# How to develop small workable embeddings for debugging

Processing language data with deep learning models often involves a lookup of a pre-trained embedding model. In the model development phase, it's very annoying that every time the entire embeddings are loaded, as embeddings could be very big (e.g., Glove) and would consume a lot of timing loading it.

Is there a way to build/find a smaller and workable embedding just for debugging?

Right now what I've done is shrink the dimensions of Glove, for example, reduce the dimension of embeddings from 300d to 30d by taking the first 30 dims, but that might potentially induce a risk that some of words will share the same embedding.

• Do you mean small as in vocab? Small as in dimension? – kbrose Nov 6 '17 at 14:54
• Did you reduce dimensionality using PCA or something similar? Or just take the first 30 dimensions? – kbrose Nov 6 '17 at 14:55
• Small as in file size. I simply take first 30 dims. – Logan Nov 6 '17 at 23:08