onemtbhashaverse

onemtbhashaverse

Supports

• Translation Across 36 × 36 Indian Language Pairs
• Discourse Translation
• Domain-Specific Translation
• Machine Translation Evaluation
• Translation Error Identification and Categorization
• Automatic Post-Editing (APE)

Cite

@misc{mujadia2024bhashaversetranslationecosystem,
      title={BhashaVerse : Translation Ecosystem for Indian Subcontinent Languages}, 
      author={Vandan Mujadia and Dipti Misra Sharma},
      year={2024},
      eprint={2412.04351},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2412.04351}, 
}

How to Run

LOCAL INSTALL

bash install_server.sh

LOCAL START

sh start.sh

Docker Build

bash docker_build.sh

Docker Run

bash docker_run.sh

Docker check

docker container ls

Docker Stop/Delete Container

docker rm -f

API URL

http://0.0.0.0:8084/onemtapi/v1/translateulca

Input (ULCA)

{
    "input": [
        {
            "source": "नमस्ते"
        }
    ],
    "config": {
        "modelId":101,
        "language": {
            "sourceLanguage": "hin",
            "targetLanguage": "eng"
            }
    }
}

Output (ULCA)

{
    "output": [
        {
            "source": "नमस्ते",
            "target": "Hi."
        }
    ],
    "config": {
        "modelId": 101,
        "language": {
            "sourceLanguage": "hin",
            "targetLanguage": "eng"
        }
    }
}

Training Code

• https://ssmt.iiit.ac.in/meitygit/ssmt/mt-model-deploy-dhruva/-/tree/master/training-code

Corpora

• https://huggingface.co/datasets/ltrciiith/bhashik-parallel-corpora-generic

supported language pairs

• 36*36