Open ntadimeti opened 4 years ago
This will allow for us to encode only non-zero data for both training and inference. Also makes the data more compact and self-contained.
This will allow for us to encode only non-zero data for both training and inference. Also makes the data more compact and self-contained.