Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
wit – Dataset by mteb | AlphaNeural AI
You can deploy this model and start earning money today!
mteb
/
wit
like
0
visual-document-retrieval
image-to-text
text-to-image
derived
multilingual
ara
bul
dan
ell
eng
est
ind
jpn
kor
tur
vie
cc-by-sa-4.0
1K<n<10K
parquet
image
Views
No views yet
Model card
Files and Versions
Community
API
WITT2IRetrieval An MTEB dataset Massive Text Embedding Benchmark
Retrieve images based on multilingual descriptions.
Task category t2i
Domains Encyclopaedic, Written
Reference
https://proceedings.mlr.press/v162/bugliarello22a/bugliarello22a.pdf
How to evaluate on this task
You can evaluate an embedding model on this dataset using the following code: import mteb
task = mteb.get_tasks(["WITT2IRetrieval"]) evaluator = mteb.MTEB(task)
model =… See the full description on the dataset page:
https://huggingface.co/datasets/mteb/wit
.
