Opus-mt-hi-ur Open Source Machine Translation Model - Free and Accurate Hindi to Urdu Translation

Opus Mt Hi Ur

Developed by Helsinki-NLP

This is a Hindi-to-Urdu machine translation model based on the transformer-align architecture, released by the Tatoeba-Challenge project.

Machine Translation

Transformers

OtherOpen Source License:Apache-2.0 #Low-resource translation #SentencePiece tokenization #Tatoeba benchmark

Downloads 41

Release Time : 3/2/2022

Model Overview

This model is specifically designed for translating Hindi text into Urdu, employing standardized preprocessing and SentencePiece tokenization techniques.

Model Features

Standardized preprocessing

Adopts a standardized preprocessing pipeline to ensure input text consistency and quality.

SentencePiece tokenization

Utilizes the spm4k tokenizer for efficient text processing.

Alignment transformer architecture

Based on the transformer-align architecture, optimizing alignment relationships between language pairs.

Model Capabilities

Text translation

Cross-language conversion

Hindi comprehension

Urdu generation

Use Cases

Language services

Document translation

Automatically translates Hindi documents into Urdu

BLEU score 12.4, chr-F score 0.393

Cross-language communication

Facilitates communication between Hindi and Urdu speakers

🚀 hin-urd Translation Model

This project focuses on the translation between Hindi and Urdu, providing a transformer-based alignment model with specific pre - processing steps and detailed benchmark results.

🚀 Quick Start

The hin - urd model is designed for Hindi to Urdu translation. You can download the original weights and test set translations from the provided links to start using the model.

✨ Features

Language Pair: Translates from Hindi (hin) to Urdu (urd).
Model Type: transformer-align model.
Pre - processing: Utilizes normalization and SentencePiece (spm4k, spm4k).

📦 Installation

To use this model, you need to download the original weights: opus-2020-06-16.zip

💻 Usage Examples

Although no specific code examples are provided, you can use the downloaded model weights to perform Hindi to Urdu translation tasks according to the general machine - translation framework.

📚 Documentation

Model Details

Source Group: Hindi
Target Group: Urdu
OPUS Readme: hin-urd

Benchmarks

Testset	BLEU	chr - F
Tatoeba - test.hin.urd	12.4	0.393

System Info

Property	Details
hf_name	hin - urd
source_languages	hin
target_languages	urd
opus_readme_url	https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/hin-urd/README.md
original_repo	Tatoeba - Challenge
tags	['translation']
languages	['hi', 'ur']
src_constituents	{'hin'}
tgt_constituents	{'urd'}
src_multilingual	False
tgt_multilingual	False
prepro	normalization + SentencePiece (spm4k, spm4k)
url_model	https://object.pouta.csc.fi/Tatoeba-MT-models/hin-urd/opus-2020-06-16.zip
url_test_set	https://object.pouta.csc.fi/Tatoeba-MT-models/hin-urd/opus-2020-06-16.test.txt
src_alpha3	hin
tgt_alpha3	urd
short_pair	hi - ur
chrF2_score	0.39299999999999996
bleu	12.4
brevity_penalty	1.0
ref_len	1618.0
src_name	Hindi
tgt_name	Urdu
train_date	2020 - 06 - 16
src_alpha2	hi
tgt_alpha2	ur
prefer_old	False
long_pair	hin - urd
helsinki_git_sha	480fcbe0ee1bf4774bcbe6226ad9f58e63f6c535
transformers_git_sha	2207e5d8cb224e954a7cba69fa4ac2309e9ff30b
port_machine	brutasse
port_time	2020 - 08 - 21 - 14:41

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご