Closed morningphase closed 2 years ago
What about np.array(low_dim_embs)
?
TSNEEmbedding
is actually a subclass of np.ndarray
, so you could treat it as such. If you want to get rid of the extra variables you can do np.array(low_dim_embs)
, as Dmitry has written, or you could do low_dim_embs.view(np.ndarray)
.
I want to use: low_dim_embs = tsne.fit(np.array(emb_matrix))
Actual behaviour is getting the class :TSNEEmbedding
what can I do