Dataset assetOpen Source CommunityMachine Translation

haoranxu/WMT22-Test

The dataset provides configurations for multiple language pairs, including cs‑en (Czech‑English), de‑en (German‑English), en‑cs, en‑de, en‑is (English‑Icelandic), en‑ru, en‑zh, is‑en, ru‑en, and zh‑en. For each configuration, features consist of string columns for the two languages, and a test split with specified byte size and number of examples. The dataset is intended for machine translation tasks.

Source

hugging_face

Created

Nov 28, 2025

Updated

Jan 17, 2024

Signals

105 views

Availability

Linked source ready

Overview

Dataset description and usage context

Dataset Overview

Configurations

cs‑en

Features:
- cs: string
- en: string
Split:
- test: 325,040 bytes, 1,448 samples
Download size: 224,193 bytes
Dataset size: 325,040 bytes

de‑en

Features:
- de: string
- en: string
Split:
- test: 403,424 bytes, 1,984 samples
Download size: 267,107 bytes
Dataset size: 403,424 bytes

en‑cs

Features:
- cs: string
- en: string
Split:
- test: 422,875 bytes, 2,037 samples
Download size: 281,086 bytes
Dataset size: 422,875 bytes

en‑de

Features:
- de: string
- en: string
Split:
- test: 442,576 bytes, 2,037 samples
Download size: 280,415 bytes
Dataset size: 442,576 bytes

en‑is

Features:
- en: string
- is: string
Split:
- test: 310,807 bytes, 1,000 samples
Download size: 197,437 bytes
Dataset size: 310,807 bytes

en‑ru

Features:
- en: string
- ru: string
Split:
- test: 598,414 bytes, 2,037 samples
Download size: 333,784 bytes
Dataset size: 598,414 bytes

en‑zh

Features:
- en: string
- zh: string
Split:
- test: 383,751 bytes, 2,037 samples
Download size: 257,805 bytes
Dataset size: 383,751 bytes

is‑en

Features:
- en: string
- is: string
Split:
- test: 248,029 bytes, 1,000 samples
Download size: 152,885 bytes
Dataset size: 248,029 bytes

ru‑en

Features:
- en: string
- ru: string
Split:
- test: 579,656 bytes, 2,016 samples
Download size: 340,830 bytes
Dataset size: 579,656 bytes

zh‑en

Features:
- en: string
- zh: string
Split:
- test: 526,074 bytes, 1,875 samples
Download size: 333,078 bytes
Dataset size: 526,074 bytes

Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio