v0.0.1

Initial release.

Make sure to look at:

https://github.com/Kakusui/TLTMI/blob/main/README.md

Full Changelog: v0.0.1-alpha...v0.0.1

v0.0.1-alpha | Initial alpha release

v0.0.1-alpha - First Alpha Release

Overview

This is the first alpha release of TLTMI (That Local Translation Model Interface), a Dockerized translation service that allows users to run machine translation models locally. This release is a proof of concept, so expect some rough edges and potential performance issues.

What's New or First

• Initial Dockerized Service:

  • Docker image available on Docker Hub.
  • Basic FastAPI-based REST API for handling translation requests.
  • Initial support for multiple translation models from Helsinki-NLP/Opus-MT.

• Translation Capabilities:

  • Supports language pair switching for Helsinki-NLP/Opus-MT models.
  • Allows translation requests via the /translate endpoint.

• Local Setup Options:

  • Docker-based setup for easy deployment.
  • Development setup using Python and FastAPI for local testing.

• Basic Documentation:

  • README includes setup instructions, usage examples, and some humor to keep you entertained during setup.

Known Issues & Caveats

• Memory Usage:

  • Each language pair switch loads a new model into memory, which may result in high memory consumption. Exercise caution on machines with limited memory.
  • No performance optimizations have been implemented yet. Running multiple models simultaneously may lead to slowdowns or crashes.

• Inconsistent Language Codes:

  • Language codes are based on the Helsinki-NLP/Opus-MT models, which may be inconsistent. Currently, TLTMI does not validate these codes, so incorrect codes may lead to errors.

• Proof of Concept Status:

  • This is an early alpha release with limited functionality. Many features are planned but not yet implemented. Use at your own risk.

How to Pull & Use

• Pull the Docker Image:

  docker pull bikatr7/tltmi-core:v0.0.1-alpha

• Run the Docker Container:

  docker run -d -p 8000:8000 bikatr7/tltmi-core:v0.0.1-alpha

• Translation Request Example:

  curl -X POST "http://localhost:8000/translate" -H "Content-Type: application/json" -d '{"text": "Hello, world!", "source_lang": "en", "target_lang": "jap"}'

Future Plans

• Performance Optimizations:

  • Implementing model unloading and memory management to prevent excessive resource usage.

• Expanded Model Support:

  • Support for more powerful translation models in the future.

• Feature Enhancements:

  • Improved error handling, language code validation, and additional API endpoints.