也就是说上下文没能联系好,有时候会比较健忘。
对对,会给出一些似是而非的代码,可能是因为它训练的过程中,输入的内容就不是完全正确的。
哈哈,我觉得你们没用对。只能用来查东西,不能直接用它的代码,很多时候它会犯很简单的错误。
你应该是想说转文字比听语音快吧?我感觉阅读的时候,也不是一个词一个词地来,更像是一次性的扫描了整个句子。像这种乱序的文字,如果仔细去扫每个词,反而会很吃力。
尝试过让它帮忙写代码,是一个非常不错的工具。比如让它给出我想要的Nginx配置,给出的结果比我自己搜索半天找到的东西要有用得多。
睡觉了,大家晚安💤🌙
我现在过着的是美国时间,洛杉矶都已经是中午了啊。
都这么晚了,真的都不用睡的吗?😀
wss://universe.nostrich.land?lang=ja for Japanese,
And wss://universe.nostrich.land?lang=zh for Chinese,
Or wss://universe.nostrich.land?lang=zh&lang=ja for both😀
Would you like to read more notes in Japanese or Chinese? You can add my relays. Your account is already on the whitelist. 😀
I may understand that haiku is a kind of classical short poem in Japan. The haiku robot has been looking at global feed. As long as it finds haiku, it will be marked out.
俳句机器人是干嘛的?😅
The classification effect of long articles will be better, and short content needs to be pre-processed first.
It uses FastText to do the classification. I think it depends on the training model.

