le.chang
le.chang
> `docker run -it -p 9000:80 \onlyoffice/documentserver` without any mounts, check that it worked and start adding mount one by one to find out which is causing this trouble hi...
hi @ShockwaveNN , I'm trying to build onlyoffice docker image on another machine and I had a question about Dockerfile. This machine is `arm64` arch, so the base image `ubuntu:20.04`...
ok @ShockwaveNN :smiley:, I built local images 'cause I tried to start onlyoffice service in another machine, but still failed, took me a long time to find the cause of...
thanks for reminding, just found the `latest-arm64` tag :sweat_smile:
Finally, it worked on my arm machine! ``` $ docker images onlyoffice/documentserver latest-arm64 bd99d7503a3b $ uname -ar Linux wesley-UOS 4.19.71-arm64-desktop #1 SMP PREEMPT Thu Jul 15 14:18:15 CST 2021 aarch64...
你说的应该是词汇表(Glossary)，不过chatgpt接口应该没有提供类似的参数
@yiran1211 分割成段落还是会造成冗余。不过按你的思路，书籍翻译需要分成三个阶段：1.先把整本书用NER工具扫一遍，需要识别人名地名专有名称，2.翻译好这些词汇做成词汇表，也就是你说的mapping tables。3.再用gpt进行翻译，充分利用gpt理解上下文的能力，提前设置好prompts，让gpt记住词汇表，用词汇表中的内容翻译匹配的名称。
想了想，NER工具也不是很靠谱，有一些需要特定翻译的语词这些工具应该扫不出来，比如古代一些官职名称
@DennySORA 希望译文不要另起一段，直接穿插在原文段落中，这样非常方便读者进行原文和译文的对照，我附了一张效果图，还有哪些我没有说清的点呢？
@DennySORA 我找到[一个解决办法](https://stackoverflow.com/a/75651428/1265973)，不过有点瑕疵