I got a supervised UMAP transformation result on my dataset.The code is as below :
reducer = umap.UMAP(n_neighbors=40,min_dist=0.01,n_epochs=1000,learning_rate=0.01)
embedding = reducer.fit_transform(allData,y=allDX)
My question is this. How can I re-use the embedding for other dataset which is not used for this clustering ? I found that there is parametric UMAP, but think that is not what I want.
I want to save the result of this transform as a file such as *.ckpt or something and see where the new data goes after getting through same transformation.