Self-Compressing Neural Networks – Model size in the loss functiontwitter.com1 point·punnerud··0 commentsOpen articleSaveView on HN