datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
parsinlu-machine-translation-en-fa-alpaca-style
ParsiNLU Machine Translation En-Fa in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the ParsiNLU original dataset.
parsinlu-machine-translation-fa-en-alpaca-style
ParsiNLU Machine Translation Fa-En in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the ParsiNLU original dataset.
chem-machine-translation-samplesshola-machine-translations
SHOLA machine translations
Machine translations of 478,822 everyday English words into the 55 African
languages Google Translate supports, from two systems:
Directory
System
Languages
Coverage
gemini-3.6-flash/
Gemini 3.6 Flash
4
all 478,822 words, 3 variants each
google-translate/
Google Translate
55
tier 1
gemma-4-31b-it/
Gemma 4 31B instruct, run on an H200 with vLLM
55
all tiers, in progress
Gemini came first and covers only Twi, Ewe, Ga and Dagbani - the… See the full description on the dataset page: https://huggingface.co/datasets/AfriSpeech/shola-machine-translations.