Training Large Neural Networks With Low-Dimensional Error Feedback

Re-assign community