Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

First class Vector/Tensor Datatype #6869

Open
Aetherall opened this issue Aug 10, 2024 · 0 comments
Open

First class Vector/Tensor Datatype #6869

Aetherall opened this issue Aug 10, 2024 · 0 comments
Assignees

Comments

@Aetherall
Copy link

Hi ! Such a great tool you built !

First of all, I am no data scientist. I am a backend developper and have no clue about what i am doing around data analysis.
I was fiddling to explore ways to extract domain knowledge from images, and I wanted to play with embeddings.
I already have a dataset of embeddings, and found no other way to use that than use 1 column per dimension of my embedding (1408).
This worked well and I did find the answer I was looking for. However, as I tried to see if I could optimize things, I found myself writing lots of python scripts ( I never worked with python until now so lots of dirty code written with copilot ) to do arithmetics over those columns, because the widgets were not designed to apply the same thing to 1408 columns.

I took a look at the codebase, and tried to add support for a vector datatype. I did succeed to make some stuff work, but it was requiring me to add code to every widget, so I know I am working in the wrong direction.

However in half a day of work I did end up with a pretty cool result.
image

It is buggy as hell, but the concept is here.

I might be completely mistaken about the way I am supposed to use the tool, but if I am not, maybe it will interest someone.
I can free up some time to work on this, but as I said, I never worked in python before although I have been programming in other languages for about 10 years, so I will need some guidance.

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants