search

quinngroup / dr1dl-pyspark

Dictionary Learning in PySpark
Apache License 2.0
1 stars 1 forks source link

Sparse vector representations #61

Open magsol opened 8 years ago

magsol commented 8 years ago

We should incorporate sparse representations wherever possible; at the very least, before sending data out onto the network. u and v could potentially benefit from this, but especially S as iterations proceed.

Initially, S could start as a dense matrix, but after a few iterations it could be converted to sparse.

magsol commented 8 years ago

v has been converted to a SparseVector.