Hi all,
I have a small issue: I’ve a numpy array of a shape of (1700, 768) - 1700 correspond to the number of images I have in my dataset and 768 to the dimension of the word vector associated to them. I want to plot the images with respect to the embedding vectors. I already used UMAP to go from 768 to 2 D and plotted a scatter plot, but I want to plot actual images, not points. Has anyone done something like that before?
Thanks!