Batch the embeddings computation.

[fissoreg]

If we're gonna index a lot of data, we will want to speed up the computation of embeddings. This can be done by proper batching, but some testing (for performance gain) is needed.

Useful links:

• https://mccormickml.com/2020/07/29/smart-batching-tutorial/
• https://github.com/jalammar/jalammar.github.io/blob/master/notebooks/bert/A_Visual_Notebook_to_Using_BERT_for_the_First_Time.ipynb