Add French Language benchmark

#78
by OmarMorsli - opened

Hello, are you expecting to add French to the leaderboard ?

Massive Text Embedding Benchmark org

@Muennighoff cool! Do you have any recap xlsx ou csv tables? Especially for Retrieval ?

Massive Text Embedding Benchmark org

@Muennighoff cool! Do you have any recap xlsx ou csv tables? Especially for Retrieval ?

Hello,

I think you can build the CSV/XLSX file from the JSON files that can be found in mteb/results repository (https://huggingface.co/datasets/mteb/results). You just have to clone the repo, retrieve all French retrieval task files (tasks and their files have the same name, e.g, for AlloProfRetrieval task, the associated file is named AlloProfRetrieval.json). Then read all files and retrieve the main metric for French.

In the meantime, I can check of we can provide you with a CSV file from our experiments. I'll let you know if it's possible.

Hello @imenelydiaker , thank you
It's already done. I was looking for the raw JSON files, and then I found them on your GitHub.

imenelydiaker changed discussion status to closed

Sign up or log in to comment