ParthJain18/fluent_flow
一个使用 Flutter 和 Python 后端构建的 Android 应用，可实时翻译对话，并利用对话历史和逻辑回归模型判断当前对话是否需要回复建议。如果需要生成回复，则使用 Google 的 Gemini API 生成回复建议。
- Stars
- 4
- Forks
- 0
- 最近推送（UTC）
- 2024年5月19日
- 项目状态
- 未归档
Parth Jain ↗
技术话题
使用的依赖
依赖清单 15 项
- flutter
{"sdk":"flutter"} - provider
any - cupertino_icons
^1.0.8 - http
^1.2.1 - shared_preferences
^2.2.3 - speech_to_text
^6.5.1 - permission_handler
^11.3.1 - flutter_sound
^9.4.6 - path_provider
^2.1.3 - html
any - camera
any - html_unescape
any - http_parser
any - flutter_test开发依赖
{"sdk":"flutter"} - flutter_lints开发依赖
^3.0.0
原始 README
以下为英文项目原文快照，最新内容请访问 GitHub。
展开 / 收起项目 README
Fluent Flow
A language translation app built using Flutter and Python backend.
It translates live conversations from one language to another, then uses the history of the conversation and a logistic regression model to determine whether the current conversation needs any reply suggestions or not (Basically, we check if a LLM can be used to generate a reply suggestion or not, because if the conversation is personal, there is no need to generate any suggestions). If the conversation needs a generated reply, Google's Gemini API is used to generate possible reply suggestions.
The user can also generate a summary for the conversation if needed, and the app can be used to translate written text using the camera.
Screenshots
| Suggestions Example | Translation Example |
|---|---|
| WhatsApp Image 2024-04-04 at 23 44 28_7158d8e1 | WhatsApp Image 2024-04-04 at 23 46 53_f1590e79 |
| Summary Example | Camera Example |
| WhatsApp Image 2024-04-04 at 23 46 03_1f69797a | WhatsApp Image 2024-05-06 at 11 08 37_2264cdc5 |
| WhatsApp Image 2024-05-06 at 11 08 37_6bbe35da |
Flowchart
image