My embeddings are 1536 or 3072 dimensions. Storage and search are expensive, and some indexes have dimension limits. How can I make them smaller?
Loading page content…

Find answers to your programming questions.
My embeddings are 1536 or 3072 dimensions. Storage and search are expensive, and some indexes have dimension limits. How can I make them smaller?
My dataset has categorical columns like 'city' and 'product type'. Can I use PCA on them?
Code that worked before now fails or prints FutureWarnings about learningrate, init, or niter.
I've seen all three used for dimensionality reduction. What's the difference, and when should I use each?
I have a 2D embedding and want a scatter plot colored by label, with a legend.
I have class labels. Can UMAP use them to separate the classes better?
Each UMAP run gives a different plot. When I set randomstate, I get a warning and it becomes much slower.
My UMAP plot shows one big blob (or many tiny fragments). Which parameters should I change?