增加识别内容的批量校对纠错，类似于翻译
Nobody has claimed this yet.
- Dominant language
- No language data
- Stars
- 1.1k
- Forks
- 108
- PR merge metrics
- No merged PRs in 30d
Description
Is your feature request related to a problem? Please describe.
中英混杂的时候，往往有些单词被识别成中文了，以及会有一些错别字。另外还有一些口吃、重复、语气词之类的。
Describe the solution you'd like
是不是可以用类似于翻译的方式，大批量地给一堆内容（更多上下文、更多语境），来进行纠正和优化。
Describe alternatives you've considered
目前是将内容扔给大模型，来给一段 Prompt 进行优化，但效率很低，且只能导出来改
Additional context
类似的处理方式
https://baoyu.io/blog/prompt-engineering/qa-how-to-correct-subtitle-with-text
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue names no files, tests, or entry points. Start by tracing the existing recognition, translation, and export flows to determine where batch text correction could fit; done should be defined as an agreed workflow that corrects large amounts of recognized content while preserving usable output.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, content
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100