Subset without embeddings and chunking
#3
by
ccdv
- opened
Hey
Is uploading an additional subset without embeddings and chunking planned?
The available subset is quite big.
Hello,
The goal of MediaTech is to publish pre-embedded datasets as open source, in order to facilitate their use in AI projects in the French civil service.
So it is not planned to upload an additional subset without embeddings and chunking.
However, you can still remove the embeddings column.
About the chunking, it is planned indeed to create a documentation on how to un-chunk MediaTech's datasets after downloading them.
Best,