Meta FLORES-200 (mr)

Meta FLORES-200 (mr)

MH Specific

Meta FLORES-200 (mr) dataset for language nlp.

Evaluate and benchmark Marathi machine translation models using FLORES-200 as a standardized test set.
HomepageHuggingFace

Quick Start

from datasets import load_dataset
ds = load_dataset('facebook/flores', 'mar_Deva', split='devtest')
for ex in ds[:5]:
    print(f"[{ex['id']}] {ex['sentence'][:80]}...")
Modality
parallel-text
Size
3,001 sentences, parallel across 200 languages
License
Format
CSV/JSON
Language
mr
Update Frequency
static
Organization
Meta AI
Category
language-nlp

Schema

FieldTypeDescription
sentencestringText sentence in Marathi (or source language)
idintSentence identifier aligned across languages

Build With This

Create a Marathi translation quality leaderboard comparing commercial and open-source MT systems on FLORES-200
Develop a fine-grained translation error analysis tool that identifies specific error types in Marathi MT output
Build a parallel corpus augmentation pipeline using FLORES-200 as seed data for mining additional Marathi translation pairs

AI Use Cases

MT evaluation benchmark
Last verified: 2026-03-07