chatGPT或者其它办法能给自动生成的字幕加上标点吗？(不要用whisper等使用本地GPU...
本帖最后由 blackantt 于 2024-4-15 09:42 编辑自动字幕往往没有标点，并且一个完整句子被放在2条 或者多条字幕里。 有啥办法能给它加上正确的标点呢？(不要用whisper等使用本地GPU的方法)
比如:
1
00:00:00,000 --> 00:00:04,000
Guess how much I love you
2
00:00:04,000 --> 00:00:11,000
One autumn morning little nut-brown hair big nut-brown hair and little field mouse
3
00:00:11,000 --> 00:00:18,000
sat bathed in the early morning light fascinated by one of Little White Owl's stories
4
00:00:18,500 --> 00:00:19,000
what a nice day!
如果不好弄，要不，问题简化一下吧。怎么样把已经加好标点的文本对应回原始没标点的srt文件里？见3楼的详细说明
.
.
. 我来瞅瞅吧！ 本帖最后由 blackantt 于 2024-4-15 09:37 编辑
goodzmq 发表于 2024-4-15 08:23
我来瞅瞅吧！
要不，问题简化一下吧。怎么样把已经加好标点的文本对应回原始没标点的srt文件里？
比如，原始文件 1.srt 如下:
1
00:00:00,000 --> 00:00:04,000
Guess how much I love you
2
00:00:04,000 --> 00:00:11,000
One autumn morning little nut-brown hair big nut-brown hair and little field mouse
3
00:00:11,000 --> 00:00:18,000
sat bathed in the early morning light fascinated by one of Little White Owl's stories
4
00:00:18,500 --> 00:00:19,000
what a nice day!
我先把它变成没有序号行，没有时间轴的文件行，如下:
Guess how much I love you
One autumn morning little nut-brown hair big nut-brown hair and little field mouse
sat bathed in the early morning light fascinated by one of Little White Owl's stories
what a nice day
然后对文本加标点如下:
Guess how much I love you.
One autumn morning, little nut-brown hair, big nut-brown hair and little field mouse
sat bathed in the early morning light, fascinated by one of Little White Owl's stories.
what a nice day!
现在，最后一步，怎么把已经有标点的(逗号，句号，问号，叹号等)文本对应写回原始时间轴里呢？ 这属于查找，定位，替换问题吗？ 有没有现成模块可用？ 没有的话，能否实现一下
就把你的需求，直接跟它说就好 不二如是 发表于 2024-4-18 19:41
就把你的需求，直接跟它说就好
怎么叫它忽略序号行及时间轴行呢？可以对文本加标点。但直接对srt加不了
页:
[1]