most of its output from your own words, so long texts are cleaned up about three times faster. The result is the same as before.
• With words in your vocabulary, numbers spoken in Chinese, such as years, amounts and percentages, are written as digits again, as they are without a vocabulary. Your vocabulary words are still written your way. With Write numbers as digits off, nothing changes.

整理更快了，有个人词库时数字也会写成阿拉伯数字。

• 打开“整理文字”后，没有语气词、也没有口头更正的听写会立即粘贴，不再等待模型。
• 需要整理时，模型大部分输出直接照抄你说的原话，所以长文字的整理速度快了大约三倍，结果和以前一样。
• 个人词库里有词时，用中文说的数字，比如年份、金额和百分比，会和没有词库时一样写成阿拉伯数字，词库里的词仍按你的写法写。关闭“把数字写成阿拉伯数字”时，一切照旧。